<?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[ImpactAI - Medium]]></title>
        <description><![CDATA[Using AI where it’s most needed - Medium]]></description>
        <link>https://medium.com/impactai?source=rss----46de92b127fe---4</link>
        <image>
            <url>https://cdn-images-1.medium.com/proxy/1*TGH72Nnw24QL3iV9IOm4VA.png</url>
            <title>ImpactAI - Medium</title>
            <link>https://medium.com/impactai?source=rss----46de92b127fe---4</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Wed, 18 Nov 2026 22:19:32 GMT</lastBuildDate>
        <atom:link href="https://medium.com/feed/impactai" rel="self" type="application/rss+xml"/>
        <webMaster><![CDATA[yourfriends@medium.com]]></webMaster>
        <atom:link href="http://medium.superfeedr.com/" rel="hub"/>
        <item>
            <title><![CDATA[Impact Driven AI applications — Session 2 at USF]]></title>
            <link>https://medium.com/impactai/impact-driven-ai-applications-session-2-at-usf-586d7624abb1?source=rss----46de92b127fe---4</link>
            <guid isPermaLink="false">https://medium.com/p/586d7624abb1</guid>
            <category><![CDATA[deep-learning]]></category>
            <category><![CDATA[social-impact]]></category>
            <category><![CDATA[machine-learning]]></category>
            <category><![CDATA[artificial-intelligence]]></category>
            <dc:creator><![CDATA[Sravya Tirukkovalur]]></dc:creator>
            <pubDate>Fri, 02 Jun 2017 03:28:00 GMT</pubDate>
            <atom:updated>2017-06-15T18:45:52.189Z</atom:updated>
            <content:encoded><![CDATA[<p>By <a href="http://fast.ai/">fast.ai</a>/<a href="https://www.usfca.edu/data-institute">Data Institute</a> students</p><p>It was great to see enthusiastic audience at last week’s <a href="https://www.meetup.com/USF-Seminar-Series-in-Analytics/events/239243015/">USF seminar — the second part of the Impact Driven AI applications</a> driven by the Data Institute/ Fast.ai students. We had a hall filled with about 160 participants! This session is part of the series which started as a discussion among a small group of students during the advanced Deep learning course offered by fast.ai earlier this year.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/828/0*ob3UDzd2LTHijSEX." /><figcaption>Discussion thread on the fast.ai forums in February</figcaption></figure><p>It is exciting to see the community growing and working together in the pursuit of leveraging Deep Learning in not just shiny cool projects but the ones which drive tremendous value around the globe by solving important and immediate problems. And we are thrilled to announce we are taking this forward with ImpactAI.org, a hands-on think tank to explore new domains and build AI driven products where they are most needed. Please watch the video to learn more about the journey so far. This is an open call for all AI/ ML practitioners who are interested in driving tremendous social value. Please reach out at sravya[at]impactai.org or anjana[at]impactai.org</p><p>I would like to take this moment to acknowledge and thank all collaborators who played an important role in making this possible: <a href="https://medium.com/u/e0622e1e83b0">Janardhan Shetty</a>, <a href="https://medium.com/u/8b897fafe57a">Anjana Prabhakar</a>, <a href="https://medium.com/u/61d9e974ecf9">Matthew Kleinsmith</a>, <a href="https://medium.com/u/34ab754f8c5e">Jeremy Howard</a>, <a href="https://medium.com/u/ee56d0bac1b7">Rachel Thomas</a>, USF — Data Institute, <a href="https://medium.com/u/a410f199331">Xinxin Li</a>, <a href="https://medium.com/u/cb3304cd3a06">Samar Haider</a>, <a href="https://medium.com/u/8e2a2d2179bf">Sara Hooker</a>, Brookie Guzder, <a href="https://medium.com/u/4d574660dab8">Yannet Interian</a>,<strong> </strong>Kirsten Keihl,<strong> </strong>Mindi Mysliwiec, Leslie Blakeman</p><p><strong>Brief summary of the talks, slide decks and video</strong></p><p>Sravya Tirukkovalur shares details about the first project at ImpactAI.org, where the team is using satellite imagery to assess structural damage after an earthquake, speeding up a critical piece in the relief coordination.</p><p>Brookie Guzder-Williams talks about “Forma 250: Near real-time detection of tree cover loss through satellite imagery”</p><p>Xinxin works with Xeed, a wearable technology company, to develop a system for Parkinson’s patient therapy management. She speaks about her work on applying deep learning to Parkinson’s disease therapeutic research, the core of which is a machine learning model trained with clinical studies data. This new system would enable a doctor to gauge patients’ symptoms, such as tremors and dyskinetic, via sensor data collected out of the clinic, rather than relying on written diaries or interviews with patient caregivers.</p><iframe src="https://cdn.embedly.com/widgets/media.html?src=https%3A%2F%2Fwww.slideshare.net%2Fslideshow%2Fembed_code%2Fkey%2F8e6Khabi28wMhG&amp;url=https%3A%2F%2Fwww.slideshare.net%2FSravyaTirukkovalur%2Fimpact-driven-ai-applications-session-2-at-usf&amp;image=https%3A%2F%2Fcdn.slidesharecdn.com%2Fss_thumbnails%2Fimpactdrivenai052617usf-170601180758-thumbnail-4.jpg%3Fcb%3D1496340602&amp;key=d04bfffea46d4aeda930ec88cc64b87c&amp;type=text%2Fhtml&amp;schema=slideshare" width="600" height="500" frameborder="0" scrolling="no"><a href="https://medium.com/media/f469d5aab4568e9f67864e8323ff2a72/href">https://medium.com/media/f469d5aab4568e9f67864e8323ff2a72/href</a></iframe><p><a href="https://www.facebook.com/plugins/video.php?href=https%3A%2F%2Fwww.facebook.com%2Fusfca.analytics%2Fvideos%2F441338442906269">Facebook</a></p><p>Update:</p><p>We checked after the talk and the statistic on number of radiologists mentioned in the talk is inaccurate. Apologies.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=586d7624abb1" width="1" height="1" alt=""><hr><p><a href="https://medium.com/impactai/impact-driven-ai-applications-session-2-at-usf-586d7624abb1">Impact Driven AI applications — Session 2 at USF</a> was originally published in <a href="https://medium.com/impactai">ImpactAI</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Thanks for the note Horea!]]></title>
            <link>https://medium.com/impactai/thanks-for-the-note-horea-f5b1818c02ae?source=rss----46de92b127fe---4</link>
            <guid isPermaLink="false">https://medium.com/p/f5b1818c02ae</guid>
