<?xml version="1.0" encoding="UTF-8"?><rss xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:content="http://purl.org/rss/1.0/modules/content/" xmlns:atom="http://www.w3.org/2005/Atom" version="2.0" xmlns:cc="http://cyber.law.harvard.edu/rss/creativeCommonsRssModule.html">
    <channel>
        <title><![CDATA[Stories by QDuality on Medium]]></title>
        <description><![CDATA[Stories by QDuality on Medium]]></description>
        <link>https://medium.com/@Qduality?source=rss-2bf9c8b9f1ed------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/proxy/1*TGH72Nnw24QL3iV9IOm4VA.png</url>
            <title>Stories by QDuality on Medium</title>
            <link>https://medium.com/@Qduality?source=rss-2bf9c8b9f1ed------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Sat, 21 Nov 2026 23:09:38 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@Qduality/feed" rel="self" type="application/rss+xml"/>
        <webMaster><![CDATA[yourfriends@medium.com]]></webMaster>
        <atom:link href="http://medium.superfeedr.com/" rel="hub"/>
        <item>
            <title><![CDATA[Future of Quantum Computing & Real-world Applications]]></title>
            <link>https://medium.com/@Qduality/future-of-quantum-computing-real-world-applications-f17bbc404483?source=rss-2bf9c8b9f1ed------2</link>
            <guid isPermaLink="false">https://medium.com/p/f17bbc404483</guid>
            <dc:creator><![CDATA[QDuality]]></dc:creator>
            <pubDate>Sun, 21 Sep 2025 04:27:37 GMT</pubDate>
            <atom:updated>2026-09-13T04:57:00.020Z</atom:updated>
            <content:encoded><![CDATA[<h3>Qucode: Day 21/21; QCC; Cohort-3.</h3><p>Final post in this QCC Cohort 3 series! We are on our way to completing the last and final day of the QCC! I will update and modify many of my existing posts in this series as I dive deeper into these concepts over a period of time.</p><p>We have touched a lot of quantum concepts in a short period of time over the last 3 weeks. Was it worth it? Does quantum computing have a future? Are there real problems that we could actually solve using quantum computing? How do we get started? All of these are questions that would be lingering on the minds of everyone starting their quantum journey. There are limitless improvements to make on existing quantum computers, algorithms to invent, and goals to accomplish in this field, so there is plenty of opportunities for everyone out there. You don’t have to worry about starting the journey late because there are no limits to quantum innovation.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/593/1*s_T6ZmMOF58O9G7M0Ji1_Q.png" /></figure><h3><strong>In Summary, quantum computing</strong> is the new revolution in computation, moving beyond the binary logic of classical digital computers. While classical computers operate using bits that are either a 0 or a 1, quantum computers use quantum bits, or qubits. Qubits can exist in a state of superposition, meaning they can be both 0 and 1 at the same time. This is a fundamental difference that allows quantum computers to perform certain calculations with significantly fewer steps.</h3><p>The power of quantum computing comes from three key principles:</p><ul><li><strong>Superposition:</strong> A qubit can be a combination of both 0 and 1 states simultaneously. When a qubit is measured, it collapses into either a 0 or a 1, with the probability of each outcome determined by its state. This is often visualized as an arrow in 3D space, where the direction of the arrow determines the probabilities.</li><li><strong>Entanglement:</strong> Qubits can become linked, or “entangled,” in a way that the state of one qubit is dependent on the state of another, no matter how far apart they are. This allows for a massive increase in the number of possible states for a system as each additional entangled qubit doubles the number of states the system can represent. An N-qubit system can be in a superposition of two to the power of N states at once.</li><li><strong>Interference:</strong> The wave-like nature of quantum mechanics allows the probabilities of different outcomes to either add together (constructive interference) or cancel each other out (destructive interference). A quantum algorithm is designed to use this principle to increase the probability of the correct answer and decrease the probability of incorrect answers.</li></ul><p>While quantum computers are not yet powerful enough to solve real-world problems more efficiently than classical computers, they have the potential to solve problems that are currently considered intractable. For example, Shor’s algorithm can efficiently find the prime factors of large numbers, a task that is exponentially difficult for classical computers and is the basis for much of today’s internet encryption. Other promising applications involve the use of quantum simulation softwares or using a quantum computer to model quantum systems like chemical reactions or the behavior of materials. This could lead to breakthroughs in fields such as medicine, energy, and materials science.</p><p>The main challenges in building a practical quantum computer are <strong>DECOHERENCE</strong> and <strong>NOISE</strong>. Decoherence occurs when qubits interact with their environment and lose their quantum properties. To combat this, quantum computers must be cooled to temperatures near absolute zero and shielded from all forms of interference. In Quantum Communication, the noise comes from sources like external interactions and not necessarily due to temperatures. Plenty of error codes, error correction schemes are being developed that use many physical qubits to create a single, more stable logical qubit. The path to building a large, error-corrected quantum computer is still a work in progress, but steady advancements are being made toward this goal. Are you ready to invest your brain in this new revolution?</p><p>Cheers!</p><blockquote><strong><em>Tags</em></strong></blockquote><p>#QuantumComputing #QuantumLearning #Probability #Statistics #BayesTheorem #Qohort3 #STEM #QuCode #QuantumIndia #NationalQuantumMission #DeepTech #FutureOfComputing #QuantumStartups #ResearchAndInnovation #LearnQuantum #QuantumAlgorithms #QuantumCommunity #NextGenTech #DigitalIndia #TechForGood</p><blockquote><strong><em>Resources/References</em></strong></blockquote><p>This article was written by understanding content from:</p><ul><li><a href="https://youtu.be/qQviI1d_hFA?si=Ypoyj0v7KCftOieG">https://youtu.be/qQviI1d_hFA?si=Ypoyj0v7KCftOieG</a></li><li><a href="https://youtu.be/-UlxHPIEVqA?si=jTVb1o_I7BaFDslE">https://youtu.be/-UlxHPIEVqA?si=jTVb1o_I7BaFDslE</a></li><li><a href="https://youtu.be/UtDllX_MTbw?si=NV930uRpt0y8D7Qk">https://youtu.be/UtDllX_MTbw?si=NV930uRpt0y8D7Qk</a></li></ul><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=f17bbc404483" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Noise, Error Correction & NISQ Era Computing]]></title>
            <link>https://medium.com/@Qduality/noise-error-correction-nisq-era-computing-6786843813f0?source=rss-2bf9c8b9f1ed------2</link>
            <guid isPermaLink="false">https://medium.com/p/6786843813f0</guid>
            <dc:creator><![CDATA[QDuality]]></dc:creator>
            <pubDate>Sat, 20 Sep 2025 17:07:58 GMT</pubDate>
            <atom:updated>2026-09-13T04:57:39.732Z</atom:updated>
            <content:encoded><![CDATA[<h3>Qucode: Day 20/21; QCC; Cohort-3.</h3><p>Welcome to Day 20! We are in the penultimate day of the QCC.</p><p>Let’s get started with Noise, Error correction, and NISQ.</p><p>We have discussed a lot about the building blocks of quantum computers like gates, circuits, and also the quantum phenomenon like superposition, entanglement, etc. Building a quantum computer isn’t all smooth sailing. While the potential is huge, there are still significant challenges to overcome.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*dLBXRhGavctgRNFRtG0kOQ.png" /></figure><h3>Creating Qubits</h3><p>There are several physical systems that can be engineered to behave as qubits. There are nearly more than a dozen of them, so for simplicity sake let us briefly recap the photon based, ion and atom based, as well as the more popular and on-demand superconducting based qubit systems.</p><ul><li><strong>Photons</strong> : A single particle of light can be used. When a photon passes through a half-silvered mirror (a beam splitter), it enters a superposition of being both transmitted and reflected. These two pathways can represent the 0 and 1 states of a qubit. Other properties of photons, such as their <strong>polarization</strong>, can also encode qubit information.</li><li><strong>Ions and Atoms</strong> : The energy levels of electrons within an atom or ion can serve as the 0 and 1 states. Lasers are used to control and manipulate these energy levels, allowing for the creation of superpositions.</li><li><strong>Superconductors</strong>: This approach involves creating “artificial atoms” using superconducting circuits. These circuits are designed to exhibit quantum mechanical properties similar to natural atoms, with engineered energy levels that function as qubits. This method is popular among major tech companies and startups in the field.</li></ul><h3>Decoherence: The Fragility of Qubits and Errors</h3><p>A significant challenge in building quantum computers is the <strong>fragility of qubits</strong>. Qubits in a superposition are highly sensitive to their environment. Unwanted interactions with the environment, referred to as <strong>noise</strong>, can easily destroy the qubit’s quantum state. This process, known as <strong>decoherence</strong>, effectively “measures” the qubit prematurely, causing it to collapse into a definite 0 or 1 state, thus destroying the valuable superposition and entanglement needed for quantum computation. This fragility was once thought to be a fundamental roadblock to building useful quantum computers, as increasing the number of qubits would seemingly introduce more errors.</p><h3>Quantum Error Correction and Fault Tolerance</h3><p>Despite the inherent fragility, a solution exists in the form of <strong>quantum error correction (QEC)</strong>.</p><h3>How Quantum Error Correction Works</h3><p>QEC draws an analogy from classical error correction, such as the redundancy in language that allows us to understand words even with typos. In QEC, the principle is to encode quantum information in a robust way:</p><ul><li><strong>Encoding</strong>: Instead of storing information on a single “physical qubit,” the data from these qubits is distributed and spread across a larger group of qubits, including additional “ancilla qubits.” This collective ensemble of physical qubits forms a <strong>logical qubit</strong>. This process provides redundancy, protecting the quantum information from localized errors.</li><li><strong>Measuring and Correcting Errors</strong>: The logical qubit is then strategically measured. Crucially, these measurements are designed to extract information about <em>what errors have occurred</em> without revealing the underlying quantum data itself, which would cause the data to collapse. Based on the measurement outcomes, a decoding process identifies the specific errors, and then quantum gates are applied to the physical qubits to correct these errors. This cycle of measurement, decoding, and correction is continuously performed to prevent errors from accumulating.</li></ul><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*2Jv6Ck_7k06IDwzb6JS_sg.png" /></figure><h3>Achieving Fault Tolerance</h3><p>While QEC makes quantum computation possible, a critical property for building large-scale quantum computers is <strong>fault tolerance</strong>. Fault tolerance means that as the size and complexity of the quantum computer increase (i.e., as more qubits are incorporated into the error-correcting code), the overall error rate of the <strong>logical qubit</strong> actually <em>decreases</em>.</p><p>There is a <strong>critical noise level</strong>, often called the <strong>threshold</strong>, for physical qubits. If the error rate of individual physical qubits can be kept below this threshold, then it is possible to achieve fault tolerance. This ensures that even with noisy physical components, the overall logical error rate can be made arbitrarily small by scaling up the error correction scheme.</p><h3>Resource Demands</h3><p>Implementing quantum error correction and achieving fault tolerance is highly <strong>resource-intensive</strong>. A single logical qubit requires many physical qubits, leading to significant complexity and resource demands for building a fault-tolerant quantum computer compared to simply working with a few physical qubits.</p><p>Ultimately, the development of quantum error correction and the concept of fault tolerance have transformed quantum computing from a theoretical possibility into a realistic engineering challenge, paving the way for future quantum applications.