BEGIN:VCALENDAR VERSION:2.0 PRODID:-//HasGeek//NONSGML Funnel//EN DESCRIPTION:A conference on AI and Deep Learning NAME:Anthill Inside 2019 REFRESH-INTERVAL;VALUE=DURATION:PT12H SUMMARY:Anthill Inside 2019 TIMEZONE-ID:Asia/Kolkata X-PUBLISHED-TTL:PT12H X-WR-CALDESC:A conference on AI and Deep Learning X-WR-CALNAME:Anthill Inside 2019 X-WR-TIMEZONE:Asia/Kolkata BEGIN:VEVENT SUMMARY:Check-in and onsite registrations DTSTART;VALUE=DATE-TIME:20191123T030000Z DTEND;VALUE=DATE-TIME:20191123T034500Z DTSTAMP;VALUE=DATE-TIME:20210227T094116Z UID:session/Ec9vpCQADwEsMGDzdPhNR1@hasgeek.com CREATED;VALUE=DATE-TIME:20191114T124110Z DESCRIPTION:\n LAST-MODIFIED;VALUE=DATE-TIME:20191114T124116Z ORGANIZER;CN="Anthill Inside":MAILTO:no-reply@hasgeek.com BEGIN:VALARM ACTION:display DESCRIPTION:Check-in and onsite registrations in 5 minutes TRIGGER:-PT5M END:VALARM END:VEVENT BEGIN:VEVENT SUMMARY:Introduction to Anthill Inside 2019 DTSTART;VALUE=DATE-TIME:20191123T034500Z DTEND;VALUE=DATE-TIME:20191123T035500Z DTSTAMP;VALUE=DATE-TIME:20210227T094116Z UID:session/PrDzFs3FZr1XLF5Xt2RJzE@hasgeek.com CREATED;VALUE=DATE-TIME:20190806T093441Z DESCRIPTION:\n GEO:12.973321659788686;77.61947496794164 LAST-MODIFIED;VALUE=DATE-TIME:20191122T060940Z LOCATION:Ballroom - Taj M G Road\, Bangalore\nBangalore\, IN ORGANIZER;CN="Anthill Inside":MAILTO:no-reply@hasgeek.com BEGIN:VALARM ACTION:display DESCRIPTION:Introduction to Anthill Inside 2019 in Ballroom in 5 minutes TRIGGER:-PT5M END:VALARM END:VEVENT BEGIN:VEVENT SUMMARY:Introduction to Naman Kumar's talk on Robotics DTSTART;VALUE=DATE-TIME:20191123T035500Z DTEND;VALUE=DATE-TIME:20191123T040000Z DTSTAMP;VALUE=DATE-TIME:20210227T094116Z UID:session/VuYzkHmRJyCidtttsY9Sjk@hasgeek.com CREATED;VALUE=DATE-TIME:20191122T061009Z DESCRIPTION:\n GEO:12.973321659788686;77.61947496794164 LAST-MODIFIED;VALUE=DATE-TIME:20191122T061137Z LOCATION:Ballroom - Taj M G Road\, Bangalore\nBangalore\, IN ORGANIZER;CN="Anthill Inside":MAILTO:no-reply@hasgeek.com BEGIN:VALARM ACTION:display DESCRIPTION:Introduction to Naman Kumar's talk on Robotics in Ballroom in 5 minutes TRIGGER:-PT5M END:VALARM END:VEVENT BEGIN:VEVENT SUMMARY:What can software learn from robots and math DTSTART;VALUE=DATE-TIME:20191123T040000Z DTEND;VALUE=DATE-TIME:20191123T044000Z DTSTAMP;VALUE=DATE-TIME:20210227T094116Z UID:session/Rda85rrDA6EbXbiRTcAQCz@hasgeek.com CATEGORIES:Full talk,Advanced,Lecture CREATED;VALUE=DATE-TIME:20191027T070555Z DESCRIPTION:If you have no clue about what is going on\, don’t worry. In this presentation\, I will try to build your intuition with a series of s imple examples. Then\, with a little bit of math\, I will demonstrate how the Kalman filter works its charm. Finally\, I will end by giving you a gl impse of its numerous applications in different fields and how you can pro bably use it in your own project.\n\n### Speaker bio\n\nhttps://www.linked in.com/in/naman-kumar-0582004a/\n GEO:12.973321659788686;77.61947496794164 LAST-MODIFIED;VALUE=DATE-TIME:20200619T062515Z LOCATION:Ballroom - Taj M G Road\, Bangalore\nBangalore\, IN ORGANIZER;CN="Anthill Inside":MAILTO:no-reply@hasgeek.com URL:https://hasgeek.com/anthillinside/2019/schedule/what-can-software-lear n-from-robots-and-math-Rda85rrDA6EbXbiRTcAQCz BEGIN:VALARM ACTION:display DESCRIPTION:What can software learn from robots and math in Ballroom in 5 minutes TRIGGER:-PT5M END:VALARM END:VEVENT BEGIN:VEVENT SUMMARY:Rigorous evaluation of NLP models for real-world deployment DTSTART;VALUE=DATE-TIME:20191123T044000Z DTEND;VALUE=DATE-TIME:20191123T052000Z DTSTAMP;VALUE=DATE-TIME:20210227T094116Z UID:session/DhgpbXL1ZBCmoVJucovFu3@hasgeek.com CATEGORIES:Full talk,Intermediate,Discussion CREATED;VALUE=DATE-TIME:20191014T064157Z DESCRIPTION:We motivate why rigorous evaluation of NLP models beyond simpl e metrics such as F1-score/accuracy are needed for real world deployment w ith a few historical use-cases/examples. We then talk about the "CleverHan s Moment for NLP" (https://www.linkedin.com/posts/sandya_nlps-clever-hans- moment-has-arrived-activity-6573894455768768512-MDVW). We discuss the late st research around model evaluation for NLP. We then take up the example o f a sentiment analysis task as a case-study and discuss the methodology fo r rigorous evaluation. We conclude by pointing out future work directions in this topic.