            <dc:creator><![CDATA[Sravya Tirukkovalur]]></dc:creator>
            <pubDate>Wed, 31 May 2017 02:37:31 GMT</pubDate>
            <atom:updated>2017-05-31T02:37:30.830Z</atom:updated>
            <content:encoded><![CDATA[<p>Thanks for the note Horea! Glad you enjoyed the talk! You can reach out to me at sravya[at]impactai.org</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=f5b1818c02ae" width="1" height="1" alt=""><hr><p><a href="https://medium.com/impactai/thanks-for-the-note-horea-f5b1818c02ae">Thanks for the note Horea!</a> was originally published in <a href="https://medium.com/impactai">ImpactAI</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[CNNs from different viewpoints]]></title>
            <link>https://medium.com/impactai/cnns-from-different-viewpoints-fab7f52d159c?source=rss----46de92b127fe---4</link>
            <guid isPermaLink="false">https://medium.com/p/fab7f52d159c</guid>
            <category><![CDATA[neural-networks]]></category>
            <category><![CDATA[machine-learning]]></category>
            <category><![CDATA[deep-learning]]></category>
            <category><![CDATA[artificial-intelligence]]></category>
            <dc:creator><![CDATA[Matt Kleinsmith]]></dc:creator>
            <pubDate>Thu, 20 Apr 2017 16:11:13 GMT</pubDate>
            <atom:updated>2017-04-20T16:11:12.572Z</atom:updated>
            <content:encoded><![CDATA[<h4>Prerequisite: Basic neural networks</h4><h4>The theme of this post</h4><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*4UPp3Tc4A32S0WXJ0pWP7g.jpeg" /></figure><h4>The<strong> image</strong></h4><figure><img alt="" src="https://cdn-images-1.medium.com/max/75/1*7zVtk9Xb3X6aQScJqxCUYw.png" /></figure><h4>The filter</h4><figure><img alt="" src="https://cdn-images-1.medium.com/max/51/1*aCBPCSInHE6vFaECSAT5oA.png" /></figure><h4>Since the filter fits in the image four times, we have four results</h4><figure><img alt="" src="https://cdn-images-1.medium.com/max/53/1*I8BERkXPnVjdVCawr7Q23A.png" /></figure><h4>Here’s how we applied the filter to each section of the image to yield each result</h4><figure><img alt="" src="https://cdn-images-1.medium.com/max/362/1*s-TmnXo69mAKbnMoMBnpMg.png" /></figure><h4>The equation view</h4><figure><img alt="" src="https://cdn-images-1.medium.com/max/428/1*-TnNZ4srYItcGXov-dTmmw.png" /></figure><p>Notice that the bias term, b, is the same for each section of the image. You can consider the bias as part of the filter, just like the weights (α, β, γ, δ) are part of the filter.</p><h4>The compact equation view</h4><figure><img alt="" src="https://cdn-images-1.medium.com/max/216/1*3IbCXOa7zllJHhJrvueG9Q.png" /></figure><h4>The neural network view</h4><figure><img alt="" src="https://cdn-images-1.medium.com/max/363/1*Pirh_aXJwGLbOsENeMNHrA.png" /></figure><h4>The matrix multiplication view</h4><figure><img alt="" src="https://cdn-images-1.medium.com/max/716/1*95lL-PY5WEeBAtfaWAIIRQ.png" /></figure><p>The matrix above is a weight matrix, just like the ones from traditional neural networks. However, this weight matrix has two special properties:</p><ol><li>The zeros shown in gray are untrainable. This means that they’ll stay zero throughout the optimization process.</li><li>Some of the weights are equal, and while they are trainable (i.e. changeable), they must remain equal. These are called “shared weights”.</li></ol><p>The zeros correspond to the pixels that the filter didn’t touch. Each row of the weight matrix corresponds to one application of the filter.</p><h4>The dense neural network view</h4><figure><img alt="" src="https://cdn-images-1.medium.com/max/354/1*jrGLWytdCPiVjrU8i4XHPg.png" /></figure><p>The gray connections correspond to the untrainable zeros. This graph is the same as the previous graph, except that it shows the untrainable zeros. This view helped me see the connection between traditional neural networks and CNNs.</p><h4>A familiar diagram</h4><figure><img alt="" src="https://cdn-images-1.medium.com/max/355/1*jGIAbxQ_jlPO7gi8xnDOxg.png" /></figure><h4>How did I learn this?</h4><ol><li><a href="http://course.fast.ai/">Jeremy Howard’s deep learning course</a>, especially <a href="https://github.com/fastai/courses/blob/master/deeplearning1/excel/conv-example.xlsx">his CNN spreadsheet</a> and <a href="https://www.youtube.com/watch?v=V2h3IOBDvrA">the corresponding video</a> (0:00 to 14:00)</li><li><a href="http://neuralnetworksanddeeplearning.com/">Michael Nielsen’s deep learning book</a>, especially <a href="http://neuralnetworksanddeeplearning.com/chap1.html">chapter 1 (NNs)</a> and <a href="http://neuralnetworksanddeeplearning.com/chap6.html">chapter 6 (CNNs)</a></li></ol><p>I found out about Michael Nielsen’s book from <a href="http://wiki.fast.ai/index.php/Lesson_2">Jeremy’s suggested readings for NNs</a> and <a href="http://wiki.fast.ai/index.php/Lesson_3#CNN_review">readings for CNNs</a>.</p><h4>Bonus: Preserving the image size with zero-padding and a 3x3 filter</h4><p>Notice that we went from a 3x3 image to a 2x2 image.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/232/1*X904PMeB5iq-r8sfRpf2LA.png" /></figure><p>With zero-padding and a 3x3 filter, we can preserve the image size.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/254/1*d2zKv2LzHmmgWXK-MDFewQ.png" /></figure><p>Here’s what zero-padding with a 3x3 filter looks like:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/451/1*m8oHuCZMBK6gghLXr1bY8Q.png" /></figure><p>The 3x3 filter allows us to center the filter on each pixel, so that each original pixel corresponds to a pixel in the result. Although, without zero-padding, our filter would get stuck in the corners, unable to center on any pixel except pixel E, and the result would be a 1x1 image.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/430/1*wIDWlaOiGQQBqBtIObPobg.png" /></figure><h4>Important things that I didn’t talk about or diagram</h4><ul><li>Activation functions</li><li>Pooling</li><li>Multiple filters</li><li>Multiple convolutional layers</li><li>Dense layers (without shared weights or untrainable zeros)</li><li>The output layer</li><li>The optimization process (i.e. gradient descent via backpropagation)</li><li><a href="http://course.fast.ai/lessons/lesson1.html">How to do useful things with CNNs</a></li></ul><h4>Important views I didn’t show</h4><ul><li>Code views (e.g. a Python view or a Keras view)</li><li><a href="http://setosa.io/ev/image-kernels/">Realistic views</a> (i.e. with real images)</li><li>Many-filter views (<a href="http://cs231n.github.io/understanding-cnn/">here</a> and <a href="https://i.stack.imgur.com/Hl2H6.png">here</a>)</li><li>Interactive views (<a href="https://github.com/fastai/courses/blob/master/deeplearning1/excel/conv-example.xlsx">here</a> and <a href="http://setosa.io/ev/image-kernels/">here</a>)</li></ul><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=fab7f52d159c" width="1" height="1" alt=""><hr><p><a href="https://medium.com/impactai/cnns-from-different-viewpoints-fab7f52d159c">CNNs from different viewpoints</a> was originally published in <a href="https://medium.com/impactai">ImpactAI</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Impact driven AI applications — Fast.ai students]]></title>