</p><h3>The Imperfections of Qubits and the Need for Error Correction</h3><p>Any physical implementation of a qubit, whether it’s based on superconducting circuits, trapped ions, or other technologies, will inevitably deviate from the theoretical ideal. These imperfections arise from several factors:</p><ul><li><strong>Non-ideal Physical Systems</strong>: Real-world quantum systems used to create qubits are rarely perfect two-level systems. For instance, superconducting qubits are often built from anharmonic oscillators that have more than two energy levels, requiring careful isolation of the desired two states (ground and first excited).</li><li><strong>Energy Differences</strong>: Unlike the abstract ideal, physical qubits often have different energies for their 0 and 1 states. This can lead to spontaneous transitions from higher energy states (1) to lower energy states (0), causing errors.</li><li><strong>Imperfect Gates</strong>: The quantum operations (gates) applied to qubits are not perfectly accurate, introducing errors during computation.</li><li><strong>Environmental Noise</strong>: Qubits constantly interact with their environment, leading to unwanted effects like radiation, thermal noise, and interactions with other qubits on the chip.</li><li><strong>Measurement Errors</strong>: The process of measuring a qubit is not always perfect, and measurements can sometimes yield incorrect results or leak information about neighboring qubits.</li></ul><p>These errors accumulate over time, especially in multi-qubit systems. Even simple algorithms like Grover’s search algorithm show a significant degradation in accuracy with just a few qubits due to noise, making the results unreliable. This necessitates methods to detect and correct errors to ensure the integrity of quantum computations.</p><h3>Understanding and Detecting Quantum Errors</h3><p>Any complex error on a qubit can be broken down into a superposition of simpler errors, specifically the <strong>Pauli operators</strong>: Identity (no error), Pauli-X (bit flip), Pauli-Y (bit and phase flip), and Pauli-Z (phase flip). Consequently, if we can detect and correct bit flips and phase flips, we can effectively address any single-qubit error.</p><ul><li><strong>Bit Flip (Pauli-X) Error</strong>: This error swaps the 0 and 1 states, analogous to flipping a classical bit.</li><li><strong>Phase Flip (Pauli-Z) Error</strong>: This error flips the sign (phase) of the superposition, for example, changing a |+⟩ state to a |-⟩ state. A Pauli-Y error is a combination of X and Z.</li></ul><h3>Classical Error Correction Analogy: The Repetition Code</h3><p>To understand QEC, it’s helpful to consider classical error correction. If a single bit of information (e.g., “yes” or “no”) is transmitted over a noisy channel, it can be corrupted. A simple solution is the <strong>repetition code</strong>: to send “no,” one would repeat “no, no, no.” If one of the repetitions is flipped by noise (e.g., “no, yes, no”), a majority vote can still recover the original message. The probability of an incorrect decoding decreases exponentially with the number of repetitions, provided the individual bit error probability is below a certain threshold.</p><p>This classical method involves:</p><ol><li><strong>Input</strong>: The information to be protected.</li><li><strong>Encoding</strong>: Transforming the information (e.g., repeating it).</li><li><strong>Errors</strong>: Random perturbations during transmission or storage.</li><li><strong>Decoding</strong>: Deducing the original input from the perturbed message.</li></ol><p>For dynamic computation, where errors occur continuously, this process involves constant decoding and re-encoding or, more efficiently, detecting and correcting errors as they arise.</p><h3>Challenges of Applying Classical Methods to Quantum Systems</h3><p>Directly applying the classical repetition code to qubits has a critical flaw: If a qubit in a superposition state (e.g., a|0⟩ + b|1⟩) is encoded using repetition (e.g., a|000⟩ + b|111⟩), directly measuring the individual qubits to detect bit flips would collapse the superposition, destroying the quantum information. This measurement-induced collapse is a fundamental aspect of quantum mechanics.</p><h3>Quantum Error Correction: Preserving Superposition</h3><p>To overcome this, QEC employs <strong>syndrome measurements</strong>. Instead of measuring the individual qubits directly, which would reveal their specific 0 or 1 states and collapse superposition, syndrome measurements extract information about the <em>relationships</em> between qubits (e.g., their parity) without revealing the individual qubit states.</p><ul><li><strong>Parity Measurement</strong>: This involves comparing pairs of qubits to determine if they are the same or different. For an encoded logical qubit (e.g., |000⟩ or |111⟩), an error will cause a pair of physical qubits to become different. These “differences” act as “flags” or “syndromes” that indicate an error location without revealing the encoded value itself.</li><li><strong>Auxiliary Qubits</strong>: Special “ancilla qubits” are used to perform these parity measurements. Controlled operations (like CNOT gates) transfer information about the parity of data qubits onto the ancilla qubits. Measuring the ancilla qubits then reveals the syndrome (e.g., 0 for no difference, 1 for a difference) without directly interacting with the data qubits in a way that collapses their superposition.</li><li><strong>Decoding Process</strong>: The syndrome measurements provide “clues” about where errors have occurred. A decoding algorithm processes these clues (which can be complex, especially over time and across a 2D or 3D grid of qubits) to infer the most likely set of errors. This often involves techniques like <strong>minimum weight perfect matching</strong> on a graph, where defects are nodes and potential error paths are weighted edges. The goal is to find the simplest explanation (minimum number of errors) for the observed syndromes.</li></ul><h3>Logical Operations and Limitations of Simple Codes</h3><p>Once errors are detected and corrected, it’s also necessary to perform computations on the protected logical qubits. Some logical operations are straightforward:</p><ul><li><strong>Logical X (Bit Flip)</strong>: For a simple repetition code, a logical bit flip can be performed by applying a physical X gate to all the constituent physical qubits. The imperfections in these physical X gates simply introduce more errors that the QEC scheme can detect and correct.</li></ul><p>However, many other crucial operations, such as <strong>rotations</strong>, are not easily implemented in a fault-tolerant manner with simple repetition codes. Performing such operations might require “taking the code apart,” performing the rotation on the unprotected qubits, and then re-encoding, leaving the quantum information vulnerable to errors during the unprotected phase.</p><p>Furthermore, the simple repetition code only protects against <strong>bit flip errors</strong> and not <strong>phase flip errors</strong>. While it preserves superpositions by avoiding direct measurement, it doesn’t actively correct phase errors. For a universal quantum computer, protection against both types of errors is essential.</p><h3>Moving Towards Universal Quantum Error Correction</h3><p>To achieve truly fault-tolerant and universal quantum computation, more sophisticated quantum codes are required. These codes must treat bit-flip and phase-flip errors symmetrically and allow for a universal set of fault-tolerant logical gates. Examples include the <strong>surface code</strong> and <strong>Shor’s 9-qubit code</strong>. The latter concatenates a code that protects against bit flips with a code that protects against phase flips to offer comprehensive protection.</p><p>While full-fledged QEC with its high qubit overhead is still a goal for the future, simpler “error mitigation” techniques are being explored in the near term to improve the performance of current noisy quantum devices. Researchers are actively using codes like the repetition code as <strong>benchmarking tools</strong> on existing quantum hardware to understand and characterize the noise properties of real qubits and validate the fundamental principles of QEC. These experiments show that even on current devices, the probability of errors decreases exponentially with increasing code distance (number of repetitions), confirming the promise of quantum error correction.</p><blockquote><strong><em>Tags</em></strong></blockquote><p>#QuantumComputing #QuantumLearning #Probability #Statistics #BayesTheorem #Qohort3 #STEM #QuCode #QuantumIndia #NationalQuantumMission #DeepTech #FutureOfComputing #QuantumStartups #ResearchAndInnovation #LearnQuantum #QuantumAlgorithms #QuantumCommunity #NextGenTech #DigitalIndia #TechForGood</p><blockquote><strong><em>Resources/References</em></strong></blockquote><p>This article was written by understanding content from:</p><ul><li><a href="https://youtu.be/c5-HhsS1orw?si=dZXloVyjx7Sd8Iea">https://youtu.be/c5-HhsS1orw?si=dZXloVyjx7Sd8Iea</a></li><li><a href="https://youtu.be/AuDfq7j_W7E?si=Z2MIIpZMYw3F9pry">https://youtu.be/AuDfq7j_W7E?si=Z2MIIpZMYw3F9pry</a></li></ul><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=6786843813f0" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Quantum Machine Learning Basics]]></title>
            <link>https://medium.com/@Qduality/quantum-machine-learning-basics-7516c525b029?source=rss-2bf9c8b9f1ed------2</link>
            <guid isPermaLink="false">https://medium.com/p/7516c525b029</guid>
            <dc:creator><![CDATA[QDuality]]></dc:creator>
            <pubDate>Fri, 19 Sep 2025 17:43:03 GMT</pubDate>
            <atom:updated>2026-09-13T04:58:21.608Z</atom:updated>
            <content:encoded><![CDATA[<h3>Qucode: Day 19/21; QCC; Cohort-3.</h3><p>Welcome to Day 19!</p><p><strong>Quantum Machine learning </strong>which is a relatively new branch of quantum is an amalgamation of Quantum and AI, where extensive research is being done across the globe.</p><p>Imagine a regular neural network that classifies certain images as digits, represented by a complex network of activation functions, weights, and tonnes of learning parameters. Now, extending that to quantum, one would have to encode the data and use it in combination with quantum circuits before making our measurements, and predictions. By harnessing quantum principles, quantum machine learning can drastically accelerate the training process and enable the use of a much larger number of features and parameters in a model, surpassing the limitations of conventional machine learning.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/320/1*n6jknAGkzoh5Duia7Y8RRg.gif" /><figcaption>source: makeagif.com</figcaption></figure><h3><strong>— Quantum Neural Networks (QNN)</strong> Intro —</h3><p>A quantum neural network is a novel approach to machine learning that integrates quantum computing principles with the structure of classical neural networks. The goal is to leverage quantum phenomena like superposition and entanglement to potentially achieve computational advantages for certain machine learning tasks.</p><p>The general framework for quantum machine learning, including QNNs, typically involves the following steps:</p><p><strong>Data Encoding</strong>: Classical data, which is usually in a numerical format, must first be transformed into a quantum state that can be processed by a quantum computer. This involves mapping the data onto qubits. A common technique is <strong>angle encoding</strong>, where each numerical feature of a data point is used to determine the angle of rotation for a corresponding qubit. For example, if you have a data point represented by a vector (x1​,x2​,…,xN​), you might apply a rotation operation to the i-th qubit based on the value of xi​.</p><p><strong>Quantum Model Processing</strong>: Once the data is encoded, a <strong>parameterized quantum circuit</strong> (often referred to as an “<strong>ansatz</strong>”) is applied to the qubits. This circuit consists of a sequence of quantum gates, some of which have adjustable parameters. These parameters are analogous to the weights and biases in a classical neural network. The circuit typically includes various types of quantum operations:</p><ul><li><strong>Single-qubit rotations</strong>: These gates rotate individual qubits around specific axes (e.g., X, Y, or Z axes on the Bloch sphere). The rotation angles are often the parameters that are optimized.</li><li><strong>Entanglement gates</strong>: These gates, such as CNOT (controlled-NOT) gates, create entanglement between qubits. Entanglement is a unique quantum correlation that is believed to be essential for quantum speedups. An example of a variational model might involve layers of single-qubit rotations followed by layers of entanglement gates, with the rotation angles being the trainable parameters.