<?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 Dr. Stylianos Kampakis on Medium]]></title>
        <description><![CDATA[Stories by Dr. Stylianos Kampakis on Medium]]></description>
        <link>https://medium.com/@stylianos-kampakis?source=rss-db4ffe57dadc------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/2*n12jAdhOMKOmBB5ycrQOxw.jpeg</url>
            <title>Stories by Dr. Stylianos Kampakis on Medium</title>
            <link>https://medium.com/@stylianos-kampakis?source=rss-db4ffe57dadc------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Sun, 22 Nov 2026 23:30:36 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@stylianos-kampakis/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[Pollen DeFi 2.0: A Decentralised Next-gen Asset Management Protocol]]></title>
            <link>https://medium.com/pollen-defi/pollen-defi-2-0-a-decentralised-next-gen-index-82666194674f?source=rss-db4ffe57dadc------2</link>
            <guid isPermaLink="false">https://medium.com/p/82666194674f</guid>
            <category><![CDATA[investment]]></category>
            <category><![CDATA[blockchain]]></category>
            <category><![CDATA[cryptocurrency]]></category>
            <category><![CDATA[avalanche]]></category>
            <category><![CDATA[finance]]></category>
            <dc:creator><![CDATA[Dr. Stylianos Kampakis]]></dc:creator>
            <pubDate>Wed, 12 Jan 2022 14:41:33 GMT</pubDate>
            <atom:updated>2022-01-24T14:29:14.382Z</atom:updated>
            <content:encoded><![CDATA[<h3>What is Pollen DeFi?</h3><p><a href="https://pollen.id/">Pollen DeFi</a> is a new and exciting technology that can help you manage your cryptocurrency investment with other community members.</p><p>Millions of people invest in the stock market every day, but not everyone has the same access to financial opportunities. At Pollen, we’re creating a fully decentralised universal investment system that provides equal opportunity for all.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/640/1*Q1GggOQAm33vvl5VoXoGzg.gif" /><figcaption>Harvest alpha and make the best decision possible.</figcaption></figure><p>Traders who have confidence in their skills build reputation &amp; rank on the leaderboard to earn rewards. Users that do not want to actively trade can delegate tokens to successful managers of the portfolios who share them with those that hold the coin.</p><p>All users can gain benefits by accessing community managed assets.</p><p>So, hopefully, the name <a href="https://pollen.id/">Pollen</a> has started to make sense. The users operate like bees in a hive, exchanging information and collaborating to reach optimal decisions. A high level overview of how the system operates is shown in the figure below.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/570/0*M6pwANwnAAe-ICLT" /></figure><h3>Pollen DeFi Reputation</h3><p>At the core of Pollen’s protocol is reputation. Reputation determines who makes wise investment decisions and who doesn’t. Over time, reputation helps accumulate more and more power to users that make optimal decisions.</p><p>Deciding who is lucky and who is a skilled investor is a very challenging task. However, this hasn’t prevented Pollen’s modelling team from coming up with a variety of solutions:</p><ol><li>An algorithm that helps us fully understand skill level, allowing us to rank users</li><li>Statistical models for measuring reputation</li><li>Traditional and new metrics to measure financial performance</li></ol><figure><img alt="" src="https://cdn-images-1.medium.com/max/396/0*CY3jW89lTh27GEJt" /></figure><p>The goal behind all these methods is to ensure that the users that can make the best decisions climb at the top of the hierarchy, providing optimal returns for the protocol.</p><h3>DeFi 2.0</h3><p>Ever since <a href="https://www.olympusdao.finance/">OlympusDAO</a> came about, <a href="https://academy.binance.com/en/articles/what-is-defi-2-0-and-why-does-it-matter">DeFi2.0</a> has become a strong narrative in the blockchain space. While OlympusDAO is focused on protocol-owned liquidity, DeFi2.0 is trying to provide a general improvement over DeFi1.0 with improvements across multiple areas:</p><ol><li>User experience: Many protocols from DeFi1.0 are very complicated, and suffer from high gas-fees.</li><li>Decentralisation: DeFi2.0 has re-popularised DAOs, as many DeFi protocols ended up being very centralised.</li><li>Security</li></ol><p>Some ideas we have seen in this space include protocol-owned liquidity, insurance against impermanent loss.</p><p>Pollen identifies as a DeFi2.0 project. First of all, Pollen has a strong focus on user experience. You don’t have to be an expert investor to participate. The protocol has been designed in a way that anyone can join it, and delegate decisions to more experienced users if they want to. Secondly, it is not centralised. It has been designed with decentralisation as a goal from the onset. Finally, the Pollen team is designing a set of methods which can help unlock additional yield, in a true DeFi2.0 fashion.</p><h3>Pollen DeFi tokenomics</h3><p>The Pollen (PLN) governance token lies at the heart of all activity in the Pollen Decentralized Finance ecosystem. Users manage their own portfolios, and get rewarded when the make good decisions, or lose Pollen when they make bad decisions. Users that want to abstain from making decisions can delegate their governance tokens to someone else, and still receive rewards.</p><p>So, what do the tokenomics and the token distribution of the project look like? I think these are best explained by the graph below. You can also find a more in-depth explanation on <a href="https://medium.com/pollen-defi/pollen-pln-tokenomics-fdf8371f53d0">Medium</a>.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/625/0*WvuEF5sn0LTuP7Eo" /></figure><h3>Where can you learn more?</h3><p>A very good start is <a href="https://pollen-defi.medium.com/">Pollen’s medium</a>. You can learn more about the team on this <a href="https://medium.com/pollen-defi/pollen-defi-meet-the-team-behind-the-protocol-ced474ed5824">post</a>. You can find the PLN token on <a href="https://traderjoexyz.com/">traderjoe</a>. And obviously, you should visit <a href="https://pollen.id/">Pollen</a>‘s website and watch the video below.</p><p><a href="https://www.youtube.com/watch?v=8U0wvjn4oGU&amp;feature=emb_imp_woyt">https://www.youtube.com/watch?v=8U0wvjn4oGU</a></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*6Qv4Blg5yC6BwHyoa6S8JQ.jpeg" /></figure><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=82666194674f" width="1" height="1" alt=""><hr><p><a href="https://medium.com/pollen-defi/pollen-defi-2-0-a-decentralised-next-gen-index-82666194674f">Pollen DeFi 2.0: A Decentralised Next-gen Asset Management Protocol</a> was originally published in <a href="https://medium.com/pollen-defi">Pollen-DeFi</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[What is the F-1 measure and why is it useful for imbalanced class problems?]]></title>
            <link>https://stylianos-kampakis.medium.com/what-is-the-f-1-measure-and-why-is-it-useful-for-imbalanced-class-problems-11448de34732?source=rss-db4ffe57dadc------2</link>
            <guid isPermaLink="false">https://medium.com/p/11448de34732</guid>
            <category><![CDATA[machine-learning]]></category>
            <category><![CDATA[artificial-intelligence]]></category>
            <category><![CDATA[big-data]]></category>
            <category><![CDATA[statistics]]></category>
            <category><![CDATA[data-science]]></category>
            <dc:creator><![CDATA[Dr. Stylianos Kampakis]]></dc:creator>
            <pubDate>Fri, 19 Feb 2021 11:15:58 GMT</pubDate>
            <atom:updated>2021-03-07T14:22:28.280Z</atom:updated>