            <link>https://medium.com/impactai/lightning-talks-impact-driven-ai-applications-fb5836d875f7?source=rss----46de92b127fe---4</link>
            <guid isPermaLink="false">https://medium.com/p/fb5836d875f7</guid>
            <category><![CDATA[machine-learning]]></category>
            <category><![CDATA[social-enterprise]]></category>
            <category><![CDATA[artificial-intelligence]]></category>
            <category><![CDATA[social-impact]]></category>
            <category><![CDATA[deep-learning]]></category>
            <dc:creator><![CDATA[Sravya Tirukkovalur]]></dc:creator>
            <pubDate>Thu, 20 Apr 2017 12:20:19 GMT</pubDate>
            <atom:updated>2017-06-06T18:42:02.646Z</atom:updated>
            <content:encoded><![CDATA[<p>A collaboration of mission-driven deep learning practitioners.</p><p>Co-written by <a href="https://medium.com/u/e0622e1e83b0">Janardhan Shetty</a></p><blockquote>“Hungry to learn and Driven by impact community”</blockquote><p>It was not just another Friday evening here at the University of San Francisco. But it is one where “Hungry to learn and Driven by impact community” of Deep learning practitioners came together to discuss various domains, what are some high impact problems solvable through AI and how do we solve them. I only peripherally mentioned in my <a href="https://medium.com/@sravsatuluri/millennials-here-are-three-ways-to-be-effective-and-creative-in-your-philanthropic-endeavors-d2bdfcbfa4ce">last post</a> about why I think Deep learning could be an excellent tool for impact. Now, I am even more convinced that this tool can very well be the greatest digital tool of this generation for impact.</p><p>In this article, we give you a glimpse of what deep learning student community of <a href="http://fast.ai/">fast.ai</a> is passionate about and the talks are classified into domain areas. Here is the <a href="https://www.youtube.com/watch?v=NwhJHzu5BYU">link</a> for the recorded talk. Please watch this space as we will be going deeper into some of these applications in the future blogs.</p><h4><strong>Medical diagnosis</strong></h4><p><a href="https://medium.com/u/e0622e1e83b0">Janardhan Shetty</a> spoke about early detection tool for Autism. Early detection of autism in kids saves ⅔ of the lifetime costs for their care. He explained how Deep Learning can be helpful to identify facial features of autistic kids, characteristic voice patterns through audio clips, behavioral video patterns and movements through video clips.</p><p><a href="https://medium.com/u/34ab754f8c5e">Jeremy Howard</a> candidly shared how his professional journey led to an impact journey and how he started <a href="http://www.enlitic.com/">Enlitic</a>, a company which uses Deep learning to augment doctors in medical diagnosis. <strong>Ljubomir Buturovic</strong> shared how his company <a href="http://clinicalpersona.com/">ClinicalPersona</a> is working on detection of breast cancer. In both these cases, medical imagery is used to detect and classify cancerous tissues.</p><p><strong>Agritech</strong></p><p><a href="https://medium.com/u/f5ed18371985">Sahil Singla</a> spoke about how <a href="http://farmguide.in/">FarmGuide</a> is digitizing and solving the information asymmetry problem in Agriculture, the biggest yet loss-making industry in India. It was fascinating to learn they actually scraped Google Earth and segmented individual farms which were essential to help farmers get loans by increasing transparency. It is amazing to see the use of advanced technology in one of the most neglected yet important areas.</p><p><strong>Deforestation</strong></p><p>My friend <a href="https://medium.com/u/8e2a2d2179bf">Sara Hooker</a>, spoke about <a href="http://www.deltanalytics.org/">Delta Analytics</a>, a non-profit she co-founded, which provided 15000 pro-bono hours to nonprofits all over the world over a period of just three years. She is now co-leading a project with <a href="https://rfcx.org/">Rain Forest Connection</a>, to cut Deforestation, a leading cause of climate change. They are building models to detect the chainsaw noise captured from a continuous audio stream sent by recycled solar powered cell phone tied to trees. She also shared how they are working around with so little ground truth data and mentioned other logistic challenges as being able to detect the direction of the noise.</p><p><strong>Disaster response and resilience</strong></p><p>I, <a href="https://medium.com/u/69a655fa6c1e">Sravya Tirukkovalur</a>, have been following the work done by <a href="http://www.digital-humanitarians.com/">Digital Humanitarians</a> during past natural disasters around the world and was very inspired by the power of crowd-sourcing and use of technology to tackle immediate problems. I spoke about my experience researching this space and proposed some paths where Deep learning can really move the needle including quickly assessing structural damage using satellite imagery (which currently takes a few days on average) and intent classification of social media and text help-line content to fast track critical information to on-field aid workers.</p><p><strong>Education and accessibility to under-served populations</strong></p><p><a href="https://medium.com/u/cb3304cd3a06">Samar Haider</a> shared how he created first embeddings for 175K Urdu words, which is key in many NLP applications. He explained how he started solving the chicken and egg problem of language Corpus &lt;-&gt; knowledge by doing unsupervised training on 150 million tokens from Wiki and Urdu corpora.