</li></ul><p><strong>Measurement</strong>: After the quantum circuit has processed the data, a measurement is performed on the qubits. This step is crucial because it extracts classical information from the quantum state. Importantly, the measurement process introduces the necessary <strong>non-linearity</strong> into the QNN. Unlike classical neural networks that use non-linear activation functions directly within their layers, the fundamental operations on qubits (unitary transformations) are linear. The act of measuring a quantum state, which yields probabilistic outcomes, is inherently non-linear and provides the expressive power needed for machine learning. For instance, measuring a qubit multiple times will yield a probability distribution for it being in certain states (e.g., state ∣0⟩ or state ∣1⟩).</p><p><strong>Classical Interpretation and Optimization</strong>: The classical measurement results are then interpreted to form a prediction or a classification label for the given machine learning task. This prediction is compared against the actual target label using a <strong>cost function</strong> (also known as a loss function), which quantifies the error of the model’s prediction. A classical optimizer, such as <strong>gradient descent</strong>, is then employed to adjust the trainable parameters within the quantum circuit to minimize this cost function. This creates a hybrid quantum-classical feedback loop, where the quantum computer performs the complex transformations and the classical computer handles the optimization.</p><p>One significant challenge in the field of quantum machine learning is the <strong>Barren Plateau problem</strong>. This problem describes a phenomenon where the cost landscape (the function that the classical optimizer tries to minimize) becomes extremely flat in high-dimensional parameter spaces. When the landscape is flat, the gradients (directions for adjusting parameters) become very small, making it incredibly difficult for classical optimizers to find the optimal parameters and train the QNN effectively. Despite these challenges, quantum neural networks remain a promising area of research with the potential to unlock new capabilities in machine learning.</p><h3>Advanced Details of Quantum Neural Networks</h3><p>Beyond the core concept of a hybrid quantum-classical model, a deeper understanding of Quantum Neural Networks (QNNs) involves several specific architectural and operational choices:</p><p><strong>Diverse Data Encoding Methods</strong> While using rotation angles to encode data is a common practice, it is not the only method. Other techniques exist to map classical data onto a quantum state, including:</p><blockquote><strong>binary embedding</strong> and <strong>amplitude embedding</strong>.</blockquote><p>These methods offer different ways to represent information within the quantum system, which can be more suitable for various types of data and applications.</p><p><strong>Architecture of Hidden Layers</strong> The “hidden layers” of a QNN, which perform the main data processing, are constructed from two primary components:</p><ul><li><strong>Parameterized Rotational Blocks</strong>: These are sets of quantum gates (like rotations around the X, Y, or Z axes) with adjustable parameters. These parameters are the trainable weights of the network.</li><li><strong>Entanglement Blocks</strong>: These gates, such as CNOT gates, create entanglement between qubits. They are crucial for introducing complex quantum correlations that cannot be replicated classically.</li></ul><p>The way these blocks are connected can vary, leading to different entanglement architectures, including <strong>linear</strong>, <strong>circular</strong>, and <strong>full entanglement</strong>. The choice of entanglement structure can significantly impact the network’s expressive power. Some researchers even omit the entanglement blocks, focusing solely on the rotational gates as the trainable part of the network.</p><p><strong>Role of Measurement in Output Extraction</strong> The final step of a QNN is to extract a classical output from the quantum state. This is more nuanced than just obtaining a probability distribution. The measurement can yield several types of classical information, including:</p><ul><li>A <strong>probability vector</strong></li><li>A <strong>measurement bit string</strong></li><li>An <strong>expectation value</strong> (the average outcome of a measurement)</li></ul><p><strong>Prediction</strong>. This classical output then goes through a final pre-processing stage to produce the network’s prediction. The ability to use different types of measurement-based outputs provides flexibility in how the QNN’s results are interpreted for various machine learning tasks.</p><p>The QISKIT GLOBAL SUMMER SCHOOL 2024 explains QML very well. Let us now borrow Gemini’s explanation of that video to make life easier for us.</p><h3>Introduction to Machine Learning</h3><p>Machine learning is a field of artificial intelligence where algorithms learn patterns from data to perform tasks without explicit programming. Mathematically, it’s viewed as an <strong>optimization and function approximation problem</strong>. The goal is to find an approximate function, f(x,θ), that closely mimics a true, unknown function G(x). Here, x represents the input data (features), and θ denotes the model’s trainable parameters (weights).</p><p>Machine learning algorithms are categorized into:</p><ul><li><strong>Supervised Learning</strong>: Models learn from <strong>labeled data</strong> to make predictions.</li><li><strong>Classification</strong>: Predicts discrete categories (e.g., “flower” or “no flower”).</li><li><strong>Regression</strong>: Predicts continuous values (e.g., house prices).</li><li><strong>Unsupervised Learning</strong>: Models find patterns in <strong>unlabeled data</strong>. Examples include clustering, dimensionality reduction, and generative models (like GANs or VAEs).</li><li><strong>Reinforcement Learning</strong>: An agent learns to make decisions by interacting with an environment, receiving rewards or penalties to maximize cumulative reward.</li></ul><p>The general workflow for supervised learning involves:</p><ol><li><strong>Data Set</strong>: Input features Xtrain​ and true labels y.</li><li><strong>Model Prediction</strong>: The model f(xtrain​,θ) produces predicted labels y^​.</li><li><strong>Loss Function</strong>: A metric that quantifies the difference between y^​ and y. A common example is the Mean Squared Error (MSE):</li></ol><figure><img alt="" src="https://cdn-images-1.medium.com/max/464/1*sCDWzxwXJ1TXevNCNQYEtA.png" /></figure><p>4. <strong>Optimization</strong>: Techniques like <strong>gradient descent</strong> are used to adjust θ in a direction that minimizes the loss function. This involves computing gradients of the loss with respect to the parameters.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/468/1*6XDjw0QMqrvGM610JQGZxw.png" /></figure><p>where η is the learning rate.</p><p>A crucial concept is the <strong>bias-variance trade-off</strong>, which describes how well a model generalizes to unseen data.</p><ul><li><strong>Underfitting</strong>: High bias, low variance. Model is too simple and cannot capture patterns in the training data.</li><li><strong>Overfitting</strong>: Low bias, high variance. Model learns the training data too well but performs poorly on new data.</li><li><strong>Optimal</strong>: A balance between bias and variance, leading to good generalization.</li></ul><h3>Quantum Machine Learning (QML) Framework</h3><p>QML typically operates in the <strong>classical data, quantum processing (CQ) regime</strong>, meaning classical data is processed by a quantum computer. QML algorithms are divided into:</p><ul><li><strong>Fault-tolerant QML</strong>: Requires error correction, beyond current hardware (e.g., HHL algorithm, quantum PCA).</li><li><strong>Near-term QML</strong>: Designed for current noisy quantum computers with limited qubits and coherence times (e.g., QSVMs, QNNs).</li></ul><p>The workflow for a near-term quantum classifier includes:</p><ul><li><strong>Classical Data</strong>: The input features.</li><li><strong>Data Encoding (Feature Map)</strong>: Transforms classical data x into a quantum state. This is often represented as a unitary operation Ux​ that prepares a quantum state ∣ϕ(x)⟩ from an initial state, typically ∣0⟩⊗n.</li></ul><figure><img alt="" src="https://cdn-images-1.medium.com/max/328/1*piQMUGFz_vYpHyODx5XpWA.png" /></figure><ul><li><strong>Parameterized Quantum Circuit (Variational Circuit/Ansatz)</strong>: A quantum circuit W(θ) with tunable parameters θ. This circuit applies further unitary operations to the encoded data state.</li></ul><figure><img alt="" src="https://cdn-images-1.medium.com/max/428/1*3KblTSFUnQUXnaipR1OhHg.png" /></figure><ul><li><strong>Measurement</strong>: Measurements are performed on the output state to extract classical information. This can be a probability distribution over bit strings or the expectation value of an observable O^.</li></ul><figure><img alt="" src="https://cdn-images-1.medium.com/max/462/1*6JaoZ42YR4JBCbNFPhizZg.png" /></figure><ul><li><strong>Classical Post-processing and Optimization</strong>: The classical outputs are used to compute a loss function, and a classical optimizer updates the parameters θ of the quantum circuit to minimize this loss.</li></ul><h4>Data Encoding Strategies:</h4><ul><li><strong>Basis Encoding</strong>: Maps binary features directly to computational basis states. If a feature is ‘1’, the corresponding qubit is flipped with an X-gate. For an n-bit feature vector, n qubits are needed.</li><li><strong>Amplitude Encoding</strong>: Encodes data into the amplitudes of a quantum state. For a data vector with N features, it requires log2​N qubits.</li></ul><figure><img alt="" src="https://cdn-images-1.medium.com/max/278/1*MAqxuyb6wYSJpKcVHDH1kw.png" /></figure><ul><li><strong>Angle Encoding</strong>: Encodes features into the rotation angles of single-qubit gates. For a data vector x=(x1​,…,xN​), each xi​ might be mapped to a rotation angle θi​.</li></ul><figure><img alt="" src="https://cdn-images-1.medium.com/max/734/1*eQydxBOMx54Lnt5UermdeA.png" /></figure><ul><li>A variant, <strong>dense angle encoding</strong>, can encode two features per qubit by placing one in the rotation angle and another in the relative phase.</li><li><strong>Higher-Order Encoding (Feature Maps)</strong>: These encode not just the raw features but also higher-order functions (e.g., products) of the features into the quantum state, often through entangled circuits. An example is the <strong>ZZ feature map</strong>, which uses Hadamard gates, Z-rotations, and CNOT gates to encode features and their products. For a 2-feature vector (x1​,x2​), it might encode a term like x1​x2​ into a rotation angle.</li></ul><h3>Quantum Support Vector Machines (QSVMs)</h3><p>Classical <strong>Support Vector Machines (SVMs)</strong> are supervised learning models used for classification (and regression). They work by finding a <strong>hyperplane</strong> that optimally separates data points into different classes, maximizing the “margin” (distance) to the nearest data points, called <strong>support vectors</strong>.</p><p>For data that is not linearly separable in its original feature space, SVMs use the <strong>kernel trick</strong>. This involves mapping the data into a higher-dimensional feature space where it <em>might</em> become linearly separable. The “trick” is that the dot product (inner product) in this high-dimensional space can often be computed efficiently using a <strong>kernel function</strong> K(xi​,xj​)=ϕ(xi​)⋅ϕ(xj​), without explicitly calculating the high-dimensional feature vectors ϕ(x). The classification rule for an SVM in its dual formulation involves kernel evaluations:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/624/1*lYwf2k4hU6kSeRPy_Gup7g.png" /></figure><p>where αi​ are Lagrange multipliers learned during training, yi​ are the true labels, and b is a bias term.</p><p><strong>Quantum Support Vector Machines (QSVMs)</strong> leverage quantum computers to compute these kernel functions. The core idea is that a quantum computer can implicitly map classical data into an extremely high-dimensional (potentially exponentially large) quantum feature space. The <strong>quantum kernel</strong> K(xi​,xj​) is then estimated by calculating the overlap (inner product) between the quantum states prepared from different data points:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/488/1*-uLNoo3vwla4pgj9f1xHJA.png" /></figure><p>Here, ∣ϕ(x)⟩ is the quantum state prepared by a feature map Ux​ from an initial state, so</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/536/1*a1ogZPhpuX2q94Eoy723EA.png" /></figure><p>This overlap can be estimated on a quantum computer by preparing the state:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/200/1*sx7Fz6j_OCxp45TMHT3Wmg.png" /></figure><p>and then measuring the probability of finding all qubits in the ∣0⟩⊗n state.</p><p>The potential advantage of QSVMs comes from the fact that certain quantum feature maps might be computationally hard to simulate classically, leading to a “quantum advantage” if the corresponding quantum kernel is difficult to estimate classically. After the quantum kernel matrix is computed, the rest of the SVM optimization (finding αi​ and b) is handled by classical algorithms.