            <content:encoded><![CDATA[<p>By Dr. Stylianos Kampakis</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/922/0*C__7qu8w2bw95dhI" /></figure><h3>Metrics in machine learning</h3><p>As some of you might know, metrics in machine learning is one of my favourite topics. I have already spoken about <a href="https://thedatascientist.com/performance-measures-rmse-mae/">RMSE and MAE</a>, <a href="https://thedatascientist.com/performance-measures-cohens-kappa-statistic/">Cohen’s Kappa</a>, the <a href="https://thedatascientist.com/performance-measures-concordance-correlation-coefficient/">concordance correlation coefficient</a> and other <a href="https://thedatascientist.com/performance-measures-predictive-modelling/">metrics</a>.</p><p>On this post, I’d like to talk about another very important metric: The <a href="https://en.wikipedia.org/wiki/F-score">F-1 score</a>.</p><p>So, what is the F1-score and why would you want to use it?</p><h3>The F1-score</h3><p>The F-1 score is simply the harmonic mean between precision and recall. What is precision and recall? Let’s do a quick refresher. In a two-class classification problem, the confusion matrix is defined as below:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/421/1*3tj8gIG7UEk000ksvpbFMg.png" /></figure><p>So, based on that, the precision is defined as</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/265/1*8-lTOCO1RrTeOS-o5wq9gg.png" /></figure><p>and the recall is defined as</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/266/1*50c1lU1OUbz1BGbWRzu_yg.png" /></figure><p>The image below from Wikipedia provides a very nice overview.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/410/1*drhXaKEpBeEo2Typ5NWzYw.png" /></figure><p>In simple terms:</p><p><strong>Precision</strong>: If I predict that something is X, then what is the probability that it is really X?</p><p><strong>Recall</strong>: Out of all instances of X in the dataset, how many we did we manage to correctly collect?</p><p>So, I mentioned that the F1-score is the harmonic mean of the precision and recall.</p><p>What is the harmonic mean? For two numbers, the harmonic mean is simply defined as:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/217/1*-eWH87EdFrElfoCiTDqTiw.png" /></figure><p>For those of you who don’t know, this is one of the <a href="https://en.wikipedia.org/wiki/Pythagorean_means">3 Pythagorean means</a>, the other ones being the classic arithmetic average and the <a href="https://www.mathsisfun.com/numbers/geometric-mean.html">geometric mean</a>.</p><p>Why are we using the harmonic mean? The harmonic mean is the correct mean to use when we are averaging out <a href="https://corporatefinanceinstitute.com/resources/knowledge/other/harmonic-mean/">ratios</a>. Precision and recall are ratios, so the right way to average them is through the harmonic mean. So, the final formula simply becomes:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/761/1*RgfwWguY6A1LOclW0gkaKg.png" /></figure><h3>So, when to use the F-1 score?</h3><p>The F-1 score is very useful when you are dealing with imbalanced classes problems. These are problems when one class can dominate the dataset.</p><p>Take the example of predicting a disease. Let’s say that only only 10% of the instances in your dataset have the actual disease. This means that you could get 90% accuracy by simply predicting the negative class all the time. But, how useful is this? Not very useful, as you wouldn’t have predicted a single instance of the actual disease! This is where the F1-score can be very helpful. In this example, the recall for the positive class would be 0, and hence the F1-score would also be 0.</p><p>So, next time you are dealing with imbalanced classes you know that the F1-score is a far superior metric compared to accuracy. Another great metric is <a href="https://thedatascientist.com/performance-measures-cohens-kappa-statistic/">Cohen’s Kappa</a>, which you should make sure to check out.</p><p><em>Originally published at </em><a href="https://thedatascientist.com/f-1-measure-useful-imbalanced-class-problems/"><em>https://thedatascientist.com</em></a><em> on February 19, 2021.</em></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=11448de34732" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[What is the best way to become a data scientist?]]></title>
            <link>https://stylianos-kampakis.medium.com/what-is-the-best-way-to-become-a-data-scientist-876d054abc25?source=rss-db4ffe57dadc------2</link>
            <guid isPermaLink="false">https://medium.com/p/876d054abc25</guid>
            <category><![CDATA[course]]></category>
            <category><![CDATA[machine-learning]]></category>
            <category><![CDATA[artificial-intelligence]]></category>
            <category><![CDATA[big-data]]></category>
            <category><![CDATA[data-science]]></category>
            <dc:creator><![CDATA[Dr. Stylianos Kampakis]]></dc:creator>
            <pubDate>Wed, 30 Dec 2020 10:18:15 GMT</pubDate>
            <atom:updated>2021-01-19T10:27:16.256Z</atom:updated>
            <content:encoded><![CDATA[<p>By Dr. Stylianos Kampakis</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/740/0*BSjvODHWR5JnMw-j" /></figure><p>I’ve met many people who want to become data scientists, but are not sure how long they need to study for to achieve that. These days, with the plethora of content out on the web, and great data science schools like <a href="http://dataly.st/datalyst-academy">Datalyst Academy</a> , it has become easier than ever.</p><p>That being said, your background can play a huge role in how much you might have to study to become a data scientist. I you come from a technical background, and you are already familiar with things like coding and programming, then you can expect that you will progress faster from someone who has no prior experience in these fields. However, not having technical experience is not a drawback, as having a diverse background can also be proven very beneficial. There are different types of data scientists, as I’ve <a href="https://thedatascientist.com/tribes-data-scientists/">written about</a> in the past, and you just have to figure out what kind of data scientists you need to be.</p><p>I decided to create a simple tool (<a href="https://thedatascientist.com/best-way-become-a-data-scientist/">https://thedatascientist.com/best-way-become-a-data-scientist/</a>)to help you understand what kind of effort you should be putting in, if you want to become a data scientist. This tool (<a href="https://thedatascientist.com/best-way-become-a-data-scientist/">https://thedatascientist.com/best-way-become-a-data-scientist/</a>) has been created as an assistant to <a href="http://dataly.st/datalyst-academy">Datalyst Academy</a>.</p><p><strong>Datalyst Academy</strong> is a data science school I’ve created which combines the best elements from different modes of learning, and it has been based on my 10+ years of experience in the field.</p><p>Make sure to <a href="https://thedatascientist.com/contact/">get in touch</a> or message me via LinkedIn if you want to become a data scientist!</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*hKT3IbOm12TM6BDpZAbbpQ.png" /></figure><p><em>Originally published at </em><a href="https://thedatascientist.com/best-way-become-a-data-scientist/"><em>https://thedatascientist.com</em></a><em> on December 30, 2020.</em></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=876d054abc25" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Bayesian Statistics vs Frequentist Statistics]]></title>
            <link>https://stylianos-kampakis.medium.com/bayesian-statistics-vs-frequentist-statistics-c2681bcb58c8?source=rss-db4ffe57dadc------2</link>
            <guid isPermaLink="false">https://medium.com/p/c2681bcb58c8</guid>
            <category><![CDATA[data-science]]></category>
            <category><![CDATA[statistics]]></category>
            <category><![CDATA[big-data]]></category>
            <category><![CDATA[machine-learning]]></category>
            <category><![CDATA[artificial-intelligence]]></category>
            <dc:creator><![CDATA[Dr. Stylianos Kampakis]]></dc:creator>
            <pubDate>Wed, 16 Dec 2020 12:06:57 GMT</pubDate>
            <atom:updated>2021-01-05T17:20:11.331Z</atom:updated>
            <content:encoded><![CDATA[<p>By Dr. Stylianos Kampakis</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/988/1*2rQFrRuvF6sCsCfg3A_cCQ.jpeg" /></figure><p><em>Note: This is an excerpt from my new book-in-progress called “Uncertainty”.</em></p><p>The most popular definition of probability, and maybe the most intuitive, is the frequentist one. According to frequentists, an event’s probability is defined as the limit of this event’s frequency in a large number of trials.</p><p>What does this mean? Let’s go back, to the example of rolling a fair coin. You said that the probability of rolling heads on a single roll is 50%. However, how do you know this to be true? What if you roll tails 10 times in a row? Would this change the probability of rolling heads? Obviously not. Intuitively, this makes sense, but why?</p><p>I ran a coin-tossing experiment (simulated in the R programming language<a href="https://thedatascientist.com/bayesian-statistics-vs-frequentist-statistics/#_ftn1">[1]</a>) and you can see the results below. The proportion of heads very quickly converges to 50%.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/373/0*Xp_00CsT31onFP_q" /></figure><p>Figure 1. Coin tossing experiment</p><p>This is the definition of frequentist probability in practice. If you execute an experiment a large number of times, then the frequencies will converge to the true probabilities.</p><p>Frequentist statistics have been the orthodox branch of statistics for most of its history. Statistics is based on the idea that you can extract a sample from a population, and then study properties of the population. If we treat each entity in this sample as an experiment, then the more samples we collect, the closer we will get to the truth.