</p><p><a href="https://medium.com/u/ee56d0bac1b7">Rachel Thomas</a> shared her story on how uncomfortable she felt with disparities in access to technology and knowledge which led her to co-found fast.ai to make Deep learning really accessible to all populations. The people who are intimately familiar with the problems are often best equipped to solve it. And the key contribution of fast.ai, which I strongly believe is to give those populations the powers to use this incredible technology to solve their problems even faster!</p><p><strong>Where do we go from here?</strong></p><p>Above projects fall in various stages of development. Some of the above are already producing state of the art results and are validated to have a huge impact and there are some which are just getting started with ambitious goals.</p><p>As with any successful software project, it is important for us to start small and iterate to make sure we are asking the right questions, solving actual problems and having a real impact. And these are hard problems which can benefit from an army of passionate individuals, especially domain and tech experts. So collaborations are very welcome!</p><p>It is important to note that many of us had little machine learning background when we started the course. Deep learning, what started as a black box and a tough task to conquer, turned out to be not that hard. The results achieved by students are phenomenal. It’s just an inspiring journey and a great validation that anyone can get started irrespective of their backgrounds as long as they are passionate to keep at it.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=fb5836d875f7" width="1" height="1" alt=""><hr><p><a href="https://medium.com/impactai/lightning-talks-impact-driven-ai-applications-fb5836d875f7">Impact driven AI applications — Fast.ai students</a> was originally published in <a href="https://medium.com/impactai">ImpactAI</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Setting up a Deep learning machine in a lazy yet quick way]]></title>
            <link>https://medium.com/impactai/setting-up-a-deep-learning-machine-in-a-lazy-yet-quick-way-be2642318850?source=rss----46de92b127fe---4</link>
            <guid isPermaLink="false">https://medium.com/p/be2642318850</guid>
            <category><![CDATA[toddlers]]></category>
            <category><![CDATA[laziness]]></category>
            <category><![CDATA[deep-learning]]></category>
            <category><![CDATA[gpu]]></category>
            <category><![CDATA[ai]]></category>
            <dc:creator><![CDATA[Sravya Tirukkovalur]]></dc:creator>
            <pubDate>Thu, 20 Apr 2017 12:19:56 GMT</pubDate>
            <atom:updated>2018-01-08T22:01:51.979Z</atom:updated>
            <content:encoded><![CDATA[<p>How about do it today?</p><p>November of last year, I took the <a href="http://course.fast.ai/">“Deep learning for coders — Part 1”</a> course and since then I have been trying out Deep networks for a variety of problems areas like Computer vision and NLP among others. I got a p2 EC2 instance ($0.90 / hr for 1 K80 GPU) and was running all my experiments there. In the process, I spent way more than I was expecting on AWS.</p><p>If I was very vigilant in using my instance just when I am running on GPU, I would have saved good amount of money. But that was not always possible without compromising on productivity or flexibility. Also K80s are not that great and realized I should invest in a better GPU to accelerate experimentation and reduce overall costs.</p><p>Although I knew I wanted to upgrade and started researching about graphics cards, parts and configurations, it took me a while before I have this beauty setup in our home.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*AD2ze8uhHRx7H0F2JH6RIw.jpeg" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*hbwqt28A4lLpw3VVTCEbtg.jpeg" /></figure><p>There are some good GTX 10 series laptops, but I decided to go with a desktop for the sake of upgradability. And I am super glad I did, as there is a new card <a href="https://www.nvidia.com/en-us/geforce/products/10series/geforce-gtx-1080-ti/">GTX 1080Ti</a> which is released exactly a week after I setup this system. So now, I can add that as my second card :-)</p><p>So if you are in the position I was where you are on the edge of building your own machine but a little unsure of time that you need to invest in getting that setup, this post is for you. And this is inspired by my fellow students, mainly <a href="https://medium.com/u/40f65e98572">Yad Faeq</a>, <a href="https://medium.com/u/b51c67c513a3">Brendan Fortuner</a> and our professor <a href="https://medium.com/u/34ab754f8c5e">Jeremy Howard</a>.</p><h3>Instructions</h3><figure><img alt="" src="https://cdn-images-1.medium.com/max/343/1*_VHJkuUvSKkEq0LTnXuiYw.png" /></figure><p><a href="https://medium.com/u/b51c67c513a3">Brendan Fortuner</a>, my classmate has written a very accessible and excellent <a href="https://medium.com/@bfortuner/building-your-own-deep-learning-box-47b918aea1eb#.r4hjq2yli">blog </a>post on building your own Deep learning machine which includes 6 key steps as outlined to the left. So stop here and skim through it now, you will need to go back to that blog post when you are doing the actual setup. I will only be covering some additional perspectives here.</p><h3>Some tips to speed up the process in a lazy way :-)</h3><h4>On choosing parts</h4><p>Start from Brendan’s <a href="https://pcpartpicker.com/list/FRp8XH">list</a>, trust me, it’s good. Some things to keep in mind when you customize the list based on your budget.</p><ol><li>Spend most of your money on GPU in this order GTX 1080Ti -&gt; GTX1080 -&gt; GTX1070</li><li>Get good RAM, I got 32GB and more room to expand</li><li>Get SSD, accounts for so much productivity boost.</li><li>Get a motherboard which supports multiple GPUs, so that you can upgrade in future.