</p><h3>Quantum Neural Networks (QNNs) in Detail</h3><p>QNNs aim to mimic classical neural networks using quantum circuits. A key challenge is introducing non-linearity, which is fundamental to the expressive power of classical neural networks but not inherent in linear quantum unitary operations.</p><p>Sources of non-linearity in QNNs include:</p><ul><li><strong>Measurements</strong>: Performing intermediate measurements within the circuit (mid-circuit measurements) or at the end. The probabilistic nature of quantum measurement is inherently non-linear.</li><li><strong>Data Re-uploading</strong>: This strategy involves repeatedly encoding the input data into the quantum circuit multiple times, interspersed with parameterized quantum layers. This iterative re-uploading of data has been theoretically shown to enhance the expressive power of QNNs, even with a small number of qubits, by introducing non-linear transformations.</li></ul><h4>Quantum Convolutional Neural Networks (QCNNs)</h4><p>Inspired by classical Convolutional Neural Networks (CNNs), <strong>Quantum Convolutional Neural Networks (QCNNs)</strong> are designed for tasks like image recognition. Classical CNNs use alternating <strong>convolutional layers</strong> (which extract local features by sliding filters over input data) and <strong>pooling layers</strong> (which reduce dimensionality and introduce non-linearity).</p><p>QCNNs adapt these concepts to quantum circuits:</p><ul><li><strong>Quantum Convolutional Layers</strong>: Consist of repeating unitary blocks that act on neighboring qubits, mimicking the local feature extraction of classical convolutions. These blocks typically contain parameterized rotation gates and entanglement gates.</li><li><strong>Quantum Pooling Layers</strong>: These layers reduce the number of qubits (dimensionality) and introduce non-linearity. Two main strategies are employed:</li><li><strong>Mid-circuit Measurement</strong>: Pairs of qubits are measured, and depending on the outcome of one qubit, a conditional operation might be applied to the other, or one qubit is discarded. This introduces non-linearity and reduces the effective number of active qubits.</li><li><strong>Tracing Out (Discarding) Qubits</strong>: One qubit from a pair is effectively removed from the computation. Mathematically, this corresponds to tracing out the subsystem, which is a non-linear operation.</li></ul><p>QCNNs are considered particularly promising for near-term quantum hardware due to:</p><ul><li><strong>Logarithmic Depth and Parameters</strong>: They often require a logarithmic number of layers and parameters with respect to the initial number of qubits, leading to shallower circuits which are less susceptible to noise.</li><li><strong>Mitigation of Barren Plateaus</strong>: QCNN architectures have been theoretically shown to be less prone to the <strong>Barren Plateau problem</strong>. This issue, where gradients vanish exponentially with the number of qubits in deep, randomly initialized circuits, makes training extremely difficult. The localized nature of QCNN operations and their ability to discard qubits contribute to a “more local” cost function, which helps avoid barren plateaus and makes them more trainable.</li></ul><p>Are we done? No not yet! Let us look into how Qiskit implements QML algorithms.</p><h3><strong>Quantum Machine Learning (QML) with QISKIT</strong></h3><p>QML with Qiskit integrates quantum computing principles into machine learning workflows, providing specialized circuits and algorithms for tasks such as classification and regression. Fully quantum machine learning algorithms are not yet practical due to hardware limitations, leading to hybrid quantum-classical approaches where quantum circuits are used for components that may offer speedups (such as state preparation and parameter optimization), while classical methods are used for parts better suited to traditional computation. QISKIT’s QML framework supports hybrid quantum-classical models, where only select tasks (e.g., state preparation, parameter optimization) are performed on quantum hardware, while other tasks are processed classically. Integration with frameworks such as PyTorch enables joint optimization of hybrid models using classical automatic differentiation routines.</p><h4>Key QML Components in QISKIT</h4><ul><li>Feature Maps &amp; Data Encoding: Classical data must be loaded and encoded into quantum states using feature maps. QISKIT provides several such maps, allowing dimensional and entanglement configuration. Feature maps transform classical feature vectors into quantum Hilbert space via parameterized unitary transformations, and different circuit architectures can be selected based on problem needs (linear, circular, or full entanglement).</li><li>Ansatz Circuits: An ansatz is a parameterized circuit representing possible quantum states, analogous to weights and biases in neural networks. Multiple types are available in the QISKIT library, with flexibility in gate choice, repetitions, and entanglement scheme.</li></ul><h4>QML Algorithms and Classes</h4><ul><li>Variational Quantum Classifier (VQC) &amp; Regressor (VQR): These classes automate the construction, training, and evaluation of hybrid machine learning models. The process involves iterative parameter updates (via classical optimization) to minimize loss and improve accuracy, with test data used to score model performance.</li><li>Quantum Support Vector Classifier (QSVC): Built on quantum kernels, these models leverage quantum feature maps for kernel trick, enabling classification of non-linear data. Kernel matrices are constructed from inner products of quantum states, displaying class clusters visually.</li></ul><h3>Workflow Example</h3><p>A typical workflow for quantum classification or regression involves:</p><ul><li>Generating or importing training data.</li><li>Encoding data using feature maps.</li><li>Building a quantum neural network with a chosen ansatz circuit.</li><li>Choosing an optimizer (classical or hybrid).</li><li>Training the network on quantum simulators or hardware, iteratively updating parameters to improve predictions.</li><li>Evaluating accuracy on test data.</li><li>Saving and reloading trained models for future use.</li></ul><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*i_WWGHq4pyXJ1JTG8G2Wkg.png" /></figure><h3>Further Exploration</h3><p>QISKIT Machine Learning provides tutorials and documentation on advanced features such as quantum generative adversarial networks, multi-dimensional data visualization, and effective integration with classical ML datasets and frameworks</p><blockquote><strong><em>Tags</em></strong></blockquote><p>#QuantumComputing #QuantumLearning #Probability #Statistics #BayesTheorem #Qohort3 #STEM #QuCode #QuantumIndia #NationalQuantumMission #DeepTech #FutureOfComputing #QuantumStartups #ResearchAndInnovation #LearnQuantum #QuantumAlgorithms #QuantumCommunity #NextGenTech #DigitalIndia #TechForGood</p><blockquote><strong><em>Resources/References</em></strong></blockquote><p>This article was written by understanding content from:</p><ul><li><a href="https://youtu.be/xL383DseSpE?si=lyGAyK6Ykj7gzMxP">https://youtu.be/xL383DseSpE?si=lyGAyK6Ykj7gzMxP</a></li><li><a href="https://youtu.be/tmpPqGUk-nI?si=24-ShyYP_lQKu-pW">https://youtu.be/tmpPqGUk-nI?si=24-ShyYP_lQKu-pW</a></li><li><a href="https://www.youtube.com/watch?v=ChPLYpUzsY4">https://www.youtube.com/watch?v=ChPLYpUzsY4</a></li><li><a href="https://www.youtube.com/watch?v=NiI0AJ6qPZ0">https://www.youtube.com/watch?v=NiI0AJ6qPZ0</a></li></ul><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=7516c525b029" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Variational Quantum Algorithms]]></title>
            <link>https://medium.com/@Qduality/variational-quantum-algorithms-bdc5a7d86fd7?source=rss-2bf9c8b9f1ed------2</link>
            <guid isPermaLink="false">https://medium.com/p/bdc5a7d86fd7</guid>
            <dc:creator><![CDATA[QDuality]]></dc:creator>
            <pubDate>Thu, 18 Sep 2025 18:23:24 GMT</pubDate>
            <atom:updated>2026-09-13T04:58:49.652Z</atom:updated>
            <content:encoded><![CDATA[<h3>Qucode: Day 18/21; QCC; Cohort-3.</h3><h3>Variational Quantum Algorithms (VQAs)</h3><p>Variational Quantum Algorithms are a class of <strong>hybrid quantum-classical algorithms</strong> designed to work with today’s noisy, intermediate-scale quantum computers. The core idea is to use a quantum computer to perform a part of the computation that a classical computer can’t do efficiently, and a classical computer to optimize the parameters of the quantum circuit. This creates a feedback loop: the quantum computer runs a circuit and gives a result, and the classical computer adjusts the circuit’s parameters to improve the result. This process is repeated until an approximate solution is found. VQAs are seen as a bridge to future, more advanced quantum computers and can be used for problems in areas like <strong>quantum chemistry</strong> and <strong>combinatorial optimization</strong>. While they offer potential speedups, they are still limited by the noise and size of current quantum hardware.</p><h3>Variational Quantum Eigensolver (VQE)</h3><p>The <strong>Variational Quantum Eigensolver (VQE)</strong> is a specific type of VQA that is used to find the minimum energy, or <strong>ground state energy</strong>, of a quantum system, such as a molecule. This is a crucial problem in chemistry and physics. VQE is particularly well-suited for current quantum computers because it can use <strong>shallow quantum circuits</strong>, which are less susceptible to noise and errors. The algorithm works by preparing a quantum state that approximates the ground state and then using a classical optimizer to minimize its energy. This involves mapping the complex physical properties of a molecule (like the positions and energies of its electrons and nuclei) onto a representation that can be processed by a quantum computer. The algorithm uses approximations and specialized mappings to simplify the problem, allowing it to be solved on existing quantum hardware.</p><blockquote><strong><em>Tags</em></strong></blockquote><p>#QuantumComputing #QuantumLearning #Probability #Statistics #BayesTheorem #Qohort3 #STEM #QuCode #QuantumIndia #NationalQuantumMission #DeepTech #FutureOfComputing #QuantumStartups #ResearchAndInnovation #LearnQuantum #QuantumAlgorithms #QuantumCommunity #NextGenTech #DigitalIndia #TechForGood</p><blockquote><strong><em>Resources/References</em></strong></blockquote><p>This article was written by understanding content from:</p><ul><li><a href="https://youtu.be/qc4zqVREb1g?si=YADMP0CAiut9oUhd">https://youtu.be/qc4zqVREb1g?si=YADMP0CAiut9oUhd</a></li><li><a href="https://youtu.be/AhEnvYgoA34?si=_-28JOmcxHGB6KKz">https://youtu.be/AhEnvYgoA34?si=_-28JOmcxHGB6KKz</a></li><li><a href="https://youtu.be/TUFovZsBcW4?si=beET4KbrHwjOx0f1">https://youtu.be/TUFovZsBcW4?si=beET4KbrHwjOx0f1</a></li></ul><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=bdc5a7d86fd7" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Grover’s Algorithm & Search Problems]]></title>
            <link>https://medium.com/@Qduality/grovers-algorithm-search-problems-0c868c7b0bc7?source=rss-2bf9c8b9f1ed------2</link>
            <guid isPermaLink="false">https://medium.com/p/0c868c7b0bc7</guid>
            <dc:creator><![CDATA[QDuality]]></dc:creator>
            <pubDate>Wed, 17 Sep 2025 18:23:50 GMT</pubDate>
            <atom:updated>2026-09-13T04:59:21.244Z</atom:updated>