</p><p>However, there are certain events for which this definition makes no sense. Many of these events happen to be very important for our lives, and they are unique. Some examples include elections and sports events. We can’t run an experiment of 1000 elections and see what happens. Also, we can’t run 1000 Champions League finals to find the true probability of Real Madrid winning it.</p><p>In those cases, the frequentist definition of probability seems to get us into trouble. This is where the Bayesian definition of probability comes to our rescue.</p><p>The term Bayesian is due to Reverend Thomas Bayes (1701?-1761), pictured below.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/225/0*arwgmCe5xBkRWNyv" /></figure><p><em>Figure 2. Portrait of Thomas Bayes. We are not 100% certain about whether this is actually his, but it is the only one we got. Got the irony there, about being uncertain about the face of the man who is responsible for one our best tools against uncertainty?</em></p><p>Bayes used <em>conditional probability</em> (which we will explain about shortly) in his essay <em>“An Essay towards solving a Problem in the Doctrine of Chances”, </em>which was presented in the Royal Society in 1763. In case you noticed the year of his death in the previous paragraph, you will realize that this essay was presented posthumously. It was Bayes’ friend Richard Price (1723–1791) who discovered Bayes’ notes and published his work.</p><p>Bayes’ work concerned the following problem: How can you know the probability of an event, based only on how many times it occurred or didn’t occur in the past? Bayes used a thought experiment to illustrate his argument.</p><p>Bayes has his back turned towards a table, and his assistant throws a ball on it. There is equal probability of the ball falling anywhere on the table. Bayes has to guess where the ball is. On that first through, Bayes experiences the maximum degree of uncertainty, as the ball could really be anywhere on the table.</p><p>In the next step, his assistant throws another ball and reports whether it fell on the left or the right side of the first one. Let’s say that this time the ball lands on the right side of the table. We can assume that the the first ball now is more likely to be on the left side. If the original ball had landed on the left side, then the right side would have more space for another ball to land.</p><p>Then, the assistant throws another ball and then this ball lands to the right again. This makes it even more likely that the original ball lies on the left. Hence, with each throw, we narrow down the position of the original thrown ball more and more.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/525/0*x9byF2Hf_8zvgY-f" /></figure><p><em>Figure 3. Depiction of Bayes’ argument. After the first ball throw (black ball), the second throw (orange ball), has more space on the right of the table than on the left. On the second row (left), you can see that there are only a few positions where the second ball could end up left of the original ball. On the second row (right) graph, you see all the positions right of the original ball, where the orange one could lie. It is clear that there is more space right of the original ball, and most of this space, lies to the right of the table. However, it is still possible for the orange ball to be right of the black ball, but on the left side of the table. So, 2 ball throws are not enough to completely pinpoint the location of the black ball.</em></p><p>In the modern world, this might remind us a bit of the Battleships board game. In battleships, each player (the game has only 2 players), has to place ships on a square board. On each round, the players specify one point in the board which they attack. If they hit part of a ship, then the opposing player informs them that they managed to hit a ship, but no other information is given. As you can see in the figure below, there are two ways to place the ships on the board (horizontally, as well as vertically). So, a player has to first figure out where a ship is, by getting an initial hit, and then guess where the rest of the ship lies. By trying out more and points on the board, the uncertainty around the location of the ships is reduced round by round.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/591/0*o40UmERCsNTBp25w" /></figure><p><em>Figure 4. A version of the Battleships board game.</em></p><p>So the main concept behind Bayes’ idea was the following:</p><p><em>Initial belief + new information = New belief</em></p><p>In modern era this is now turned into:</p><p><em>Prior + Likelihood = Posterior</em></p><p>We will explain more about this below. Interestingly enough, Bayes never focused on the theorem that now bears his name. The mathematicians of his time were not very happy with his approach, for two reasons. First, guessing didn’t sound too rigorous of a process. Secondly, in absence of information, Bayes assumed that all outcomes are equally likely. In modern parlor, we would say that the prior is uniform. Having to assign a prior probability of belief, seemed like an additional hurdle.</p><p>The theorem was independently discovered by one of the most prominent mathematician of all times: Pierre-Simon de Laplace. Laplace re-discovered this principle and published it in 1774. Eventually, he learned of Bayes’ discovery in 1781 when Price visited Paris. Laplace improved his formulation and decided to test it out.</p><p>What Laplace studied was whether the observation that more boys than girls were being born was a law of nature or just a statistical anomaly. He collected records from London, Paris, St. Petersburg, rural areas in France, Egypt and Central America. Using his theorem, he managed to conclude that indeed this seems to be a law of nature<a href="https://thedatascientist.com/bayesian-statistics-vs-frequentist-statistics/#_ftn2">[2]</a>.</p><p>Laplace went on to make major contributions to other scientific fields, like Astronomy. There is a famous expression attributed to Laplace. Story says, that Napoleon asked him why he did not include God in his explanations of the movement of celestial objects, to which Laplace famous answered: <em>Je n’avais pas besoin de cette hypothèse-là. (“I had no need of that hypothesis.”</em></p><p>Laplace was part of the human era, alongside Newton, that instigated a scientific revolution, with mathematics and reason being the main tools to explain away uncertainty, in place of religious and metaphysical explanations.</p><p>While Laplace did most of the work on Bayes’ theorem, his name was never attached to it.</p><p>Laplace later on discovered the <strong>central limit theorem</strong>, one of the most powerful and significant findings in modern mathematics (we explain more about this later). Upon discovering this, he realized that once we possessed large amounts of data, the Bayesian approach converged to the traditional frequentist approach. And so Laplace converted to frequentism, which he abided to until the end of his life.</p><p><a href="https://thedatascientist.com/bayesian-statistics-vs-frequentist-statistics/#_ftnref1">[1]</a> The R language ( <a href="https://www.r-project.org/">https://www.r-project.org/</a>) is the most popular language for statistics. Python, however, seems to have overtaken it for machine learning.</p><p><a href="https://thedatascientist.com/bayesian-statistics-vs-frequentist-statistics/#_ftnref2">[2]</a> We now know the ratio to be approximately 105 boys to 100 girls: <a href="http://www.searo.who.int/entity/health_situation_trends/data/chi/sex-ratio/en/">http://www.searo.who.int/entity/health_situation_trends/data/chi/sex-ratio/en/</a></p><p><em>Originally published at </em><a href="https://thedatascientist.com/bayesian-statistics-vs-frequentist-statistics/"><em>https://thedatascientist.com</em></a><em> on December 16, 2020.</em></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=c2681bcb58c8" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[How Accepting Cryptocurrency Can Benefit Your SME]]></title>
            <link>https://stylianos-kampakis.medium.com/how-accepting-cryptocurrency-can-benefit-your-sme-52a615219a55?source=rss-db4ffe57dadc------2</link>
            <guid isPermaLink="false">https://medium.com/p/52a615219a55</guid>
            <category><![CDATA[fintech]]></category>
            <category><![CDATA[business]]></category>
            <category><![CDATA[bitcoin]]></category>
            <category><![CDATA[cryptocurrency]]></category>
            <category><![CDATA[blockchain]]></category>
            <dc:creator><![CDATA[Dr. Stylianos Kampakis]]></dc:creator>
            <pubDate>Mon, 21 Sep 2020 11:34:33 GMT</pubDate>
            <atom:updated>2020-10-05T12:54:53.340Z</atom:updated>