</li></ol><h4>Hardware assembly</h4><p>You have two options here:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*IIt_oou-5GXJAOOSQtrLjg.jpeg" /><figcaption>Our toddler “helping out” with the setup</figcaption></figure><ol><li>Get your system assembled at a place like <a href="https://centralcomputers.com/">Central computers</a> for like 80$, they also do a burn test which can save you some pain if your parts come broken. I used this service as I figured assembling it myself with a toddler at home would be a lot more fun than I would want it to be. The service was great and on time. Got great recommendations for motherboards which support multiple graphic cards, PSU required for your current and future parts. They install required drivers. They even install Windows, but I wanted Ubuntu so I did that part myself. A bonus, I liked their case selection :-)</li><li>If you are on a tight budget, buy a used Dell workstation. Order a GPU online and you can plug it in yourself.</li></ol><h3>Setup workflow and test within your LAN</h3><figure><img alt="" src="https://cdn-images-1.medium.com/max/486/1*PhRO63lzeQPlkNQQ54WQFw.jpeg" /><figcaption>Src: <a href="https://www.qnap.com/images/products/Application/TS109AP05.htm">https://www.qnap.com/images/products/Application/TS109AP05.htm</a></figcaption></figure><p>Assuming you would want to not always work from home and would like to access your machine remotely from your laptop, you will need to setup remote workflow. This picture gives a brief overview of WAN and LAN and what needs to happen. But first, I would recommend setting up your workflow within LAN to test everything works within LAN.</p><ol><li>Make sure you are able to ssh within your LAN.</li><li>I mostly use terminal and Jupyter notebooks for coding. If you have a simialar workflow, make sure you are able to run Jupyter workflow from your laptop</li></ol><pre>Server:  jupyter notebook — no-browser — port=8888<br>Laptop: Put this in your ~/.bash_profile using your ACCOUNT@LAN_IP<br>___________________<br>desktop_ip=”ACCOUNT@LAN_IP”<br>remote_jupyter() {<br>ssh -NL $1:localhost:$1 $desktop_ip<br>}<br>___________________<br>Laptop: $remote_jupyter 8888</pre><h4>Password-less ssh</h4><p>Although this step is not required, I would hate to type my long password every time I connect, and also I have better trust on strong SSH keys than my passwords.</p><p>Follow this if you already have ssh setup on your laptop.</p><pre>a@A:~&gt; ssh b@B mkdir -p .ssh<br>b@B&#39;s password:<br>a@A:~&gt; cat .ssh/id_rsa.pub | ssh b@B &#39;cat &gt;&gt; .ssh/authorized_keys&#39;<br>b@B&#39;s password:</pre><p>If not follow instructions <a href="http://www.linuxproblem.org/art_9.html">here</a> to setup your keys. And disable password ssh login on the server for better security unless you plan to share your work station with others. Update the following parameters in your /etc/ssh/sshd_config</p><pre>ChallengeResponseAuthentication no<br>PasswordAuthentication no<br>UsePAM no</pre><p>and restart SSH deamon/etc/init.d/sshd restart</p><h4>Reserve LAN IP in your DHCP</h4><p>Internal IPs are also dynamically allocated, so to avoid IP conflict, reserve your internal IP in your DHCP settings, which would essentially bind one of your LAN IPs to the MAC of your machine. See bullet 2 below for how to do it.</p><h3>Make it work from outside of LAN</h3><ol><li>Open ssh port for access from outside LAN : We use TP-Link router at home, its Admin page is at 192.168.0.1 with default credentials: admin/admin. You can check instructions for your specific router <a href="https://portforward.com/router.htm">here</a>. Once you get to the admin console, they all look pretty similar. All that you need is to open up port 22 for ssh.</li><li>Address Reservation: DHCP -&gt; Address reservation and bind MAC to an internal IP</li><li>Update LAN_IP with WAN_IP in your ~/.bash_profile . To get your WAN_IP do $dig +short myip.opendns.com @resolver1.opendns.com</li></ol><p>Finally, test your work flow from outside of LAN. Simple way is to connect your laptop to your phone hotspot (so that it is not on LAN), and test the workflow. And my models train in half the time now with just one GTX 1080 compared to AWS p2 instance. Yayy!</p><p>So far, my WAN IP has not changed, so I decided to not worry about static IP yet. Especially since I do not have any production models deployed on my home machine and only occasionally use it remotely. Although I think, I will look into static ip options like <a href="http://www.noip.com/">noip</a> if needed.</p><p>Would love to hear if you have setup a similar machine. And please share your favorite productivity tips in these kind of setups.</p><p><em>Update</em>: I got requests asking for list of components used, so attaching it here for reference. Although, please note that at the rate at things are moving, there might be better bets now.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*mcSXmqZ5hg9QhRIzzv75lA.jpeg" /></figure><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=be2642318850" width="1" height="1" alt=""><hr><p><a href="https://medium.com/impactai/setting-up-a-deep-learning-machine-in-a-lazy-yet-quick-way-be2642318850">Setting up a Deep learning machine in a lazy yet quick way</a> was originally published in <a href="https://medium.com/impactai">ImpactAI</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Millennials — Here are three ways to be effective (and creative) in your philanthropic endeavors]]></title>
            <link>https://medium.com/impactai/millennials-here-are-three-ways-to-be-effective-and-creative-in-your-philanthropic-endeavors-d2bdfcbfa4ce?source=rss----46de92b127fe---4</link>
            <guid isPermaLink="false">https://medium.com/p/d2bdfcbfa4ce</guid>
            <category><![CDATA[tech-for-good]]></category>
            <category><![CDATA[effective-altruism]]></category>
            <category><![CDATA[impact-investing]]></category>
            <category><![CDATA[philanthropy]]></category>
            <category><![CDATA[social-entrepreneurship]]></category>
            <dc:creator><![CDATA[Sravya Tirukkovalur]]></dc:creator>
            <pubDate>Thu, 20 Apr 2017 12:18:26 GMT</pubDate>
            <atom:updated>2016-12-21T20:56:30.508Z</atom:updated>