            <content:encoded><![CDATA[<h3>Qucode: Day 17/21; QCC; Cohort-3.</h3><p>Welcome to Day 17 of the QCC!</p><p>Both <strong>Grover’s algorithm</strong> and <strong>quantum walks</strong> are quantum computing techniques designed to solve search problems more efficiently than classical computers.</p><h3>Grover’s Algorithm</h3><p>Grover’s algorithm is a powerful method for <strong>unstructured search</strong>. Imagine you have a huge, unsorted phone book and you want to find a person’s name by their phone number. A classical computer would have to check, on average, half of the entries, and in the worst case, every single one. Grover’s algorithm, however, can find the correct entry significantly faster, with a <strong>quadratic speedup</strong>. It works by using the principles of quantum mechanics. It starts with all possible entries in a state of <strong>superposition</strong>, meaning they all exist at once. The algorithm then repeatedly applies a clever two-step process:</p><ol><li>It subtly changes the quantum state of the correct entry to make it stand out.</li><li>It then amplifies this change, increasing the probability of measuring the correct entry while decreasing the probability of all the incorrect ones.</li></ol><p>By repeating these steps a specific number of times, the quantum state rotates until it is heavily weighted towards the right answer, making it highly likely you will find it on the first try.</p><h3>Quantum Walks</h3><p>A <strong>quantum walk</strong> is the quantum equivalent of a classical <strong>random walk</strong>. In a classical random walk, like a drunk person stumbling around a city, you move from one location to another in a random fashion. This can be used to solve a search problem, for example, finding a specific building by wandering around randomly. The time it takes is called the <strong>hitting time</strong>.</p><p>A quantum walk, on the other hand, can explore a graph or a network of locations much more effectively due to the ability of a quantum system to be in multiple places at once (<strong>superposition</strong>). This allows it to find a marked or desired item in a graph with a <strong>quadratic speedup</strong>, in a time proportional to the square root of the classical hitting time. This technique is particularly useful for problems that can be represented as a network, and it has been generalized to find items in any type of network, regardless of the number of marked items. This makes quantum walks a versatile tool for quantum software development.</p><blockquote><strong><em>Tags</em></strong></blockquote><p>#QuantumComputing #QuantumLearning #Probability #Statistics #BayesTheorem #Qohort3 #STEM #QuCode #QuantumIndia #NationalQuantumMission #DeepTech #FutureOfComputing #QuantumStartups #ResearchAndInnovation #LearnQuantum #QuantumAlgorithms #QuantumCommunity #NextGenTech #DigitalIndia #TechForGood</p><blockquote><strong><em>Resources/References</em></strong></blockquote><p>This article was written by understanding content from:</p><ul><li><a href="https://youtu.be/Ezc0cjsBG-M?si=6koBLWNtl0DYwreL">https://youtu.be/Ezc0cjsBG-M?si=6koBLWNtl0DYwreL</a></li><li><a href="https://youtu.be/c30KrWjHaw4?si=-N5krOGMnKdrTN61">https://youtu.be/c30KrWjHaw4?si=-N5krOGMnKdrTN61</a></li><li><a href="https://youtu.be/hnpjC8WQVrQ?si=PH1lOjx-wMAO8Obo">https://youtu.be/hnpjC8WQVrQ?si=PH1lOjx-wMAO8Obo</a></li></ul><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=0c868c7b0bc7" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Shor’s Algorithm & Quantum Cryptography]]></title>
            <link>https://medium.com/@Qduality/shors-algorithm-quantum-cryptography-9eca1605a362?source=rss-2bf9c8b9f1ed------2</link>
            <guid isPermaLink="false">https://medium.com/p/9eca1605a362</guid>
            <dc:creator><![CDATA[QDuality]]></dc:creator>
            <pubDate>Tue, 16 Sep 2025 15:26:03 GMT</pubDate>
            <atom:updated>2026-09-13T04:59:51.149Z</atom:updated>
            <content:encoded><![CDATA[<h3>Qucode: Day 16/21; QCC; Cohort-3.</h3><p>Welcome to Day 16!</p><p>We have only briefly looked at Shor’s algorithm in the previous article. We’ll continue from where we left off and explain how Shor’s algorithm really works by using Qiskit, and in parallel we’ll also look at Quantum Cryptography and how it could enable or disrupt security and encryption.</p><p>Let’s get started.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/848/1*VLzzvr-egHa7UGXxfpsQ-Q.png" /></figure><p>Peter Shor is a theoretical computer scientist who came up with his famous Shor’s algorithm in 1994. His algorithm shows the power of quantum computation compared to classical computation today. This is because it provides exponential speed up of finding large prime factors of large numbers, but only theoretically as of today. Why is prime factoring important?</p><p>Finding prime factors for large numbers is the basis for <strong>RSA encryption</strong>, which is a public key sharing scheme used for secure data transactions like credit card transactions.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/678/1*6T0sSkOFOu0TS4FzaAe6WQ.png" /></figure><p>Under RSA encryption, messages or data can be encrypted with a code called the <strong>public key: which is a large number </strong>that is shared openly. Once the message is encrypted and shared along with the public key, it can only be decrypted using the <strong>private key: which is a prime factor of that large number.</strong> The larger the number the more computational resources one requires to decrypt the data, and the security of RSA lies in the fact that prime factoring a large number takes a long period of time and is resource intensive. Consider a 232 digit decimal semi prime number. A semi prime number is the product of two prime numbers and is the hardest to factorize. It takes 1000’s of years to factorize this number on a laptop. Any classical or quantum computer does not yet have the capability to break RSA encryption so we don’t have to worry about our financial transactions <strong><em>yet, </em></strong>but let’s take that with a grain of salt.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/600/1*V6KKmvsGzE9tchLwslPuJA.jpeg" /></figure><p>Shor’s algorithm is a quantum algorithm that can break RSA encryption theoretically. Let us look into how to implement it using Qiskit.</p><h3><strong>Implementing Shor’s algorithm in Qiskit</strong></h3><p>As we already know, Shor’s algorithm leverages quantum principles, specifically the Quantum Fourier Transform (QFT), to find the period of a function, which in turn helps find the prime factors of a composite number. The implementation involves both a quantum circuit and classical post-processing.</p><blockquote>Source Code taken from Gemini 2.5 Flash Model</blockquote><pre>import matplotlib.pyplot as plt<br>import numpy as np<br>from qiskit import QuantumCircuit, Aer, transpile, assemble<br>from qiskit.visualization import plot_histogram<br>from math import gcd<br>from numpy.random import randint<br>import pandas as pd<br>from fractions import Fraction<br><br># --- 1. The Classical Part: Choose a number &#39;a&#39; ---<br>N = 15<br>a = 7 # A random number between 1 and N-1, coprime to N<br><br># --- 2. The Quantum Part: Period Finding ---<br><br>def qpe_amod15(a):<br>    &quot;&quot;&quot;<br>    Quantum circuit for Quantum Phase Estimation (QPE)<br>    to find the period of a^x mod 15.<br>    &quot;&quot;&quot;<br>    n_count = 8  # Number of counting qubits<br>    qc = QuantumCircuit(n_count + 4, n_count)<br><br>    # Initialize counting qubits in superposition<br>    for q in range(n_count):<br>        qc.h(q)<br><br>    # Initialize target qubits to |1&gt;<br>    qc.x(n_count + 3)<br><br>    # Controlled-U operations<br>    for q in range(n_count):<br>        qc.append(c_amod15(a, 2**q),<br>                  [q] + [i+n_count for i in range(4)])<br><br>    # Inverse Quantum Fourier Transform<br>    qc.append(qft_dagger(n_count), range(n_count))<br><br>    # Measure counting qubits<br>    qc.measure(range(n_count), range(n_count))<br>    return qc<br><br>def c_amod15(a, power):<br>    &quot;&quot;&quot;Controlled multiplication by a^power mod 15&quot;&quot;&quot;<br>    if a not in [2, 7, 8, 11, 13]:<br>        raise ValueError(&quot;&#39;a&#39; must be 2, 7, 8, 11, or 13&quot;)<br>    U = QuantumCircuit(4)<br>    for _iteration in range(power):<br>        if a in [2, 13]:<br>            U.swap(0, 1)<br>            U.swap(1, 2)<br>            U.swap(2, 3)<br>        if a in [7, 8]:<br>            U.swap(2, 3)<br>            U.swap(1, 2)<br>            U.swap(0, 1)<br>        if a == 11:<br>            U.swap(1, 3)<br>            U.swap(0, 2)<br>        if a in [7, 11, 13]:<br>            for q in range(4):<br>                U.x(q)<br>    U = U.to_gate()<br>    U.name = f&quot;{a}^{power} mod 15&quot;<br>    c_U = U.control()<br>    return c_U<br><br>def qft_dagger(n):<br>    &quot;&quot;&quot;n-qubit QFT†&quot;&quot;&quot;<br>    qc = QuantumCircuit(n)<br>    for qubit in range(n//2):<br>        qc.swap(qubit, n-qubit-1)<br>    for j in range(n):<br>        for m in range(j):<br>            qc.cp(-np.pi/float(2**(j-m)), m, j)<br>        qc.h(j)<br>    qc.name = &quot;QFT†&quot;<br>    return qc<br><br># --- 3. Simulate and Get Results ---<br>qc = qpe_amod15(a)<br>backend = Aer.get_backend(&#39;qasm_simulator&#39;)<br>t_qc = transpile(qc, backend)<br>qobj = assemble(t_qc, shots=1)<br>result = backend.run(qobj, memory=True).result()<br>readings = result.get_memory()<br>print(f&quot;Register reading: {readings[0]}&quot;)<br><br>phase = int(readings[0], 2) / (2**8)<br>print(f&quot;Corresponding phase: {phase}&quot;)<br><br># --- 4. Classical Post-Processing: Find Factors ---<br># Use Continued Fractions to find the period &#39;r&#39;<br>frac = Fraction(phase).limit_denominator(N)<br>s, r = frac.numerator, frac.denominator<br>print(f&quot;Fraction: {s}/{r}&quot;)<br><br>guesses = [gcd(a**(r//2)-1, N), gcd(a**(r//2)+1, N)]<br>print(f&quot;Guessed factors: {guesses}&quot;)<br><br># --- Verification ---<br>if guesses[0] != 1 and guesses[0] != N and (N % guesses[0] == 0):<br>    print(f&quot;Success! Factors are {guesses[0]} and {N // guesses[0]}.&quot;)<br>else:<br>    print(f&quot;Factor found is trivial: {guesses[0]}. Trying another &#39;a&#39;.&quot;)<br><br>if guesses[1] != 1 and guesses[1] != N and (N % guesses[1] == 0):<br>     print(f&quot;Success! Factors are {guesses[1]} and {N // guesses[1]}.&quot;)<br>else:<br>     print(f&quot;Factor found is trivial: {guesses[1]}. Trying another &#39;a&#39;.&quot;)</pre><h3>How the Code Works</h3><p>The above python implementation breaks down Shor’s algorithm into four main parts.</p><h3>1. Picking a number</h3><ul><li><strong>N = 15</strong>: This is the number we want to factor since quantum computers today can’t factorize large numbers.</li><li><strong>a = 7</strong>: We pick a random number a between 1 and N-1. A key condition is that a must be coprime to N (their greatest common divisor must be 1). For N=15, a=7 is a valid choice.</li></ul><h3>2. Period-Finding</h3><p>The goal of the quantum circuit is to find the <strong>period </strong><strong>r</strong> of the function f(x)=ax(modN). This r is the smallest positive integer such that ar≡1(modN).</p><ul><li><strong>qpe_amod15(a)</strong>: This function builds the main quantum circuit. It uses <strong>Quantum Phase Estimation (QPE)</strong> to find the phase, which is directly related to the period r.</li><li>It initializes a set of “counting” qubits into a superposition using Hadamard (H) gates.</li><li>It then performs a series of controlled-U operations. In our case, U is the operation that performs the modular multiplication (x→ax(mod15)). The controlled versions apply this operation a specific number of times.</li><li><strong>qft_dagger(n)</strong>: The <strong>Inverse Quantum Fourier Transform (QFT†)</strong> is applied to the counting qubits. This is the crucial step that transforms the quantum state so that when we measure it, we get a result that reveals the period.</li></ul><h3>3. Simulation and Measurement</h3><ul><li><strong>Aer.get_backend(&#39;qasm_simulator&#39;)</strong>: We use a classical simulator to run our quantum circuit.</li><li><strong>backend.run(...)</strong>: We execute the circuit. Since the algorithm is probabilistic, even a single &quot;shot&quot; can yield the correct answer, but multiple runs might be needed in a real scenario.</li><li><strong>result.get_memory()</strong>: This retrieves the measurement outcome from the counting qubits as a classical binary string (e.g., &#39;10000000&#39;).</li></ul><h3>4. Classical Post-Processing</h3><p>The quantum part gives us a phase, but not the factors directly. We need some classical math to finish the job.</p><ul><li><strong>Phase Calculation</strong>: The binary string from the measurement is converted into a decimal fraction. This phase is approximately equal to some integer s divided by the period r (s/r).</li><li><strong>Continued Fractions Algorithm</strong>: This classical algorithm is a very effective way to find the fraction s/r from its decimal approximation. The Fraction(phase).limit_denominator(N) line does this efficiently. The denominator of the resulting fraction is our best guess for the period r.</li><li><strong>Finding Factors</strong>: Once we have the period r, we can calculate the factors using the formula:</li></ul><blockquote>p,q=gcd(ar/2±1,N)</blockquote><ul><li>This last step gives us the prime factors of N. For our example, with N=15 and a=7, the period r is 4. The factors are then calculated as gcd(74/2−1,15)=gcd(48,15)=3 and gcd(74/2+1,15)=gcd(50,15)=5.</li></ul><p><strong>Future Risk</strong>: “Store now, decrypt later” attack is a major concern. A malicious actor could intercept and save encrypted data today. In the future, perhaps in the 2050s or 2060s, they could use a powerful quantum computer to break the now-outdated RSA encryption and access the stored information. To tackle this issue, new methods of <strong>post-quantum encryption</strong> are already being developed and deployed.</p><h3>Quantum Cryptography</h3><p>Quantum cryptography is a field of cybersecurity that uses the principles of quantum mechanics to create and distribute cryptographic keys. Unlike traditional cryptography, which relies on complex mathematical problems that are difficult for classical computers to solve, quantum cryptography’s security is based on the fundamental laws of physics.