            <content:encoded><![CDATA[<p>By Dr. Stylianos Kampakis</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/844/1*oGdqIQQlNo7EwKUHP9gcJg.png" /></figure><h3>Cryptos for smaller organisations</h3><p>With blockchain technology gaining greater influence across a range of industries, I’ve written on <a href="https://thedatascientist.com/sme-benefit-blockchain-technology/">the subject of ‘How Your SEM Can Benefit From Blockchain Technology’</a>. In doing so, I alluded to numerous possibilities, including blockchain-based cloud storage, smart contracts, and handling digital payments. Here, I’m going to dig a little bit deeper into that last idea, and discuss some of the ways in which choosing to accept payment in cryptocurrency can boost your business.</p><p><strong>1. Tapping Into Young Markets</strong></p><p>While not <em>every</em> SME is aiming to attract young consumers, most would agree that it can only be a positive thing. And it’s certainly one of the benefits that can come of accepting cryptocurrency. According to a <a href="https://www.cnbc.com/2018/07/30/more-millennials-than-older-adults-say-bitcoin-is-the-best-way-to-save.html">CNBC write-up on interest in bitcoin</a>, millennials are up to five times as likely to say that bitcoin is “the best way to invest for the future” as older adults. That doesn’t necessarily mean that the same millennials are looking to <em>spend</em> bitcoin. But the interest and trust in cryptocurrency is there, and it sets the stage for more spending activity. Putting it simply, by accepting cryptocurrency payments you’ll likely make your SME more appealing to some young consumers.</p><p><strong>2. Increasing Transaction Volume</strong></p><p>The implication of attracting young consumers is that there will be more business to go around. And this is another general benefit your SME could enjoy by making the move to accept cryptocurrency. An <a href="https://www.fisglobal.com/en-gb/merchant-solutions-worldpay/large-enterprise-business/b2c-ecommerce/digital-content">overarching look at digital payment methods by FIS Global</a> points to the various ways in which digital methods can actually increase transactions. The write-up mentions improved transactions across borders, for instance, as well as the fact that simplified crypto payments (like crypto card acceptance) can lead to more new customers. Basically, we’d posit that there are numerous ways in which accepting cryptos can increase transaction volume, without any negative trade-off (or much effort on your part).</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/960/0*kwtyf-KaI2TSkoI_.jpg" /></figure><p><em>Cryptocurrency adoption. Source: </em><a href="https://www.forbes.com/sites/billybambrough/2019/12/08/bitcoins-massive-multi-trillion-dollar-question/#6b1d5e8e5dd8"><em>Forbes</em></a><em>.</em></p><p><strong>3. Setting Up the Future</strong></p><p>Preparation is also something to consider. Even if you don’t fully buy into the notion that accepting cryptocurrency will meaningfully impact your SME now, doing so will position you to adapt to future developments in the crypto world. So, for example, if bitcoin gains more traction in the next few years, or if new crypto options emerge that are favoured for payments for whatever reason, you’ll be equipped to quickly accommodate shifting consumer preferences.</p><p><strong>4. Speeding Up Transactions</strong></p><p>Yet another positive effect of embracing cryptocurrency with your SME is that you may stand to offer faster transaction times to consumers. This may not be a particularly big issue for in-person payments, if your business has a retail component. But where online transactions are concerned, speed tends to be one of the main reasons that some favour cryptos. Quick processing is not just preferred, but <em>expected</em> by many modern consumers, and by at least providing a crypto option you’ll be embracing and catering to that expectation.</p><p><strong>5. Lowering Transaction Costs</strong></p><p>Last but not least is the fact that you may also stand to lower the costs of the digital transactions your SME processes. These costs never seem like much in the moment, but over time businesses tend to lose quite a lot of money in processing fees paid to services like PayPal and its counterparts — whereas crypto payments occur at no cost. According to <a href="https://medium.com/blockdata/merchants-can-reduce-transaction-processing-costs-by-70-with-crypto-payments-6d1d343f3f75">Blockdata’s look at crypto savings for merchants</a>, businesses can “reduce transaction processing costs by 70%” by embracing crypto payments.</p><h3>The future of crypto-payments for SMEs</h3><p>So, does it make sense for an SME to accept crypto payments? I think it does. Adoption rates of cryptocurrencies, and especially Bitcoin, increasing. It is only a matter of time until consumers expect cryptocurrency payments to be available, pretty much like card payments in retail stores. The SMEs that make this step early, are the ones that will benefit the most. So, if you are the owner of an SME, make sure you learn more about blockchain now, before the competition gets ahead of you.</p><p><em>Originally published at </em><a href="https://thedatascientist.com/how-accepting-cryptocurrency-can-benefit-your-sme/"><em>https://thedatascientist.com</em></a><em> on September 21, 2020.</em></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=52a615219a55" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[What is Consciousness-based AI and why you should care]]></title>
            <link>https://stylianos-kampakis.medium.com/what-is-consciousness-based-ai-and-why-you-should-care-e57556d358e7?source=rss-db4ffe57dadc------2</link>
            <guid isPermaLink="false">https://medium.com/p/e57556d358e7</guid>
            <category><![CDATA[data-science]]></category>
            <category><![CDATA[artificial-intelligence]]></category>
            <category><![CDATA[machine-learning]]></category>
            <category><![CDATA[big-data]]></category>
            <category><![CDATA[computer-science]]></category>
            <dc:creator><![CDATA[Dr. Stylianos Kampakis]]></dc:creator>
            <pubDate>Tue, 07 Jul 2020 10:52:38 GMT</pubDate>
            <atom:updated>2020-07-28T13:57:00.462Z</atom:updated>
            <content:encoded><![CDATA[<p>By Dr. Stylianos Kampakis</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/617/1*V5WzCWYei0GY8qI5aw4IfA.png" /></figure><h3>A short history lesson on AI</h3><p>The holy grail of AI research is to create <a href="https://thedatascientist.com/general-ai/"><strong>Artificial General Intelligence</strong></a> (AGI). AGI is what we will get once we get a machine to operate on the same level of intelligence as a human.</p><p>There have been countless theories and approaches over the years trying to achieve this goal, but, so far, none of them have succeeded.</p><p>The first wave of theories around the 50s and the 60s, was focused on using logical rules in order to represent knowledge and make machines reason like humans. This was based on the assumption that the property that makes humans special is the ability to reason in a logical manner. For example, humans can easily understand logical statements of the form</p><p>“If all men are mortal and Socrates is a man, then Socrates is a mortal”.</p><p>This approach is now termed “ <a href="https://en.wikipedia.org/wiki/Symbolic_artificial_intelligence#:~:text=John%20Haugeland%20gave%20the%20name,Old%2DFashioned%20Robotics%22).">Good Old-Fashioned AI</a>”. It was invented by researchers like <a href="https://en.wikipedia.org/wiki/John_McCarthy_(computer_scientist)">John McCarthy</a> and <a href="https://thedatascientist.com/marvin-minsky-reflections-ai-research/">Marvin Minsky</a>. While this approach helped birth AI as a scientific field, it, unfortunately, ended up being very limited.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/349/1*GqPZ5hTSgX8XgqVoShQNdg.png" /></figure><p>Research then moved on to other approaches like <a href="https://thedatascientist.com/machine-learning-vs-ai/">machine learning</a> and neural networks. The 90s saw the rise of what was called “computational intelligence”. This was an approach inspired by biology. Computational intelligence had three main “arms”: neural networks, genetic algorithms (or evolutionary computation) and fuzzy logic.</p><p>The idea behind <a href="https://cis.ieee.org/about/what-is-ci#:~:text=Computational%20Intelligence%20(CI)%20is%20the,Fuzzy%20Systems%20and%20Evolutionary%20Computation.">computational intelligence</a> was very attractive. Since intelligence exists in nature, we should be able to replicate intelligence if we mimic natural processes. For example, ants display a kind of intelligence in the way they are organised. By mimicking the way they organise we can create algorithms which display the same kind of intelligence.</p><p>A very common analogy was that of the birds’ flight. While feathers is a way for an organism to develop the ability to fly, it is not a necessary prerequisite, as there can be other ways to generate flying organisms or machines. However, by creating machines that imitate birds, we can identify the core features that let us create machines that can fly.