            <content:encoded><![CDATA[<h3>Here are three ways to be effective (and creative) in your philanthropic endeavors</h3><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*MEv9UT8MDAJWpZvVWynUxw.png" /></figure><p>Around this time last year, I was experiencing some of the most beautiful moments in my life cuddling with our baby boy. I used to think the greatest unconditional love is what moms have for their kids until I had a baby. No matter how much I love my baby, he makes me feel that he loves me more. So, I guess it’s an infinite loop of positive reinforcement. As I was getting used to the new found love, one day, I came across <a href="https://www.theguardian.com/world/2015/sep/02/shocking-image-of-drowned-syrian-boy-shows-tragic-plight-of-refugees">this</a> image of a 2-year-old boy in a red t-shirt lying face down lifeless on the coasts of Turkey. I cried, feeling helpless. What did the boy do to deserve this, and how unlucky the parents must be to lose such a precious angel? Why do the place and conditions you are born into have to be such big factors in deciding your fate? It is not the first time these questions occurred to me. But this time it hit me harder.</p><blockquote>3 million children under five years died in 2015 due to conditions that could be prevented or treated with access to simple, affordable interventions</blockquote><p>Donating to organizations helping with refugee crisis was one option, but <a href="http://blog.givewell.org/2013/11/12/6-tips-on-disaster-relief-giving/">how do I know which organization is having an impact and not adding to the disruption</a>? And what about day to day less publicized suffering? To put things in context, about 3 million children under five years died in 2015 due to conditions that could be prevented or treated with access to <a href="http://www.who.int/mediacentre/factsheets/fs178/en/">simple, affordable interventions </a>— that is around five children every second. There are around <a href="http://data.worldbank.org/topic/education">60 million kids</a> out of primary school. Almost half the world, over three billion people, live on <a href="http://www.globalissues.org/article/26/poverty-facts-and-stats">less than $2.50 a day</a>. <a href="http://www.theendofpoverty.com/">20% of the planet’s population uses 80% of its resources and consumes 30% more than the planet can regenerate</a>. With so much inequity in the world, each one of us has an obligation to give back and be active forces in leveling the playing field. While you are here, check out <a href="https://www.givingwhatwecan.org/">Giving What We Can</a>’s <a href="https://www.givingwhatwecan.org/get-involved/how-rich-am-i/">calculator</a> to find out how rich you are on a global scale.</p><blockquote><strong>But where do we start? Do we make donations or work directly for causes? How do we know if we are optimizing for impact? These are some hard questions which hinder giving and effective altruism.</strong></blockquote><p>Charitable donations are a form of investment, which has social returns instead of financial ones. Yet, we rarely embark on the same amount of due diligence and analysis as compared to buying stocks, for example. To have the biggest impact both empathy and critical thinking are necessary. I believe our generation has some great resources and frameworks at our disposal to have a tremendous impact. In this article, I make an attempt to collect some of the compelling and nontraditional paths millennials are taking to boost their impact.</p><h3>Effective giving and earning to give</h3><blockquote>How can I do the most good with my donation?</blockquote><p>Traditionally, charities are rated based on what percentage of the money goes towards program related activities versus operational costs. I believe this is not the best metric. It is entirely possible a charity has minimal overhead costs but does not have a real impact. See Dan Pallotta’s talk on “<a href="https://www.ted.com/talks/dan_pallotta_the_way_we_think_about_charity_is_dead_wrong">The way we think about charity is dead wrong</a>.” And moreover,<a href="http://highexistence.com/1000-times-more-altruistic-with-a-single-dollar/"> some charities are 100 times or 1000 times more effective than others</a>. In this regard, I love what <a href="http://www.givewell.org/">GiveWell</a> and <a href="https://www.givingwhatwecan.org/">Giving What We Can</a> do. They are organizations which perform rigorous evidence-based research on the outcomes of charities and rate them by their <a href="http://www.givewell.org/international/technical/criteria/impact">impact</a>. They mainly use randomized control trials to check for evidence of successful interventions - Much like how pharmaceutical companies vet out new drugs. These organizations importantly help us answer the most crucial question in giving — How can I do the most good with my donation? For the more general form of the same question “How can I do the most good in my lifetime?”, I strongly encourage checking out <a href="https://www.effectivealtruism.org/">Effective Altruism</a> and <a href="https://80000hours.org/">80,000 hours</a> (Yes, you guessed it right, that is the number of hours we spent in our day jobs).</p><p>This leads us to an interesting case where you might be able to optimize the impact by deliberately taking a high earning job which is morally neutral or positive and donate a very significant portion of your earning. The premise here is that some people have skills which are better suited for earning money than doing good directly. This approach is most efficient for cause areas where there is a funding gap. For example, long lasting insecticide treated nets (LLIN) is one of the most cost-effective ways to prevent Malaria and it is <a href="http://www.givewell.org/charities/against-malaria-foundation">funding bounded</a> right now. Very roughly, <a href="http://www.givewell.org/international/technical/programs/insecticide-treated-nets#HowcosteffectiveisLLINdistribution">cost of saving a life is ~$3000</a>. Also, check out this <a href="https://twitter.com/BillGates/status/749619467466973184">beautiful map</a> which shows the progress made so far. Just take a moment here and think about saving a person’s life who is close to your heart. Saving a stranger is not much different when you consider their loved ones.</p><p>For inspiration, check out this <a href="http://www.today.com/news/couple-donates-half-income-every-year-treat-world-health-t51916">Boston couple who donate half their income</a> every year and so far donated more than 400K.