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/480/1*qCyw0vY1hVKayQLmpW9gRw.gif" /><figcaption>sourced from Giphy</figcaption></figure><p>The core of quantum cryptography, and its most widely known application, is <strong>Quantum Key Distribution (QKD)</strong>. QKD allows two parties, traditionally named Alice and Bob, to create a shared, secret key in a way that is provably secure against any eavesdropper, whom we’ll call Eve. The security comes from three key quantum principles:</p><ul><li><strong>Superposition:</strong> A quantum particle, such as a photon, can exist in multiple states at once (e.g., both 0 and 1).Information is encoded onto these qubits (quantum bits).</li><li><strong>No-Cloning Theorem:</strong> It’s physically impossible for Eve to make a perfect copy of an unknown quantum state.</li><li><strong>Measurement Disturbance:</strong> The act of measuring a quantum particle fundamentally changes its state.</li></ul><p>Using these principles, Alice sends a stream of photons, each with a specific quantum state, to Bob. If Eve attempts to intercept or measure these photons, her actions will inevitably alter their states. Alice and Bob can then compare a small portion of their measurements over a public channel to check for these disturbances. If they detect a high error rate, they know their communication has been compromised and they can discard the key. If the error rate is low, they can proceed to use the remaining part of the key to securely encrypt and decrypt their data using a traditional encryption algorithm.</p><h3>Quantum Cryptography vs. Post-Quantum Cryptography</h3><p>It’s important to distinguish between <strong>quantum cryptography</strong> and <strong>post-quantum cryptography (PQC)</strong>.</p><ul><li><strong>Quantum cryptography</strong> uses quantum mechanics to create a secure key distribution system. It is considered “future-proof” because its security isn’t based on the computational difficulty of a problem, but on the laws of physics. However, QKD requires a dedicated infrastructure, such as fiber optic cables, and is limited by distance. Therefore you need to use quantum network topolgies, repeaters, entanglement swapping and other techniques to build a robust communication channel from point A to point B.</li><li><strong>Post-quantum cryptography </strong>on the other hand is NOT a quantum method<strong>. </strong>It is a set of new, classical cryptographic algorithms that are designed to be secure against attacks from both classical and future quantum computers. These algorithms are based on different mathematical problems than those currently used, and are intended to be deployed on existing computer networks.</li></ul><p>While PQC is a practical solution for securing current communications against the threat of future quantum computers, quantum cryptography offers a theoretically unhackable method for key exchange. Both are crucial for the future of cybersecurity.</p><p>New post-quantum cryptography (PQC) standards were released by the National Institute of Standards and Technology (NIST). NIST’s call for new encryption methods led to the selection of four quantum-safe algorithms, all of which were co-developed by IBM scientists. The algorithms include Crystal-Kyber, Crystal-Dilithium, Falcon, and Sphinx+. These new standards puts everyone in the right track for global cybersecurity and provides guidance to organizations on how to begin the transition to PQC now to avoid future issues.</p><blockquote><strong><em>Tags</em></strong></blockquote><p>#QuantumComputing #QuantumLearning #Probability #Statistics #BayesTheorem #Qohort3 #STEM #QuCode #QuantumIndia #NationalQuantumMission #DeepTech #FutureOfComputing #QuantumStartups #ResearchAndInnovation #LearnQuantum #QuantumAlgorithms #QuantumCommunity #NextGenTech #DigitalIndia #TechForGood</p><blockquote><strong><em>Resources/References</em></strong></blockquote><p>This article was written by understanding content from:</p><ul><li><a href="https://youtu.be/lvTqbM5Dq4Q?si=6x4CFUSPyjG_WQON">https://youtu.be/lvTqbM5Dq4Q?si=6x4CFUSPyjG_WQON</a></li><li><a href="https://youtu.be/EdJ7RoWcU48?si=O2MG4HPS1BJ4wFlM">https://youtu.be/EdJ7RoWcU48?si=O2MG4HPS1BJ4wFlM</a></li><li><a href="https://youtu.be/gSy0PCS4m6E?si=oEG7x1tnwc7yd7C6">https://youtu.be/gSy0PCS4m6E?si=oEG7x1tnwc7yd7C6</a></li><li><a href="https://www.youtube.com/watch?v=YJ00O4gBs0I&amp;ab_channel=IBMResearch">https://www.youtube.com/watch?v=YJ00O4gBs0I&amp;ab_channel=IBMResearch</a></li><li><a href="https://www.scribd.com/document/634097421/Untitled">https://www.scribd.com/document/634097421</a></li></ul><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=9eca1605a362" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Quantum Fourier Transform (Subroutine in Quantum Phase Estimation)]]></title>
            <link>https://medium.com/@Qduality/quantum-fourier-transform-subroutine-in-quantum-phase-estimation-9e3c2016015b?source=rss-2bf9c8b9f1ed------2</link>
            <guid isPermaLink="false">https://medium.com/p/9e3c2016015b</guid>
            <dc:creator><![CDATA[QDuality]]></dc:creator>
            <pubDate>Mon, 15 Sep 2025 14:22:04 GMT</pubDate>
            <atom:updated>2026-09-13T05:00:16.330Z</atom:updated>
            <content:encoded><![CDATA[<h3>Qucode: Day 15/21; QCC; Cohort-3.</h3><p>We are a week away from finishing our QCC!</p><p>What a journey it’s been and we have now positioned ourselves deep inside the quantum realm.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/997/1*ou_b_5UDDE6dl2pYYXv_2w.jpeg" /></figure><p>Today we will try and understand <strong>Quantum Phase Estimation (QPE) </strong>and the usage of <strong>Quantum Fourier Transform (QFT) </strong>in QPE, and Shor’s algorithm that uses <strong>QPE</strong>. In addition to the provided videos, I have also borrowed some material from PennyLane’s tutorials as well, so feel free to read on or access any of the materials directly from the references section.</p><p>The <strong>Quantum Fourier Transform (QFT)</strong> is a quantum algorithm that is the quantum equivalent of the classical Discrete Fourier Transform (DFT), transforming a quantum state into a superposition of other states by manipulating their probability amplitudes. It is used in the Quantum Phase Estimation algorithm.</p><h3>Quantum Phase Estimation</h3><p>Quantum Phase Estimation is a key quantum algorithm used to find the <strong>eigenvalue</strong> of a <strong>unitary operator</strong>, U, given one of its eigenvectors (or eigenstates if you need to call it that). All eigenvalues in quantum computing can be represented as a complex phase on the unit circle, in the form of e^iθ. The goal of the algorithm is to find the value of θ.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/236/1*dkLJsiN2aXkXdKcwFyTQjg.png" /></figure><p>where θ or (phi in this case) is the <em>phase</em> of the eigenvalue (remember, unitaries have eigenvalues with an absolute value of 1). The goal is to estimate hence the name <em>phase estimation</em>. Our challenge is to design a quantum algorithm to solve this problem.</p><p>The algorithm works by using a quantum circuit with two registers:</p><ol><li>A control register with m qubits, which will store the final phase approximation.</li><li>An eigenvector register with n qubits, which holds the eigenvector, ∣v⟩.</li></ol><h4>Steps in QPE algorithm</h4><ul><li>The circuit starts in the below state</li></ul><figure><img alt="" src="https://cdn-images-1.medium.com/max/148/1*flFCLU4YqLUbn3RuQsV4GA.png" /></figure><ul><li>A Hadamard gate is applied to each qubit in the control register, creating a superposition of all possible states.</li><li>A series of <strong>controlled-U operations</strong> are applied. The first control qubit controls a U gate, the second controls a U² gate, the third controls a U⁴ gate, and so on, up to U^2(m-1). These operations cause what is called a “<strong>phase kickback</strong>,” where the eigenvalue’s <strong>phase is transferred</strong> to the control qubits.</li></ul><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*nKF-Db2Ha4CPdACSxXq-3g.png" /></figure><ul><li>After these operations, the state of the control register is:</li></ul><figure><img alt="" src="https://cdn-images-1.medium.com/max/346/1*eZ4Q_g0ms7xyOhLE-WUlrQ.png" /></figure><ul><li>An <strong>Inverse Quantum Fourier Transform (represented as QFT dagger), </strong>is then applied to the control register. This transformation converts the state from a sum of complex exponentials into a state where the basis vectors correspond to the binary representation of the phase.</li><li>Finally, the control register is measured. The measured result, an integer J, provides an approximation of the phase. The eigenvalue is then calculated from the relationship:</li><li>θ=2πJ</li></ul><h3>Shor’s Algorithm</h3><blockquote>John Watrous from IBM has one of the best in-depth technical explanation of QPE and Shor’s algorithm which is available in the references section, so please do watch if you can set aside an hour.</blockquote><p><strong>Shor’s algorithm</strong> is a quantum algorithm that efficiently finds the prime factors of a large integer, N, by reducing the problem to <strong>order finding</strong>. The order of an integer a modulo N is the smallest positive integer, r, such that:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/302/1*1gkPg-riNYPVa57d4pw2uA.png" /></figure><p>The order can be found using the Phase Estimation algorithm. The unitary operator, U, in this case is defined to be multiplication by a modulo N. The eigenvectors of this operator are a superposition of states related to the powers of a:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/662/1*Zsqe1Kx9oVak2T--L5uh1w.png" /></figure><p>The corresponding eigenvalues are e^2πij/r, where j is an integer. By performing phase estimation on these states, the algorithm can find an approximation of a random fraction, j/r. A classical continued fraction algorithm is then used to efficiently recover the order, r.</p><p>The key to Shor’s algorithm is that while the quantum part solves the order-finding problem efficiently, the classical part then uses the order to find the prime factors of the original number. The algorithm’s efficiency comes from its ability to use quantum parallelism to find the order in polynomial time, unlike classical algorithms which take exponential time.</p><h3>How many qubits are needed for Shor’s algorithm?</h3><blockquote>This below material is directly sourced from <a href="https://www.classiq.io/insights/shors-algorithm-explained"><strong>ClassiQ’s</strong></a> blog explaining the Shor’s algorithm, and I stitched it together with the above section for the sake of the audience understanding how many qubits are really needed for Shor’s, and where we are practically with our quantum computers to have the quantum advantage.</blockquote><p>A distinction first needs to be made between physical and logical qubits to answer this question. All present-day qubits are physical qubits. They are extremely “noisy,” which means they are error-prone. The results of quantum computation of any significance make it impossible to distinguish correct answers from incorrect ones. Every possible solution has the same probability of being right, which is the same as having no answers at all.</p><p>That’s where logical qubits come into play. Physical qubits need to be connected and structured in ways that they will collectively provide enough error correction to each other to be considered “fault-tolerant” collectively. At that point, these qubit collectives become known as “logical qubits,” or sometimes “perfect qubits.”</p><p>These logical qubits are abstractions. A quantum circuit today with five qubits represents five highly-noisy, error-prone physical qubits. Quantum algorithm designers of the near future will want those five qubits to represent logical, fault-tolerant, error-free qubits. The estimates vary as to how many physical qubits will be needed to represent one logical qubit, but a reasonable number to work with is 1,000. Most estimate that a quantum computer will need around 1,000 physical qubits to represent just one logical qubit.</p><p>To appreciate how challenging that is, the largest quantum computer today has only 127 physical qubits. And while IBM has a goal of unveiling a 1,000-qubit device by the end of this decade, that is still only 1,000 physical qubits. Much work remains to engineer the first logical qubit.</p><p>Estimates of the number of qubits Shor’s Factoring Algorithm needs vary considerably. First, care must be taken, as noted, to discern estimates for logical qubits and estimates for physical qubits. Depending on the researchers’ objectives, the number of required qubits can be reduced, sacrificing time of execution and circuit depth. On the other hand, the number of qubits can be significantly expanded, reducing runtime and shedding circuit depth. The differences are that lower qubit counts will become available sooner, while the quickest runtimes will be the most advantageous. A selection of estimates follows.