</p><p>Unfortunately, computational intelligence didn’t manage to give us true general AI. It lacked a proper mathematical theoretical framework which could guide research and provide satisfactory explanations as to how these methods worked.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/372/1*zhmZM7aUxQc-mCqqZr3qng.png" /></figure><p>These days, the latest iteration in the effort to create general AI is <a href="https://thedatascientist.com/deep-neural-network-go/">deep learning</a>. But, even this might not be enough. While deep learning has been very successful in multiple tasks like computer vision and natural language processing, it has been criticised as simply a very powerful curve fitting machine, but nothing. Judea Pearl claims that it is simply a more powerful version the the same thing that algorithms have been doing for the last 150 years. According to his opinion, the only way to generate real intelligence, is to create an algorithm which can understand causality.</p><p>In any case, it looks that the debate is far from over. So, let’s go back to the title of this post. What is consciousness-based AI and what does this have to do with general AI?</p><h3>Consciousness-based AI</h3><p>Computational intelligence tried to imitate nature in order to generate AI and it failed. Other approaches have tried to use mathematical wizardry, in order to do the same.</p><p>I don’t believe that the approach behind computational intelligence was wrong. I think that you can create real AI through imitation. But you need to focus on the right thing to imitate. It is not nature that displays intelligence. It is consciousness.</p><p>So, let me elaborate on that. What do I mean by <strong>consciousness</strong>? There are many definitions, and the truth is that it is something very difficult to bring into words. But all of us, more or less, understand what it is to have a subjective conscious experience.</p><p>What kind of entities unequivocally display intelligent behaviour? What is the common theme amongst all of them? If you think about it for a second, you will soon discover that it is consciousness.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/281/1*80NwcInJJcD3-YXf2K4WHA.png" /></figure><p>From a microscopic ant, to birds and humans, any kind of creature that is conscious, is also intelligent, at least to a small extent. Are all intelligent organisms conscious? Well, if you believe that a bee hive is an intelligent organism, then the answer would be no. But the <em>components </em>of the bee hive, the bees themselves, are conscious.</p><p>So, what does this have to do with artificial intelligence? What if we focused our efforts into replicating the elements of consciousness? Would this lead to the creation of an artificially intelligent machine? I believe it would.</p><p>Think in how many ways you can break down consciousness and the substrates that affect it. We can talk about the cognitive layer, the neural layer (brain), the biological layer (gut-brain axis, etc.), and potentially add other layers the deeper we go.</p><p>Once you start analysing all these pieces, then the aggregation provides the basis for intelligence.</p><h3>A new dawn for Conscious AI</h3><p>Can AI really be conscious? If we could break down all components of consciousness, then I believe we could replicate intelligence. And we could even use tools like <a href="http://integratedinformationtheory.org/">Integrated Information Theory</a> in order to create a <strong>consciousness equivalent</strong> in a machine. Integrated Information Theory gives us the tool to measure the quantity of consciousness in any organism. The theory implies some kind of panpsychism, where, for example, organisms like plants are conscious, but have a lower <strong>Phi</strong> (the measure of consciousness) than humans or animals.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/389/1*nGS6imz_2BZfoB0voQL_6Q.png" /></figure><p>What if we could measure someone’s Phi and then create an algorithm that possesses an equivalent Phi? While this machine might not be conscious, it would be a <strong>consciousness equivalent</strong> of someone’s brain. And this artificial construct should be as intelligent as the real person.</p><p>But are there any instances of intelligence where there is no consciousness? Our autonomic nervous system, for example, can react to changes in the internal or external environment, but we wouldn’t call it conscious (even though it might be). This is a simple kind of intelligence, and I would argue that higher levels of intelligence, usually imply some kind of conscious experience.</p><p>Is this thesis the one to lead us to <strong>Artificial General Intelligence</strong>? I think it might well be. That’s why I decided to become part of <a href="http://withinu.ai/">Within[u]</a>, so that we can do the necessary research to find out. I can’t share many details for now, but what I can say is that we are looking for researchers on all kinds of field relating to AI, cognitive science, neuroscience and biology.</p><p>If you are curious about this, just drop me a <a href="https://thedatascientist.com/contact/">message</a>.</p><p><em>Originally published at </em><a href="https://thedatascientist.com/consciousness-based-ai/"><em>https://thedatascientist.com</em></a><em> on July 7, 2020.</em></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=e57556d358e7" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[The Different Types of AI Explained]]></title>
            <link>https://stylianos-kampakis.medium.com/the-different-types-of-ai-explained-93e750d061bf?source=rss-db4ffe57dadc------2</link>
            <guid isPermaLink="false">https://medium.com/p/93e750d061bf</guid>
            <category><![CDATA[big-data]]></category>
            <category><![CDATA[machine-learning]]></category>
            <category><![CDATA[artificial-intelligence]]></category>
            <category><![CDATA[data-analysis]]></category>
            <category><![CDATA[data-science]]></category>
            <dc:creator><![CDATA[Dr. Stylianos Kampakis]]></dc:creator>
            <pubDate>Thu, 02 Jul 2020 10:51:06 GMT</pubDate>
            <atom:updated>2020-08-24T12:29:32.512Z</atom:updated>
            <content:encoded><![CDATA[<p>By <a href="https://thedatascientist.com/">Dr. Stylianos Kampakis</a></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/735/1*gvgqmVTiqpBvLf0kCPlOTA.png" /></figure><p>Artificial intelligence (AI) is not only an emerging technology but it is also highly complex. There isn’t even a settled definition of AI, not least because AI is a catch-all term that encompasses various types of technology. This provides plenty of opportunity to theorise as to what the future will hold. The news is very much focused on what <a href="https://thedatascientist.com/ai-startups-disrupting-economy-work/">technology can do now</a>, whether it is about <a href="https://thedatascientist.com/deep-neural-network-go/">AI beating humans in games</a>, asking whether <a href="https://thedatascientist.com/should-we-fear-ai/">AI will steal our jobs</a>, or discussing about the potential of <a href="https://thedatascientist.com/the-rise-of-ai-for-good/">AI to improve society</a>. It is good to take a step back and try and look at the bigger picture of what can be achieved within the next 100 years.</p><p>We’ll look at two different methods of categorisation, one comparing the technology to the human mind and the other to the functions that AI can perform.</p><h3>Classifying AI by Comparing It to the Human Mind</h3><p>Under this method of classification, the types of AI include:</p><p>* <strong>Artificial Narrow Intelligence (ANI)</strong> — ANI is the level of AI that we currently have. It will probably also cover much of the AI that will be developed for many, many years to come. ANI machines can autonomously perform tasks in a similar way to humans.</p><p>* <strong>Artificial General Intelligence (AGI</strong>) — AGI machines are perceived as machines that can function like a human. So rather than being able to perform and get better at a single task, it will be able to perform multiple tasks, learn completely new competencies, and solve problems it has never encountered before.</p><p>* <strong>Artificial Super Intelligence (ASI)</strong> — Like AGI, ASI is purely hypothetical. It is an AI that is far more intelligent than even the most brilliant minds on Earth. This is the type of AI that causes the most fear.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/437/0*80XThNMRBjekhiqL" /></figure><h3>The Four Types of AI Based on the Functions It Can Perform</h3><p>Let’s now move to the other categorisation of AI-defining AI based on the functions it can perform. Under this method of categorisation, there are four main types:</p><p><strong>1. Reactive Machines</strong></p><p>Reactive machines base their decisions and actions on the existing situation. This means they can’t use past experiences to improve their decision-making process. In other words, they can’t learn.