</p><h3>Pro bono Volunteering and Social Intrapreneurship</h3><blockquote>Volunteers get to contribute their expertise and time, which might be more valuable than just donating money</blockquote><p><a href="http://www.idealist.org/info/Volunteer/Skilled">Pro bono volunteering</a> refers to companies and individuals volunteering their professional skills to assist nonprofit organizations in creating or improving their business practices. This is good as volunteers get to contribute their expertise and time, which might be more valuable than just donating money. For example, as most organizations do, nonprofits also can benefit from cutting-edge technology, but it is often expensive to hire full-time technical teams. But domain knowledge is also crucial to have most impact, so just tech skills alone would not help. Pro bono volunteering helps with this problem by connecting social sectors and tech workers on a semi long term projects. But I think there is immense value if you can channelize tech challenges from social sector to tech expertise in your company, there by amplifying the impact. This liaison is where <a href="http://www.forbes.com/sites/ashoka/2014/01/24/2014s-most-valuable-employee-the-social-intrapreneur/">social intrapreneurship</a> can be effective. You can help run your internal philanthropic efforts and specifically drive the tech-for-good aspects of it to drive product roadmap features which help small businesses and non profits. For example, <a href="http://www.forbes.com/sites/danielrunde/2015/08/12/m-pesa-and-the-rise-of-the-global-mobile-money-market/#7d490f6523f5">mPesa, </a>a mobile phone-based platform which contributed to a dramatic increase in the percentage of adults with access to formal financial services in Kenya, <a href="http://www.forbes.com/sites/danielrunde/2015/08/12/m-pesa-and-the-rise-of-the-global-mobile-money-market/#7d490f6523f5">was started by a few social intrapreneurs at Vodafone</a>. If you want to find more resources about social intrapreneurship, check out <a href="http://www.leagueofintrapreneurs.com/resources">The League of Intrapreneurs</a> and the impact work <a href="https://www.netimpact.org/sites/default/files/documents/impact-work-toolkit.pdf">toolkit</a> from Net Impact.</p><p>Being a data engineer, I am familiar with pro bono opportunities in data science and related fields. So here are some pointers: <a href="http://www.datakind.org/">Datakind</a>, <a href="http://www.bayesimpact.org/">Bayes Impact</a>, <a href="http://www.deltanalytics.org/">DeltaAnalytics</a> and <a href="http://community.amstat.org/statisticswithoutborders/home">Statistics Without Borders</a> provide with medium term opportunities to collaborate with nonprofits. There are more competition-style opportunities like <a href="https://www.kaggle.com/c/the-nature-conservancy-fisheries-monitoring">Kaggle competitions for Govt and Nonprofits</a> and <a href="https://www.drivendata.org/competitions/">Drivendata</a>. There are also more academic style opportunities like “<a href="https://dssg.uchicago.edu/">Data Science for Social good fellow</a>” at the University of Chicago. Also, check out “<a href="http://www.kdnuggets.com/2015/07/guide-data-science-good.html">Datalook’s Definitive Guide</a>” for more details on these opportunities.</p><p>As we talk about optimizing impact and skills, I would like to mention technologies like AI, specifically Deep learning, which I firmly believe can be great tools for massive impact in the coming years. Deep learning, through multiple layers of hidden neurons, allows us to model very complex nonlinear functions, which by definition of <a href="https://en.wikipedia.org/wiki/Universal_approximation_theorem">Universal approximation theorem</a> can be used to model almost anything. With the ubiquity of GPUs and higher level language abstractions to run neural networks, innovation in the field is flourishing. And it is proving to be successful in socially relevant problems such as detecting <a href="https://blogs.nvidia.com/blog/2016/09/19/deep-learning-breast-cancer-diagnosis/">Breast Cancer</a> and D<a href="http://jamanetwork.com/journals/jama/fullarticle/2588763">iabetic Retinopathy</a>(most common cause of blindness among diabetics). Here are some deep learning <a href="http://www.deeplearningweekly.com/pages/open_source_deep_learning_curriculum">resources</a> you might want to check out. I strongly recommend fast.ai’s MOOC — <a href="http://course.fast.ai/index.html">“<strong><em>Practical Deep Learning For Coders”</em></strong></a><strong><em> </em></strong>which Jeremy Howard and Rachael Thompson are offering from this week. I just completed the in-person version of the course and can vouch for it. Not just the content is very hands on and current, but there is also an emphasis on social relevance and ideas being exchanged on how to use this technology for greater good. Bonus — all proceeds from the course are donated to <a href="http://www.hollows.org/au/home?lang=en-au">Fred Hollows</a>, an organization which can restore vision for an individual for as little as $25.</p><h3>Social Entrepreneurship and Impact Investing</h3><p>From Wikipedia — “<em>A social enterprise is an organization that applies commercial strategies to maximize improvements in human and environmental well-being — this may include maximizing social impact alongside profits for external shareholders.</em>” Just to be clear, a social enterprise can either be a for-profit or a non-profit entity. This is an entrepreneurial option where you can lead and innovate a solution to a problem by establishing a company focused on solving that problem for the poorest socio-economic group or “Bottom of the pyramid” (BoP), as it is often referred to. This approach is especially useful in areas where there is no established scalable solution to the problem, and it is ripe for innovation. (You might want to check out an argument for talent gap versus funding gap <a href="https://80000hours.org/2015/11/why-you-should-focus-more-on-talent-gaps-not-funding-gaps/)">here</a>.) A well-celebrated success in this space is <a href="https://en.wikipedia.org/wiki/Muhammad_Yunus">Muhammad Yunus</a> who was an early social entrepreneur who pioneered micro credits. Some other more recent examples include <a href="https://www.wearethorn.org/">Thorn</a>, which drives technology innovation to fight the sexual exploitation of children; <a href="http://simpanetworks.com/">Simpa networks</a>, making clean energy affordable and accessible to everyone; <a href="http://www.unima.com.mx/nsite/index.html">Unima</a>, taking diagnostics to the 2 billion people in developing world. For more role models in this space check out this new age social entrepreneurs who won accolades from the <a href="https://www.weforum.org/agenda/2016/03/11-social-entrepreneurs-who-are-changing-the-world/">World Economic Forum</a>, <a href="http://www.forbes.com/special-report/2014/30-under-30/social-entrepreneurs.html">Forbes 30 Under 30</a>, and the <a href="http://widgets.weforum.org/social-entrepreneurs-2016/">Schwab Foundation</a>.