</p><p>Most recently, in an April 13, 2021 paper titled, “<a href="https://arxiv.org/pdf/1905.09749.pdf">How to factor 2048 bit RSA integers in 8 hours using 20 million noisy qubits</a>” by Craig Gidney and Martin Eker, the authors estimate that breaking RSA encryption, specifically, would require roughly 20,000 logical qubits, or approximately 20 million physical qubits. The authors take one step further and quantify the runtime for their configuration as only eight hours. Compare that to the estimate for the world’s most powerful supercomputers to break RSA encryption, which is well beyond any human lifetime, and the power of Shor’s Factoring Algorithm becomes clear. For other estimates of how many qubits for Shor’s algorithm, check out the references for this paper.</p><p>Again, Shor’s algorithm runtime is inversely proportional to the number of logical qubits available. It also depends on the qubit technology in use, since different quantum computers execute operations at vastly different speeds. The fewer qubits needed, the longer the algorithm will take to run, but the sooner modern cryptosystems transition from obsolescent to obsolete. Conversely, the more qubits available, the faster modern cryptosystems will be cracked. But fortunately, those days are even further away.</p><blockquote><strong><em>Tags</em></strong></blockquote><p>#QuantumComputing #QuantumLearning #Probability #Statistics #BayesTheorem #Qohort3 #STEM #QuCode #QuantumIndia #NationalQuantumMission #DeepTech #FutureOfComputing #QuantumStartups #ResearchAndInnovation #LearnQuantum #QuantumAlgorithms #QuantumCommunity #NextGenTech #DigitalIndia #TechForGood</p><blockquote><strong><em>Resources/References</em></strong></blockquote><p>This article was written by understanding content from:</p><ul><li><a href="https://pennylane.ai/qml/demos/tutorial_qpe">https://pennylane.ai/qml/demos/tutorial_qpe</a></li><li><a href="https://www.classiq.io/insights/shors-algorithm-explained">https://www.classiq.io/insights/shors-algorithm-explained</a></li><li><a href="https://youtu.be/Ex96GyRIFes?si=PUtzLNjt6oakkPcN">https://youtu.be/Ex96GyRIFes?si=PUtzLNjt6oakkPcN</a></li><li><a href="https://youtu.be/4nT0BTUxhJY?si=SBaa5Qcoj9k_Qe1U">https://youtu.be/4nT0BTUxhJY?si=SBaa5Qcoj9k_Qe1U</a></li></ul><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=9e3c2016015b" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Quantum Programming Basics]]></title>
            <link>https://medium.com/@Qduality/quantum-programming-basics-ae3c72dffcb6?source=rss-2bf9c8b9f1ed------2</link>
            <guid isPermaLink="false">https://medium.com/p/ae3c72dffcb6</guid>
            <dc:creator><![CDATA[QDuality]]></dc:creator>
            <pubDate>Sun, 14 Sep 2025 18:17:16 GMT</pubDate>
            <atom:updated>2026-09-13T05:00:47.531Z</atom:updated>
            <content:encoded><![CDATA[<h3>Qucode: Day 14/21; QCC; Cohort-3.</h3><p>Welcome to Day 14 of the QCC, where we will introduce how to program and interface with quantum computers, primarily Qiskit.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/286/1*92fvvkKk4br7xUDRD6DHJQ.png" /></figure><h3>QISKIT 2.0</h3><p>Qiskit has been around for 7 years already.</p><p><strong>Jin-Sung Kim </strong>and <strong>Nick Bronn </strong>(Nick who I met in person at Quantum India Bangalore 2025)<strong> </strong>were part of the initial iteration of Qiskit training videos released on youtube and I started my quantum journey by watching <strong>Jin-Sung Kim</strong>’s videos in 2018.</p><p>Researchers can now use IBM’s Qiskit to try and implement their research ideas on a real quantum computer in this NISQ era. Though the utility quantum computers in this NISQ (Noisy Intermediate Scale Quantum) era have some noise or errors associated with them, there are mitigation strategies that make the existing quantum computers reliable and usable for a limited set of problems that can’t be experimented on traditional classical systems. Here is the evolution of Qiskit 1.0 where the circuit depth memory usage while using 127 qubits has reduced over time.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*2Ng4QiXfX177TFjWLa_c8w.png" /></figure><p>In June 2023, IBM published a paper to showcase the evidence for the utility of quantum computing before fault tolerance. In the <strong>Utility era,</strong> we are seeing a few high impact scientific papers that use all 127 qubits on IBM’s existing 127 qubit systems.</p><p>Proliferation of methods like circuit cutting, air mitigation that require integration of a hybrid mechanism which was not possible before is now a reality.</p><p>Below a 100 qubits, any task performed on a quantum computer can be simulated classically. Beyond that, you have to be able to understand how to deal with a quantum computer. Simulators are software programs that can <strong>MIMIC</strong> what quantum computers can do, including noise. But they are not Quantum computers and they cannot show the same results like a quantum computer.</p><blockquote><strong><em>QISKIT</em></strong></blockquote><p>I will attempt to show you how to use <strong>Qiskit 2.0</strong> and not 1.0 since 1.0 version is now deprecated.</p><p>To install <strong>Qiskit 2.0 </strong>and above, you’ll need <strong>Python 3.8 </strong>or<strong> </strong>later. The recommended way is to use Python’s package manager, pip. Open your terminal or command prompt and run the following command:</p><p><strong>Bash</strong></p><pre>pip install qiskit</pre><p>Once installed, you can write your first “Hello World” program. This program creates a simple quantum circuit, measures a qubit, and runs it on a basic simulator.</p><pre>from qiskit import QuantumCircuit<br>from qiskit.providers.basic_provider import BasicProvider<br><br># Create a quantum circuit with one qubit and one classical bit<br>qc = QuantumCircuit(1, 1)<br><br># Apply a Hadamard gate to create a superposition<br>qc.h(0)<br><br># Measure the qubit and store the result in the classical bit<br>qc.measure(0, 0)<br><br># Get the basic provider<br>provider = BasicProvider()<br>simulator = provider.get_backend(&#39;basic_simulator&#39;)<br><br># Run the job and get the results<br>job = simulator.run(qc, shots=1000)<br>result = job.result()<br>counts = result.get_counts(qc)<br><br># Print the measurement outcomes<br>print(&quot;Counts:&quot;, counts)</pre><p>The above code will output a dictionary showing the measurement results, likely close to a 50/50 split between {&#39;0&#39;: 500, &#39;1&#39;: 500} due to the superposition created by the Hadamard gate. In other words, it shows the probability distribution of obtaining one of the outputs.</p><h3>Qiskit Primitives</h3><p>In <strong>Qiskit 2.0</strong>, <strong>primitives</strong> are the new fundamental building blocks for building quantum applications. They are designed to be more efficient and easier to use than the old transpiler and execute flow. Primitives are a high-level API that allows you to specify what you want to compute, rather than how to compute it. They abstract away the complexities of running circuits on different backends.</p><p>There are two main types of primitives:</p><p><strong>Sampler</strong>: This primitive takes a quantum circuit as input and returns the quasi-probability distribution of its measurement outcomes. It’s used to sample from the output of a quantum circuit.</p><pre>from qiskit import QuantumCircuit<br>from qiskit.primitives import Sampler<br><br># Create a simple circuit<br>qc = QuantumCircuit(2)<br>qc.h(0)<br>qc.cx(0, 1) # Create an entangled state<br><br># Instantiate the Sampler primitive<br>sampler = Sampler()<br><br># Run the circuit and get the quasi-probabilities<br>job = sampler.run(qc)<br>result = job.result()<br>print(&quot;Quasi-probabilities:&quot;, result.quasi_dists[0])</pre><p>Above code will output a dictionary of quasi-probabilities for the measurement outcomes, such as {0: 0.5, 3: 0.5}, indicating a 50% chance of measuring 00 and a 50% chance of measuring 11.</p><p><strong>Estimator</strong>: This primitive is used to calculate the expectation values of quantum operators (observables). It takes a quantum circuit, an observable, and a set of parameter values as input and returns the expectation value.</p><pre>import numpy as np<br>from qiskit import QuantumCircuit<br>from qiskit.quantum_info import SparsePauliOp<br>from qiskit.primitives import Estimator<br><br># Create a parameterized circuit<br>qc = QuantumCircuit(1)<br>qc.rx(np.pi / 2, 0)<br><br># Define an observable to measure<br>observable = SparsePauliOp(&quot;X&quot;)<br><br># Instantiate the Estimator primitive<br>estimator = Estimator()<br><br># Run the job to get the expectation value<br>job = estimator.run(circuits=qc, observables=observable)<br>result = job.result()<br>print(&quot;Expectation value:&quot;, result.values[0])</pre><p>This code will output the expectation value of the X observable, which in this case will be ~1.0, as the rx gate puts the qubit in the |-&gt; state which is the eigenstate of the X-Pauli operator with an eigenvalue of +1.</p><p>Last but not the least, how do you contribute to Qiskit, which is esentially open source (for the most part at least)?</p><h3>Would you like to contribute to Qiskit?</h3><p>Qiskit is an open-source project, and contributions are welcome from everyone. The new <strong>Qiskit 2.0</strong> ecosystem is modular, with different repositories for different parts of the stack.</p><blockquote><strong>Find a task</strong>: Look for issues on the Qiskit GitHub repositories (e.g., qiskit-terra, qiskit-ibm-runtime, qiskit-algorithms). Good places to start are issues labeled good first issue or help wanted.</blockquote><blockquote><strong>Fork the repository</strong>: Create your own copy of the repository on GitHub.</blockquote><blockquote><strong>Clone the fork</strong>: Clone your forked repository to your local machine.</blockquote><blockquote><strong>Install dependencies</strong>: Install the necessary development dependencies. For most repositories, this is as simple as running pip install -e .[dev] within the repository folder.</blockquote><blockquote><strong>Make changes</strong>: Create a new branch, write your code, and commit your changes.</blockquote><blockquote><strong>Run tests</strong>: Make sure all tests pass before submitting your pull request. You can run tests using pytest or tox.</blockquote><blockquote><strong>Open a pull request</strong>: Push your changes to your fork and open a pull request against the main repository. Be sure to describe your changes clearly.</blockquote><blockquote><strong><em>Tags</em></strong></blockquote><p>#QuantumComputing #QuantumLearning #Probability #Statistics #BayesTheorem #Qohort3 #STEM #QuCode #QuantumIndia #NationalQuantumMission #DeepTech #FutureOfComputing #QuantumStartups #ResearchAndInnovation #LearnQuantum #QuantumAlgorithms #QuantumCommunity #NextGenTech #DigitalIndia #TechForGood</p><blockquote><strong><em>Resources/References</em></strong></blockquote><p>This article was written by understanding content from a series of Qiskit lectures on youtube.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=ae3c72dffcb6" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Quantum Computing Models]]></title>
            <link>https://medium.com/@Qduality/quantum-computing-models-773ea4c6eca1?source=rss-2bf9c8b9f1ed------2</link>
            <guid isPermaLink="false">https://medium.com/p/773ea4c6eca1</guid>
            <dc:creator><![CDATA[QDuality]]></dc:creator>
            <pubDate>Sat, 13 Sep 2025 16:41:06 GMT</pubDate>
            <atom:updated>2026-09-13T05:01:15.670Z</atom:updated>