</p><p>One of the best-known examples of a reactive machine is <a href="https://en.wikipedia.org/wiki/Deep_Blue_(chess_computer)">Deep Blue</a>. In 1996/1997, it became the first computer programme to beat a human at chess when it defeated Gary Kasparov, the Russian chess grandmaster. Deep Blue was programmed to predict the moves it should make. It also predicted the moves its opponent might make, but it could never improve beyond this originally programmed capability.</p><p><strong>2. Limited Memory Machines</strong></p><p>This is the type of AI that we have today, such as autonomous vehicles, personal assistants like Siri, chatbots, face recognition technology, etc. In simple terms, limited memory machines are programmed to perform specific tasks, but they are also able to learn. This learning helps them improve.</p><p>Why do they use limited memory rather than unlimited though? Autonomous vehicles are a good example as they store some bits of data for a limited period of time. This includes the distance to other vehicles, the speed of other vehicles, the current speed limit, etc. When they are in a situation where this information is no longer relevant, this data is discarded. Think how this contrasts to a human brain, which can keep memories for years and years, and can retrieve them when in the right context and time.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/602/0*yBaPujWwDEHHBAt_" /></figure><p><strong>3. Theory of Mind</strong></p><p>With <a href="https://en.wikipedia.org/wiki/Theory_of_mind">theory of mind</a> AI, we are back in the realm of the hypothetical. It involves developing machines that can understand emotions, desires, needs, and beliefs. Think of this as a machine that you can have a social interaction with in much the same way that you interact with another person. This is more of the type of AI you expect to see in movies.</p><p><strong>4. Self-Aware</strong></p><p>Taking the theory of what may be possible as far as it can go is self-aware AI, although it is purely hypothetical at the moment. Self-aware AI not only understands human emotions, needs, and beliefs but it will have emotions, desires, needs, and beliefs of its own.</p><h3>AI Today and in the Future</h3><p>Whether the hypothetical iterations of AI are possible or not, there are still substantial developments occurring in the AI technologies we have today. It is very likely we will have an AI that possesses some theory of mind within this century. However, creating a self-aware AI or some kind of super-intelligence is something which still sits in the realm of science fiction. While it cannot be precluded, no-one knows how to get there. The one thing that is certain, is that the future looks exciting and it is going to be much different to what we see now!</p><p><em>Originally published at </em><a href="https://thedatascientist.com/different-types-ai-explained/"><em>https://thedatascientist.com</em></a><em> on July 2, 2020.</em></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=93e750d061bf" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[How Your SME Can Benefit from Blockchain Technology]]></title>
            <link>https://stylianos-kampakis.medium.com/how-your-sme-can-benefit-from-blockchain-technology-6e0d5dd71ad?source=rss-db4ffe57dadc------2</link>
            <guid isPermaLink="false">https://medium.com/p/6e0d5dd71ad</guid>
            <category><![CDATA[blockchain]]></category>
            <category><![CDATA[cryptocurrency]]></category>
            <category><![CDATA[finance]]></category>
            <category><![CDATA[blockchain-technology]]></category>
            <category><![CDATA[investing]]></category>
            <dc:creator><![CDATA[Dr. Stylianos Kampakis]]></dc:creator>
            <pubDate>Wed, 10 Jun 2020 11:58:31 GMT</pubDate>
            <atom:updated>2020-07-04T13:38:38.028Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/686/1*tfbUZ9f7qVN0SHWVEinvIg.png" /></figure><p><a href="https://thedatascientist.com/blockchain-technology-matured-since-2018-crypto-bubble/">Blockchain technology</a> is more commonly talked about in relation to corporations. This includes everything from banks streamlining their transactions and the sale of products to manufacturers improving their supply chain and global tech companies utilising blockchain technologies for any number of applications. What about SMEs though? Are there any benefits to investing in blockchain, and are there any applications or practical use examples that deliver a return on investment? After all, blockchain is still in its early stages. Unlike corporations, SMEs can’t afford the cost and don’t have the resources to take a punt on something so new. That said, we are now starting to see blockchain technologies, also known as distributed ledger technologies, that are both practical and cost-effective for SMEs and that deliver a return.</p><h3>Why Should You Care About Blockchain?</h3><figure><img alt="" src="https://cdn-images-1.medium.com/max/458/0*xCUeBqMDAnJOViq4" /></figure><p>As an SME, you get all the same benefits from blockchain that a corporation gets. This includes:</p><ul><li>The decentralised nature of the technology, which removes middlemen.</li><li>Complete transparency.</li><li>Immutable records that can’t be changed once they are created, improving the auditing process.</li><li>Efficiency plus the enablement of fast transactions.</li><li>Reducing costs when implemented properly.</li><li>Using highly effective encryption technology so it can improve security.</li></ul><h3>How Can You Use Blockchain Technologies as an SME?</h3><figure><img alt="" src="https://cdn-images-1.medium.com/max/510/0*dgSdWZLJeT6ozc5w" /></figure><p>What about practical examples of how SMEs can implement distributed ledger technologies? Here are some that are possible uses.</p><ol><li><strong>Transferring Money, Making Payments, and Receiving Payments</strong></li></ol><p>One of the first uses of blockchain technology is related to finance and the transfer of money. Businesses are increasingly utilising the opportunities this presents. Accepting payments in currencies like <a href="https://bitcoin.org/">Bitcoin</a> is the obvious example. You can also use tools and applications to pay suppliers or remote employees in a way that is faster, more secure, and lower cost than traditional methods. This usually happens through public blockchains, which face some issues, like scalability. However, there are improvements taking place in this area. Plus, there are some solutions like <a href="https://www.ripple.com/">Ripple</a>, which are not public blockchains, but they benefit from very fast transaction speeds. One major reason an SME might want to do this, is to facilitate payments from other countries and continents, where money transfers might be take a long time.</p><ol><li><strong>Cloud Storage Based on Blockchain Technology</strong></li></ol><p>Cloud storage solutions built using blockchain technology are known as <strong>distributed cloud storage solutions.</strong> A standard cloud storage provider rents out the storage space it has in a data centre. Distributed cloud storage, however, is essentially a peer-to-peer network. So a company or individual can rent out their unused hard-drive space, with blockchain technology ensuring its security. For most SMEs, this means you have access to lower-cost cloud storage solutions as distributed cloud storage is much cheaper than mainstream cloud storage providers. They also deliver excellent performance in relation to download and upload speeds. This is because the data is distributed so there are multiple downloads and uploads happening at the same time. Some example solutions in this space are <a href="https://storj.io/">Storj</a> and <a href="https://sia.tech/">Sia</a>.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/623/0*cQm5R1bLr20lyf4n" /></figure><p>With a smart contract, all parties involved in the contract set and agree on the conditions. There is full transparency throughout the process, with all entries to the ledger becoming available in real time. In addition, the contract is highly secure, plus the blockchain technology ensures the conditions are met before the contract can move to the next stage (e.g., releasing payment). Furthermore, the terms and penalties that are agreed in the contract execute automatically, without the involvement of a middleman. There is also no need for multiple approval processes as approvals are automated. The result is a more transparent and efficient process that saves money, saves time, improves cashflow, minimises disputes, and more. The first blockchain to use smart contracts was the <a href="https://ethereum.org/">Ethereum</a> blockchain. Smart contracts are easier to develop that most people think, with tools like the <a href="https://remix.ethereum.org/">Remix IDE</a> being freely available. In some cases, you don’t even have to code. For example, the <a href="http://erc721.org/">erc-721</a> standard, lets you create non-fungible tokens, which can be used to identify asset ownership. You could use this, for example, to store ownership of a <a href="https://thedatascientist.com/limits-ai-creativity/">work of art</a> on the blockchain, or of some other asset your company possesses.</p><h3>Blockchain for SMEs</h3><p>The above are just three examples of how SMEs can use blockchain technologies today. The number of practical applications available-and the take-up of these applications-is set to increase. If you want to find out whether you would actually benefit from using blockchain make sure to check out the <a href="https://tesseract.academy/free-tools-data-science-blockchain/">Blockchain Framework by the Tesseract Academy</a>.</p><p><em>Originally published at </em><a href="https://thedatascientist.com/sme-benefit-blockchain-technology/"><em>https://thedatascientist.com</em></a><em> on June 10, 2020.</em></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=6e0d5dd71ad" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[How COVID-19 will affect the AI industry]]></title>
