<?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 Latha Samala on Medium]]></title>
        <description><![CDATA[Stories by Latha Samala on Medium]]></description>
        <link>https://medium.com/@lathasamala?source=rss-15fec25ce67d------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/0*ARe5u9ymedId9euQ.</url>
            <title>Stories by Latha Samala on Medium</title>
            <link>https://medium.com/@lathasamala?source=rss-15fec25ce67d------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Thu, 19 Nov 2026 08:53:06 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@lathasamala/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[Intelligent Automation]]></title>
            <link>https://lathasamala.medium.com/intelligent-automation-624c09ae45b9?source=rss-15fec25ce67d------2</link>
            <guid isPermaLink="false">https://medium.com/p/624c09ae45b9</guid>
            <dc:creator><![CDATA[Latha Samala]]></dc:creator>
            <pubDate>Wed, 05 May 2021 02:06:40 GMT</pubDate>
            <atom:updated>2024-03-23T22:02:22.701Z</atom:updated>
            <content:encoded><![CDATA[<h3>Intelligent Automation of Software</h3><p>Anticipation surrounded the prediction that self-driving cars would become a reality by 2020. However, as 2020 came and went, this expectation remained unfulfilled. In 2024, LLMs have significantly advanced automation capabilities, they still have limitations and may require human oversight. <strong>While automating business processes through software is relatively straightforward, determining what and how to automate remains a human-driven decision.</strong></p><p>Software applications now provide automated functionalities to end users, especially with the rise of platform and infrastructure services that abstract and automate operational tasks. However, despite these advancements, system maintenance still requires expert knowledge.</p><p>Creating a user-friendly and self-maintaining software system involves considering several key aspects that enhance usability, efficiency, and autonomy. Let’s explore how a system progresses with levels of automation:</p><p><strong>Level1-Manual: </strong>All work is done manually, leveraging technology but without full automation.<br><strong>Level2-Attended Automation: </strong>Introduction of helper bots to assist in tasks.<br><strong>Level3-Batch Automation: </strong>Industrialization of automation processes.<br><strong>Level4-Unattended Automation: </strong>Automation starts managing tasks independently.<br><strong>Level5-Intelligent Automation: </strong>Automation begins actively learning and adapting.<br><strong>Level6-Fully Automated: </strong>A future state where systems operate autonomously.</p><h3>System1-Car: User experience of Driving Car</h3><blockquote><strong>Level1-Manual: </strong>Traditional cars are operated manually despite using automotive technology.</blockquote><blockquote><strong>Level2-Attended Automation: </strong>Cruise control and lane departure warnings in cars.</blockquote><blockquote><strong>Level3-Batch Automation: </strong>Automated parking in cars.</blockquote><blockquote><strong>Level4-Unattended Automation: </strong>Automated highway driving in cars.</blockquote><blockquote><strong>Level5-Intelligent Automation: </strong>Tesla’s advanced automation features in cars.</blockquote><ul><li>Cars of the future won’t have steering wheels. Environment and the user are perceived by the car.</li><li><strong>LLMs</strong> can be applied in various ways to enhance the user experience of driving a car:</li><li><strong>Voice Assistants: </strong>Integrate LLM-powered voice assistants into car systems for natural language interactions, such as controlling features, obtaining directions, and accessing real-time information while driving.</li><li><strong>Personalized Recommendations:</strong> Utilize LLMs to analyze user driving patterns and preferences, then provide personalized route suggestions, music playlists, and nearby service recommendations tailored to the individual driver.</li><li><strong>Adaptive User Interfaces: </strong>Develop interfaces using LLMs to understand user intents and dynamically adjust dashboard layouts, settings, and notifications based on real-time driving conditions and user preferences.</li><li><strong>Natural Language Understanding (NLU):</strong> Implement LLM-based NLU systems to interpret and respond to spoken commands, queries, and feedback from drivers, enhancing hands-free interaction and overall driving experience.</li></ul><h3>System2-End User Functions for Computer Vision Systems: Classify an image.</h3><blockquote><strong>Level 1 — Manual:</strong> Operators manually tag each acquired image using Software Systems.</blockquote><blockquote><strong>Level 2 — Attended Automation:</strong> Algorithms are trained with specific rules, such as matching a particular logo in an image. Operators intervene only for images that don’t match exactly using Software Systems.