            <content:encoded><![CDATA[<h3><strong>Qucode: Day 13/21; QCC; Cohort-3;</strong></h3><p>Welcome to Day 13.</p><p>Circuit based models are very common and we basically use the gates that we went over the last few articles to build our quantum circuits. So I won’t be diving too deeply into circuit based models but let us look at measurement based and adiabatic quantum computation models.</p><p>Let’s get started!</p><h3>Measurement based computation</h3><blockquote>Measurement-based computation is an alternative quantum computing paradigm that relies on a series of measurements instead of the more traditional gate-based quantum circuits. The core idea is that any gate operation can be replicated by preparing a specific two-qubit state and then performing two successive measurements.</blockquote><h3><strong>Key Concepts and Equations</strong></h3><p>The central problem addressed is how to perform an arbitrary unitary transformation U on a single qubit state |psi&gt; using only measurements. The method uses two qubits: a qubit of interest (initialized to |psi&gt;) and an ancilla qubit (initialized to | 0&gt;).</p><p><strong>First Measurement:</strong> A measurement of a two-qubit observable, B, is performed. The observable is defined by the following Hermitian operator, which has eigenvalues of +1 and -1:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/501/1*BPMHVN-ljEwGxrSub74cnQ@2x.jpeg" /></figure><p>Here, sigma_x and sigma_y are Pauli matrices, and U is the desired unitary transformation.</p><p><strong>Second Measurement: </strong>A second measurement is performed on the ancilla qubit.</p><ul><li>If the result is -1 (corresponding to the basis state | 1&gt;), the process is successful, and the qubit of interest is now in the desired state U|psi&gt;.</li><li>If the result is +1 (corresponding to | 0&gt;), the process has failed, and the circuit must be repeated. On average, this process will succeed after two attempts.</li></ul><p>This probabilistic, measurement-based approach is as universal and powerful as gate-based quantum computation.</p><h3>Adiabatic Quantum Computation</h3><blockquote><strong>Adiabatic</strong> quantum computation is an alternative to the more common circuit model. The central idea is to solve a problem by slowly evolving a quantum system from a known, simple ground state to a final ground state that encodes the solution to a complex problem.</blockquote><h3><strong>Key Concepts and Equations</strong></h3><ul><li>Adiabatic Process: The system starts in the ground state of a simple initial Hamiltonian, H_0, and is slowly transformed into a final Hamiltonian, H_f, whose ground state is the solution to the problem. This transformation is described by the equation:</li></ul><blockquote>H(t) = (1-t)*H_0 + t*H_f; where t is the time parameter that goes from 0 to 1.</blockquote><ul><li>Quantum Adiabatic Theorem: This theorem states that if the transformation from H_0 to H_f happens slowly enough, the system will remain in its ground state throughout the process.</li><li>Time Complexity: The total time, T, required for the computation is inversely related to the square of the minimum energy gap between the ground state and the first excited state, known as the spectral gap, G_{min}. The smaller the gap, the more slowly the system must evolve, making the computation take longer. The relationship is:</li></ul><figure><img alt="" src="https://cdn-images-1.medium.com/max/173/1*-3Xrw4CCmFa4UVZpLbr7RA@2x.jpeg" /></figure><p>where T is the computation time and G_{min} is the minimum spectral gap.</p><p><strong>Practical Challenges</strong></p><ul><li>While this model has potential, the spectral gap can become exponentially small for certain problems, requiring exponentially long computation times, thus eliminating any potential quantum speed-up.</li></ul><blockquote><strong><em>Tags</em></strong></blockquote><p>#QuantumComputing #QuantumLearning #Probability #Statistics #BayesTheorem #Qohort3 #STEM #QuCode #QuantumIndia #NationalQuantumMission #DeepTech #FutureOfComputing #QuantumStartups #ResearchAndInnovation #LearnQuantum #QuantumAlgorithms #QuantumCommunity #NextGenTech #DigitalIndia #TechForGood</p><blockquote><strong><em>Resources/References</em></strong></blockquote><p>This article was written by understanding content from:</p><iframe src="https://cdn.embedly.com/widgets/media.html?src=https%3A%2F%2Fwww.youtube.com%2Fembed%2F7tf3I8gYkOs%3Ffeature%3Doembed&amp;display_name=YouTube&amp;url=https%3A%2F%2Fwww.youtube.com%2Fwatch%3Fv%3D7tf3I8gYkOs&amp;image=https%3A%2F%2Fi.ytimg.com%2Fvi%2F7tf3I8gYkOs%2Fhqdefault.jpg&amp;type=text%2Fhtml&amp;schema=youtube" width="854" height="480" frameborder="0" scrolling="no"><a href="https://medium.com/media/26f16cccf38746a13b88d2f2b112b672/href">https://medium.com/media/26f16cccf38746a13b88d2f2b112b672/href</a></iframe><iframe src="https://cdn.embedly.com/widgets/media.html?src=https%3A%2F%2Fwww.youtube.com%2Fembed%2FvFpNNrt-baE%3Ffeature%3Doembed&amp;display_name=YouTube&amp;url=https%3A%2F%2Fwww.youtube.com%2Fwatch%3Fv%3DvFpNNrt-baE&amp;image=https%3A%2F%2Fi.ytimg.com%2Fvi%2FvFpNNrt-baE%2Fhqdefault.jpg&amp;type=text%2Fhtml&amp;schema=youtube" width="854" height="480" frameborder="0" scrolling="no"><a href="https://medium.com/media/8ef02ee1b9c5bb6ceea65e629e4bb24e/href">https://medium.com/media/8ef02ee1b9c5bb6ceea65e629e4bb24e/href</a></iframe><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=773ea4c6eca1" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Quantum Measurement & No-Cloning Theorem]]></title>
            <link>https://medium.com/@Qduality/quantum-measurement-no-cloning-theorem-6d29c3cd851f?source=rss-2bf9c8b9f1ed------2</link>
            <guid isPermaLink="false">https://medium.com/p/6d29c3cd851f</guid>
            <dc:creator><![CDATA[QDuality]]></dc:creator>
            <pubDate>Thu, 11 Sep 2025 18:41:57 GMT</pubDate>
            <atom:updated>2026-09-13T05:01:44.427Z</atom:updated>
            <content:encoded><![CDATA[<h3>QuCode: Day 12/21; QCC; Cohort-3.</h3><p>Hi and Welcome to Day 12, and we are already half way through our QCC! Today we will be looking into quantum measurement, specifically projective measurement, and will also understand what we mean by collapse of the wave function.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*gsRlWYVKlcfzAuQt7bSexg.png" /></figure><p>Let’s get started!</p><p><strong>Wave Function: What is it?</strong></p><p>We have seen Schrödinger’s wave function equation a few times in this series already, but let’s truly understand what a wave function actually is, what it does, and why it is required.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*W0BfLngE8Nmk4V_k8OJV4A.png" /></figure><p>Schrödinger’s equation helps determine the probability of finding a quantum object in a given state by treating it as a wave function.</p><h3>The No-Cloning Theorem</h3><p><strong>No-Cloning Theorem </strong>states that it is impossible to create a unitary operator that can perfectly clone an arbitrary unknown quantum state. This means there is no universal “copy machine” for qubits.</p><p>There are two proofs for this theorem, both of which rely on fundamental principles of quantum mechanics:</p><p><strong>Proof by Linearity:</strong> This proof demonstrates that a hypothetical cloning operator, Uc​, would not be a linear operator, which contradicts the fact that all quantum gates are linear. A linear operator acts on a superposition as a sum of its actions on the individual components.</p><p>If a cloning operator existed, applying it to a superposition ∣ψ⟩=α∣0⟩+β∣1⟩ would result in:</p><blockquote>Uc​(α∣0⟩+β∣1⟩)=α2∣00⟩+αβ∣01⟩+βα∣10⟩+β2∣11⟩</blockquote><ul><li>However, if you applied the operator to the individual components first and then summed them up, you would get a different result:</li></ul><blockquote>αUc​∣0⟩+βUc​∣1⟩=α∣00⟩+β∣11⟩</blockquote><ul><li>Since these two results are not equal, the operator is not linear and therefore cannot exist as a quantum gate.</li></ul><p><strong>Proof by Inner Product:</strong> This proof uses the property that unitary operators preserve the inner product of two state vectors. The inner product of two states before the operation must equal the inner product after.</p><ul><li>Let’s assume a cloning operator exists that can clone two different states, ∣ψ⟩ and ∣ϕ⟩.</li><li>The inner product of the two states before cloning is:</li></ul><blockquote>⟨ψ∣ϕ⟩</blockquote><ul><li>After cloning, the inner product of the resulting states is:</li></ul><blockquote>⟨ψψ∣ϕϕ⟩=⟨ψ∣ϕ⟩⟨ψ∣ϕ⟩=(⟨ψ∣ϕ⟩)2</blockquote><ul><li>Since a unitary operator must preserve this inner product, we have:</li></ul><blockquote>⟨ψ∣ϕ⟩=(⟨ψ∣ϕ⟩)2</blockquote><ul><li>This equation is only true if ⟨ψ∣ϕ⟩ is either 0 or 1. This means a cloning operator can only clone states that are either identical or completely orthogonal to each other. Since a cloning operator cannot clone an arbitrary state, the theorem is proven.</li></ul><h3>Measurement and Wave Function Collapse</h3><p>When a measurement is performed, two things happen: the measurement result is an <strong>eigenvalue</strong> of the measured quantity, and the wave function of the system “collapses” into the corresponding <strong>eigenstate</strong>.</p><ul><li><strong>Position Measurement:</strong> The position operator is denoted by a capital X. When it acts on a position eigenstate Ψx0​​, it returns the eigenvalue x0​ times the same eigenstate:</li></ul><blockquote>XΨx0​​=x0​Ψx0​​</blockquote><ul><li>The position eigenstate is a <strong>Dirac Delta function</strong>, which is an infinitely narrow and tall spike at the measured position, x0​.</li><li><strong>Momentum Measurement:</strong> The momentum operator is denoted by P. The momentum eigenstate Ψp0​​ is a complex exponential:</li></ul><blockquote>Ψp0​​=eℏip0​x​</blockquote><ul><li>The momentum operator is defined as:</li></ul><blockquote>P=iℏ​dxd​</blockquote><ul><li>When this operator acts on the momentum eigenstate, it returns the momentum eigenvalue p0​ times the eigenstate:</li></ul><blockquote>PΨp0​​=p0​Ψp0​​</blockquote><p><strong>Heisenberg Uncertainty Principle</strong> can be shown through a simulation. When position is measured, the position wave function becomes a sharp spike, meaning position is known with high certainty. However, the momentum wave function becomes flat, indicating high uncertainty in momentum. The opposite occurs when momentum is measured.</p><h3>Multibit Systems and Projective Measurements</h3><p>How do we represent a system with multiple qubits and how do we perform measurements on them?</p><ul><li>A system with n qubits has a state vector that exists in a 2n dimensional space. The basis states are represented by tensor products of single-qubit states. For example, a two-qubit system has four basis states: ∣00⟩,∣01⟩,∣10⟩,∣11⟩.</li><li>A <strong>projective measurement</strong> collapses the wave function by projecting the state onto a subspace consistent with the measurement outcome.</li><li>The probability of a measurement outcome is the squared magnitude of the amplitude for that state. For a two-qubit state ∣ψ⟩=α00​∣00⟩+α01​∣01⟩+α10​∣10⟩+α11​∣11⟩, the probability of measuring ∣00⟩ is ∣α00​∣2.</li><li>For a <strong>partial measurement</strong>, where only a subset of qubits is measured, the system’s state collapses to a superposition of only the states consistent with the measurement outcome. This resulting state is then re-normalized. For example, if you measure the first qubit and get ∣0⟩, the post-measurement state is a normalized superposition of the states ∣00⟩ and ∣01⟩:</li><li>∣α00​∣2+∣α01​∣2</li><li>​α00​∣00⟩+α01​∣01⟩​</li></ul><h3>Interpretations of Quantum Mechanics</h3><p>There are also philosophical interpretations of quantum mechanics, specifically addressing the concept of <strong>wave function collapse</strong>.</p><ul><li><strong>The Problem:</strong> The instantaneous collapse of a wave function seems to contradict special relativity if a single quantum object, described as a wave, is measured at two different, distant locations. Einstein famously called this “spooky action at a distance.”</li><li><strong>Copenhagen Interpretation:</strong> This is the most widely accepted view, stating that the wave function represents a probability and only collapses instantaneously upon measurement. Since the wave function itself is not a physical object, the instantaneous collapse is not considered a problem.</li><li><strong>Hidden Variables:</strong> This interpretation suggests that quantum mechanics is a deterministic theory with unknown variables that we have not yet discovered. However, experiments based on <strong>Bell’s Theorem</strong> have effectively ruled out any local hidden variables.</li><li><strong>Many-Worlds Interpretation:</strong> This view avoids collapse by suggesting that every possible outcome of a quantum measurement happens in a different, parallel universe.</li><li><strong>Continuous Spontaneous Localization:</strong> This interpretation proposes that the wave function continuously and spontaneously collapses due to an unknown source, potentially gravity, which interacts with the quantum system. This theory is testable, but experiments to date have not found the predicted signals, thus narrowing the possibility of this interpretation being correct.</li></ul><blockquote><strong><em>Tags</em></strong></blockquote><p>#QuantumComputing #QuantumLearning #Probability #Statistics #BayesTheorem #Qohort3 #STEM #QuCode #QuantumIndia #NationalQuantumMission #DeepTech #FutureOfComputing #QuantumStartups #ResearchAndInnovation #LearnQuantum #QuantumAlgorithms #QuantumCommunity #NextGenTech #DigitalIndia #TechForGood</p><blockquote><strong><em>Resources/References</em></strong></blockquote><p>This Article was written by understanding content from:</p><ul><li><a href="https://youtu.be/1pLynSfLucs?si=uKI6WQzZRvtiv9us">https://youtu.be/1pLynSfLucs?si=uKI6WQzZRvtiv9us</a></li><li><a href="https://youtu.be/VxoMYQIzA6k?si=1HqyEyUnncY8Aqc9">https://youtu.be/VxoMYQIzA6k?si=1HqyEyUnncY8Aqc9</a></li><li><a href="https://youtu.be/QmZJf3ipW7g?si=upqr78ZvO4AO1nZz">https://youtu.be/QmZJf3ipW7g?si=upqr78ZvO4AO1nZz</a></li><li><a href="https://www.youtube.com/watch?v=qtjdyAziXr8">https://www.youtube.com/watch?v=qtjdyAziXr8</a></li></ul><p><a href="https://medium.com/u/334b8f0a8fb1">Qucode</a></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=6d29c3cd851f" width="1" height="1" alt="">]]></content:encoded>
        </item>
    </channel>
</rss>