</p><p>More links for interested folks:</p><ul><li>M<a href="http://www.creatorsforgood.com/2016/03/28/9-must-read-books-for-social-entrepreneurs/">ust-read books</a> for aspiring social entrepreneurs</li><li>Resourceful networks like <a href="https://www.ashoka.org/">Ashoka Changemakers</a> and <a href="http://skoll.org/">Skoll Foundation</a>.</li><li><a href="http://blog.givewell.org/2015/10/15/charities-wed-like-to-see/">Ideas for Social enterprises</a> from GiveWell</li></ul><blockquote><a href="https://thegiin.org/impact-investing/need-to-know/">Impact investing</a> challenges the long-held views that social and environmental issues should be addressed only by philanthropic donations, and that market investments should focus exclusively on achieving financial returns</blockquote><p>But it is often harder to make profits when you are catering to the poorest of the poor. That is the main reason most of the economy revolves around the affluent for bigger margins. <a href="https://thegiin.org/impact-investing/need-to-know/">Impact investing</a> challenges this long-held view that social and environmental issues should be addressed only by philanthropic donations, and that market investments should focus exclusively on achieving financial returns. It does so by optimizing not just for financial returns but also for social-environmental impact. More capital flowing into Social sector encourages long-term planning for sustainability as well as fosters innovation.</p><p>Again the fundamental difficulty here is — it is easy to measure financial returns, but it is rather hard to measure impact. In the Impact Investing world, there are various <a href="http://impactinvesting.marsdd.com/social-impact-measurement/">frameworks and tools</a> to gauge the impact like GIIRS, <a href="https://iris.thegiin.org/">IRIS</a>, B Certification, Demonstrating Value, etc. I see some of them are more outcome based and some are more efficiency and transparency based. It would be interesting to see progress in this area. Ultimately, this research and tooling around Impact measurement would be tremendously impactful if it makes it easy for all organizations to make impact driven decisions just like how natural it is today to make decisions based on financial value.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*uKK-6i1BJg9uNnrL3z_plg.png" /><figcaption>Source: <a href="http://impactinvesting.marsdd.com/social-impact-measurement/measurement-dashboard/">http://impactinvesting.marsdd.com/social-impact-measurement/measurement-dashboard/</a></figcaption></figure><p><a href="https://ssir.org/">Stanford Social Innovation Review</a> is an excellent source to follow the SoCap(Social capital) world as well as other innovations in philanthropy space. As more shareholders value impact I think it becomes imperative for all companies to report on impact along with financial gains. It turns out you can in fact set up legal structures to formally define intentions for the triple bottom line: Profits, environmental impact and social impact — L3Cs and Benefit corporations fall into this category. You can also chose more traditional legal structures such as LLC (like Chan-Zuckerberg Initiative), Nonprofit or Hybrid (A nonprofit and a for profit like Omidyar Network).</p><p>Eventually, I wish all enterprises would be social enterprises, measuring not just their profits but also their impact. So impact conscious millennials can feed in this parameter in their career decisions. A simpler starting point maybe for Glassdoor to include the“Impact” criteria for employees to rate on just like work/life balance and CEO ratings. I am sure Cloudera, my previous employer, would come up pretty high among mid sized tech companies — its product has a big impact in the Social sector, and it provides various pro bono opportunities to its employees through Cloudera Cares.</p><h3>Closing remarks</h3><p>This list is in no way an exhaustive list of paths for optimal impact, but just a starting point. Activism, public policy making and influencing, basic sciences and political sciences research, meta-charities, and traditional local volunteering are some of the other forms. We can choose our impact strategy just like how we choose our investment options in our portfolio based on our risk tolerance and current financial and personal situations. We can either chose one path, or try out different ways over a period.</p><p>Either way, I think it pays off to start early and evaluate your strategy periodically to maximize the impact over the lifetime. As Melinda Gate’s mother said, <em>“</em><a href="https://journal.thriveglobal.com/melinda-gates-on-the-invaluable-advice-her-mom-gave-her-about-success-as-a-teenager-95f93320b71d#.9fhk82ost"><em>To know that even one life has breathed easier because you have lived, that is success.</em></a><em>”</em> Here’s wishing everyone an even more successful (read impactful) 2017!</p><p><strong>Acknowledgements:</strong> Thanks to my wonderful husband/critic <a href="https://medium.com/u/d97a4e2970f6">Venu Satuluri</a> and my best mentor/friend <a href="https://medium.com/u/32bc04bcb7e5">Anjana Prabhakar</a> for providing feedback on an earlier draft. And special thanks to all the wonderful people who inspire me every day!</p><p>Update: I have received messages mentioning that this advice applies to non millennials as well! In hind sight, I feel so too. So updating the title to reflect the same.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=d2bdfcbfa4ce" width="1" height="1" alt=""><hr><p><a href="https://medium.com/impactai/millennials-here-are-three-ways-to-be-effective-and-creative-in-your-philanthropic-endeavors-d2bdfcbfa4ce">Millennials — Here are three ways to be effective (and creative) in your philanthropic endeavors</a> was originally published in <a href="https://medium.com/impactai">ImpactAI</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
    </channel>
</rss>