            <link>https://stylianos-kampakis.medium.com/how-covid-19-will-affect-the-ai-industry-2f30a0925225?source=rss-db4ffe57dadc------2</link>
            <guid isPermaLink="false">https://medium.com/p/2f30a0925225</guid>
            <category><![CDATA[covid19]]></category>
            <category><![CDATA[coronavirus]]></category>
            <category><![CDATA[artificial-intelligence]]></category>
            <category><![CDATA[big-data]]></category>
            <category><![CDATA[technology]]></category>
            <dc:creator><![CDATA[Dr. Stylianos Kampakis]]></dc:creator>
            <pubDate>Thu, 28 May 2020 10:59:28 GMT</pubDate>
            <atom:updated>2020-06-22T12:23:01.067Z</atom:updated>
            <content:encoded><![CDATA[<p>By Dr. Stylianos Kampakis</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/355/1*F_IFIhfmL7stsEunQ8w62w.png" /></figure><h3>COVID-19 is impacting the economy</h3><p>COVID-19 had a massive impact on the world economy so far . We know that many things are going to change form now on. First of all, companies will have to adjust to the <a href="https://www.bbc.co.uk/news/business-51706225">economic impact</a>. Secondly, it is possible that the virus might stay for longer than expected, and measures like social distancing might become part of our lives.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/508/0*yDU5V4tOGq6zDv2e" /></figure><p>Thirdly, it is very likely that companies from now on will have to try and be “pandemic-proof”. While the resurgence of a new pandemic is a low probability event, it can be a catastrophic event for many industries. The restaurant industry now is suffering. The airline industry as well. <a href="https://www.theguardian.com/business/live/2020/may/28/easyjet-job-cuts-30percent-workforce-shrink-fleet-covid19-coronavirus-stocks-markets-jobless-pound">EasyJet</a> is planning to cut a significant part of its workforce, as it is not expecting to recover before 2023.</p><p>So, we are looking towards some fundamental changes in the way we conduct business. How can AI help us out?</p><h3>COVID-19 and AI opportunities</h3><p>I believe that artificial intelligence and data science will play a massive role in re-designing whole sectors of the economy. I believe one of the biggest contribution will be around automation. I have already spoken in the past how AI can help facilitate an <a href="https://thedatascientist.com/ai-economy-really-look-like/">automation economy</a>. Until now, automation was a nice-to-have for many organisations, but maybe businesses didn’t care too much about it. But now, automation seems to becoming a great solution to make a business pandemic proof.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/640/0*-YoCUM_qaXpL35dF" /></figure><p>Machines don’t get sick, and they don’t require social distancing measures. If a business can use robots for some of its operations, then it will ensure it will keep functioning, in the face of any pandemic.</p><p>Another great opportunity for AI is in health tech. I already wrote an article about how <a href="https://thedatascientist.com/ai-data-science-help-fight-covid-19/">AI can be used to fight COVID-19</a>. Drug and vaccine discovery through the use of AI is going to dominate the headlines on the tech news for the next 3–5 years. Something else that came out of this crisis is that there are more applications of AI in medicine, besides the traditional application of using computer vision for diagnoses. <a href="https://www.haaretz.com/israel-news/.premium-israeli-researchers-hope-ai-can-tame-covid-19-and-they-want-our-help-1.8681366">Israel</a>, for example, is using AI to predict where the next outbreak will take place.</p><p>All this will help improve the awareness of AI’s possibilities in health. Investment will be much easier to find. We are sure to see many more companies in this area.</p><p>Also, we might see greater use of health monitoring devices. The current crisis might make people more health aware. Fitness trackers, like FitBit, have already been out for many years, but they leave many interesting variables out, like heart rate variability. I wouldn’t be surprised if we see more advanced trackers coming out in the next few years, with a strong focus on health and disease prevention, leading to a new explosion of health-related data. This could help create massive opportunities for data science and AI companies.</p><h3>AI and the new economy</h3><p>AI had already started to <a href="https://thedatascientist.com/ai-startups-disrupting-economy-work/">transform our economy</a>. The coronavirus pandemic will simply speed up the pace of change. Companies that are aware of the changes that AI is going to bring, and are taking appropriate measures, will get ahead of the curve. Companies that have ignored the benefits of automation and related technologies are now being hit in a very hard way.</p><p>If you want to know more about how AI feel free to <a href="https://thedatascientist.com/contact/">get in touch</a>, check out the <a href="https://thedatascientist.com/data-science-services/">services</a> I provide, or visit the <a href="http://tesseract.academy/">Tesseract Academy</a>’s webpage where you can find workshops on topics like <a href="https://thedatascientist.com/data-science-coaching/">data strategy and coaching</a>.</p><p><em>Originally published at </em><a href="https://thedatascientist.com/covid-19-ai-industry/"><em>https://thedatascientist.com</em></a><em> on May 28, 2020.</em></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=2f30a0925225" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[What will the AI economy really look like?]]></title>
            <link>https://stylianos-kampakis.medium.com/what-will-the-ai-economy-really-look-like-9532fd97eb1a?source=rss-db4ffe57dadc------2</link>
            <guid isPermaLink="false">https://medium.com/p/9532fd97eb1a</guid>
            <category><![CDATA[economics]]></category>
            <category><![CDATA[big-data]]></category>
            <category><![CDATA[data-science]]></category>
            <category><![CDATA[machine-learning]]></category>
            <category><![CDATA[artificial-intelligence]]></category>
            <dc:creator><![CDATA[Dr. Stylianos Kampakis]]></dc:creator>
            <pubDate>Wed, 27 May 2020 11:57:38 GMT</pubDate>
            <atom:updated>2020-06-02T11:32:35.295Z</atom:updated>
            <content:encoded><![CDATA[<p>By Dr Stylianos Kampakis</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/504/1*xVFs8Don__ztpaUwPK4N0Q.png" /></figure><h3>How is AI impacting the economy?</h3><p>Over the last few years, there has been lots of conversation around how AI is going to affect our lives, and the economy. A quick google search returns multiple studies. For example, <a href="https://www.pwc.co.uk/services/economics-policy/insights/the-impact-of-artificial-intelligence-on-the-uk-economy.html">PWC</a> says that the UK’s economy GDP will increase by at least 5% as the result of AI. A report by <a href="https://blogs.wsj.com/cio/2018/11/16/the-impact-of-artificial-intelligence-on-the-world-economy/">McKinsey</a> says that the annual productivity will be increasing by 1.2% each year. There are also some <a href="https://www.amazon.co.uk/Prediction-Machines-Economics-Artificial-Intelligence/dp/1633695670/ref=sr_1_1?adgrpid=48760788850&amp;gclid=Cj0KCQiAmsrxBRDaARIsANyiD1rdMmlyyZJ2xJ-QLFXaw1_wRRs3pr2q1hWf8Ji6EsRtS50L56IIvp0aApslEALw_wcB&amp;hvadid=259067772460&amp;hvdev=c&amp;hvlocphy=9045995&amp;hvnetw=g&amp;hvpos=1t1&amp;hvqmt=e&amp;hvrand=17267363312310990666&amp;hvtargid=aud-858007574774%3Akwd-446246170447&amp;hydadcr=18465_1772314&amp;keywords=prediction+machines&amp;qid=1580378101&amp;sr=8-1">books</a> that have started to come out on that subject. I am also talking about this topic in my upcoming book called <strong>Uncertainty</strong>. I had also done some of my own <a href="https://thedatascientist.com/ai-startups-disrupting-economy-work/">predictions</a> in the past.