</blockquote><blockquote><strong>Level 3 — Batch Automation:</strong> Algorithms are trained using examples to detect patterns in computer vision systems.</blockquote><blockquote><strong>Level 4 — Unattended Automation:</strong> Algorithms undergo training with diverse data variations for pattern detection in computer vision systems.</blockquote><blockquote><strong>Level 5 — Intelligent Automation:</strong> The system engages in active learning and noise handling within computer vision systems.</blockquote><ul><li><strong>LLMs</strong> can be applied to enhance the image classification capabilities of computer vision systems in the following ways:</li><li><strong>Fine-Tuning Models: </strong>Use LLMs to fine-tune image classification models for specific tasks or domains, improving accuracy and performance in handling diverse image classifications.</li><li><strong>Natural Language Queries:</strong> Enable users to describe image classification needs using natural language, and leverage LLMs to convert these descriptions into actionable tasks for the system, enhancing user accessibility and usability.</li><li><strong>Semantic Image Search:</strong> Develop semantic search capabilities powered by LLMs to understand user queries about image features, objects, or attributes, and retrieve relevant classified images from databases efficiently.</li><li><strong>Automated Annotation: </strong>Utilize LLMs to automate image annotation processes by generating descriptive labels and metadata for classified images, streamlining data organization and retrieval in computer vision systems.</li></ul><h3>System3-System Functions for Computer Vision Systems (or any system): Elasticity for hardware environments, data, and functional configurations.</h3><blockquote><strong>Level 1 — Manual:</strong> Developers manually review logs generated by disparate systems, identify errors, and seek solutions from external sources for resolution.</blockquote><blockquote><strong>Level 2 — Attended Automation:</strong> Logs, metrics, and events from all systems are centralized for easy access. Contextual help is provided alongside these events to assist developers in troubleshooting.</blockquote><blockquote><strong>Level 3 — Batch Automation:</strong> Centralized logs, metrics, and events are accompanied by contextual application data, aiding in understanding the events comprehensively. Suggestions for manual adjustments to the system and data are also provided.</blockquote><blockquote><strong>Level 4 — Unattended Automation:</strong> Similar to Level 3, but with the added capability of automatically adjusting data or system errors before suggesting manual interventions. This streamlines the resolution process.</blockquote><blockquote><strong>Level 5 — Intelligent Automation:</strong> Comprehensive error and report viewing is integrated across all system components. Contextual help and application data provide deep insights into system events. Automated adjustments are attempted with a learning component, allowing the system to improve based on resolved errors.</blockquote><ul><li><strong>LLMs</strong> can contribute to achieving elasticity for hardware environments, data, and functional configurations in computer vision systems (or any system) in the following ways:</li><li><strong>Natural Language Understanding for Requirements Gathering: </strong>Build conversational interfaces (chatbots, virtual assistants) for users to interact with the platform using natural language. Generate contextualized help content, troubleshooting guides, and recommendations based on user queries and system data. Capture user feedback and system performance data to identify areas for improvement</li><li><strong>Predictive Maintenance and Monitoring: </strong>LLMs can analyze system logs, metrics, and performance data to predict potential failures, anomalies, or performance degradation in software systems. LLMs can recommend preventive maintenance actions, system optimizations, or scaling strategies based on predictive analytics, improving system reliability and uptime.</li></ul><p><strong>Context</strong> is User, Feature, Value of the feature at that instance for the user.</p><p><strong>In conclusion</strong>, achieving a fully automated system requires understanding and adapting to user context, predicting needs, and allowing for user-configurable adjustments. This self-adaptive approach enhances user experience and system performance across various contexts.</p><p>Reference:</p><p><a href="https://www.linkedin.com/pulse/6-levels-automation-remastered-alex-lyashok/">https://www.linkedin.com/pulse/6-levels-automation-remastered-alex-lyashok/</a></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=624c09ae45b9" width="1" height="1" alt="">]]></content:encoded>
        </item>
    </channel>
</rss>