</p><p>However, while we hear many opinions about the subject, the truth is that <em>no-one really knows what is going to happen</em>. All data-related and AI technologies are characterised by some interesting features which makes it difficult to understand how they might affect a sector:</p><ol><li>They are <em>general purpose</em>. the same techniques that are used in autonomous vehicles, might be used in radiology for example (<a href="https://www.sciencedirect.com/science/article/pii/S0720048X19300919">here</a> is a good summary on radiology for anyone interesting in this subject).</li><li>They are becoming <em>easier and easier</em> to access. The big tech companies, like <a href="https://azure.microsoft.com/en-gb/services/machine-learning/">Microsoft</a> and <a href="https://cloud.google.com/ai-platform/docs/technical-overview">Google</a>, are releasing APIs for any possible use case.</li><li>They can become <em>better in doing what humans are already doing</em> (in some cases). For example, autonomous vehicles might completely replace all human drivers in the near future.</li><li>We are still trying to discover things and issues about current models. We don’t know what might come next. A good example of this are <strong>adversarial attacks on neural networks</strong>.</li></ol><figure><img alt="" src="https://cdn-images-1.medium.com/max/454/0*6zPKMOvhNkuuSJVA" /></figure><p><em>Example of an adversarial attack and how it works. Image taken from </em><a href="https://www.nature.com/articles/d41586-019-03013-5"><em>Nature</em></a><em>.</em></p><p>When you get all these four features combined, you can get some unpredictable consequences. Most predictions are focused on the third feature: <strong>automation</strong>. Autonomous vehicles is the most prominent example. Many companies are heavily investing in creating truly autonomous vehicles. Once the technology reaches a certain level of maturity, and society accepts it, then most professional drivers will simply be replaced. This is an obvious example.</p><p>However, the when we combine all four features, we can get <em>unpredictable consequences</em>. As technology becomes more accessible, we get higher probability that it will be used in a novel ways and problems. Remember, that the AI algorithms created, are usually general purpose. And they are becoming more and more accessible. This means that any disruptive applications might actually come, not from the creator of the algorithms, but from specialists in different areas, who just realise that they’ve found a new tool to use.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/549/0*HQ1rCDMMg91koWrD" /></figure><p>That being said, there are still things that we can do, in order to analyse the impact that AI will have on the economy. A very good article by <a href="https://thegradient.pub/the-economics-of-ai-today/">The Gradient</a>, summarised the work that was done in the <a href="https://www.economicsofai.com/nber-conference-toronto-2017">economics of AI conference in 2017</a>. In brief, there are 4 main themes (quoting the Gradient):</p><p>“</p><ol><li><strong><em>Macro View</em></strong><em>: Impact of AI on aggregate economic variables like productivity, employment or inequality</em></li><li><strong><em>Meso View</em></strong><em>: Impact of AI on individual sectors such as scientific research or regulation</em></li><li><strong><em>Micro View</em></strong><em>: Impact of AI on the behaviors of organizations and individuals</em></li><li><strong><em>Meta View</em></strong><em>: Impact of AI on the data and methods that economists use to study AI</em></li></ol><p>“</p><p>One of the interesting points coming from this article is that AI might actually have a negative effect on productivity in some cases. As jobs are being displaced, alternative jobs need to be found. If the alternative jobs are low productivity jobs, then the net impact could be negative. Also, there is a high risk of AI polarising the market.</p><p>It also looks like that the jobs that AI will create will either be highly technical jobs, or low-skilled jobs. An example of a technical job is a data scientist. An example, of a low-skilled job that has been created by AI, is using <a href="https://www.mturk.com/">mechanical Turk</a> to label images.</p><p>Another important point that is made, is that whether business will benefit from AI depends on <em>intangibles</em>, like culture and processes. I feel that this really validates my work which has touched upon many of those intangibles: from <a href="https://thedatascientist.com/creating-data-driven-culture/">building a data-driven culture</a>, to understanding <a href="https://thedatascientist.com/decision-makers-handbook-data-science/">data science on the organisational level</a>, to <a href="https://thedatascientist.com/data-science-courses/">courses and coaching</a>. In order to make sure that many industries can benefit from new technological developments, it is important to ensure that all businesses have access to the knowledge required to adopt these technologies.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/447/0*QEgraQYavQZXoPXT" /></figure><h3>So, how will AI affect the economy?</h3><p>So, now let’s try to answer the question that we set out in the beginning. I said that it is going to be difficult to predict the actual details as to how AI will affect the economy. However, we can map out some of the ways this will take place. We can map out some of our blind spots, and some of the features which this revolution will bring.</p><p>My opinion pretty much coincides with the opinion expressed in the <a href="https://thegradient.pub/the-economics-of-ai-today/">Gradient</a>.</p><blockquote><em>“The impact of AI on the economy (and society) will look more like the internet, rather than some kind of singularity event.”</em></blockquote><p>Think about the point I raised earlier, about AI being now easily accessible, and that soon anyone might be able to come up with ideas for data products. This is exactly what happened with the internet.</p><p>Let’s take commerce for example. Some retailers realised that they can sell online. So, they had both brick-and-mortar and online stores. Jeff Bezos realised that he can build a whole enterprise based only on selling online. A few of decades after the invention of the internet, anyone can create an e-shop, and promote it through social media channels, like Instagram. At the same time, these social media channels are also blamed for causing poor mental health and <a href="https://www.sciencealert.com/attention-deficit-hyperactivity-disorder-correlation-social-media-use-adolescents">attention deficit disorder</a>.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/451/0*HVaz6nSPOB1D-WHi" /></figure><p>Who would have predicted the rise of <strong>Amazon</strong> as an <strong>AI-first company</strong> that would dominate global commerce, as well as the rise of the solo lifestyle entrepreneur, promoting brands and products on some medium of communication that hadn’t even existed in the 90s? Probably no-one, and this is where we stand with predictions around the way through which AI affect the economy. We can’t make accurate predictions, but there are some pointers about where to look, like the ones suggested in the beginning of this article.</p><p>While automation is the obvious case, and most reports (like the one by <a href="https://www.pwc.co.uk/services/economics-policy/insights/the-impact-of-artificial-intelligence-on-the-uk-economy.html">PWC</a>) are focused on it, I believe that many of us will be surprised by applications of AI which are unforeseen and disruptive <a href="https://thedatascientist.com/implement-data-science-company/">business models</a>. It is also important, like explained in the previous section, to understand that many of the changes that will take place, might have counterintuitive effects. Countries that are fast to adopt <strong>AI algorithms</strong> might witness their GDP <em>growing</em>, but their <em>productivity getting worse</em>, as vast numbers of people end in low-productivity jobs.</p><p>One things that’s for certain is that the individuals and the companies that don’t educate themselves about the upcoming changes are going to be <em>left behind</em>. That’s why I keep running my <a href="https://thedatascientist.com/data-science-courses/">courses</a> and I am so happy to keep providing <a href="https://thedatascientist.com/data-science-services/">services</a> in this area. Because I believe that it is something that society really needs.</p><p><em>Originally published at </em><a href="https://thedatascientist.com/ai-economy-really-look-like/"><em>https://thedatascientist.com</em></a><em> on May 27, 2020.</em></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=9532fd97eb1a" width="1" height="1" alt="">]]></content:encoded>
        </item>
    </channel>
</rss>