\n\n### Speaker bio\n\nSandya Mannarswamy (https://www.linke din.com/in/sandya/) is an independent NLP researcher. She was previously a senior research scientist at Conduent Labs India in the Natural Language Processing research group. She holds a Ph.D. in computer science from Ind ian Institute of Science\, Bangalore. Her research interests span natural language processing\, machine learning and compilers. Her research career spans over 19 years\, at various R&D labs\, including Hewlett Packard Ltd \, IBM Research etc. She has co-organized a number of workshops including workshops at International Conference on Data Management\, Machine Learni ng Debates workshop at ICML-2018 etc. Her current research is focused on s oftware testing and evaluation of Natural Language Processing application s. She has a number of international research publications and patents in the area of natural language processing (https://scholar.google.co.in/cita tions?hl=en&user=i27nd3oAAAAJ&view_op=list_works&sortby=pubdate) She co-au thored a paper at International Conference on Artificial Intelligence (IJC AI) 2018\, which focused on the challenges in taking AI applications from research to real world. Her current research is focussed on rigorous eval uation of NLP applications (“using NLP to evaluate NLP”). She is the author of the popular “CodeSport” column in Open Source For You magazi ne. (https://opensourceforu.com/tag/codesport/).\n GEO:12.973321659788686;77.61947496794164 LAST-MODIFIED;VALUE=DATE-TIME:20200619T062515Z LOCATION:Ballroom - Taj M G Road\, Bangalore\nBangalore\, IN ORGANIZER;CN="Anthill Inside":MAILTO:no-reply@hasgeek.com URL:https://hasgeek.com/anthillinside/2019/schedule/rigorous-evaluation-of -nlp-models-for-real-world-deployment-DhgpbXL1ZBCmoVJucovFu3 BEGIN:VALARM ACTION:display DESCRIPTION:Rigorous evaluation of NLP models for real-world deployment in Ballroom in 5 minutes TRIGGER:-PT5M END:VALARM END:VEVENT BEGIN:VEVENT SUMMARY:Poster session: Model interpretability\, explainable AI and the Ri ght to Information (RTI) DTSTART;VALUE=DATE-TIME:20191123T052000Z DTEND;VALUE=DATE-TIME:20191123T055000Z DTSTAMP;VALUE=DATE-TIME:20210227T094116Z UID:session/B986KGjGMnGc38PSHCC3yr@hasgeek.com CATEGORIES:Crisp talk,Beginner,Discussion,None,Accepted as poster session CREATED;VALUE=DATE-TIME:20191021T104704Z DESCRIPTION:Consequential machine decision making is now pervasive. Automa ted decisions (to different degrees of automation) are now applied in fiel ds of welfare allocation\, policing and criminal justice\, finance and ins urance and online content moderation\, among others. Many of these tools u se complex algorithmic systems\, including machine learning techniques\, w hich are conventionally difficult to interpret. Efforts toward interpretat ion have traditionally focused on model interpretation through explaining the 'black box' of algorithmic systems (for example through local linear e xplanations or models). However\, these techniques of interpretability hav e limited significance where end-users are concerned\, for a number of rea sons\, including the ability of a lay citizen to parse technical models\, as well as the limited information it provides for achieving instrumental purposes of explanation (for example\, the ability to use an explanation t o overturn a decision). Some techniques have focused on explainability wit hout opening the black box\, including through methods like counterfactual explanations. However\, limited work exists on how the non-interpretabili ty of machine decisions impacts important constitutional concepts of due p rocess and the right to information as well as legal mechanisms like the R TI Act which actualise these rights. The RTI Act\, in particular\, places positive obligations upon the state to explain certain decisions\, includi ng administrative decisions taken that impact individuals. The extent to w hich techniques of explainability in AI can be incorporated to ensure that the RTI remains a robust instrument for holding government systems accoun table will be the focus of this session.\n\n### Speaker bio\n\nI am a lawy er and a legal researcher\, working in the field of technology policy. I h ave researched and written extensively on issues of internet openness and digital rights. In my role as a technology policy fellow at the Mozilla Fo undation\, I am focussing on creating policy for improving machine decisio n making systems in India.\n GEO:12.973321659788686;77.61947496794164 LAST-MODIFIED;VALUE=DATE-TIME:20191115T071806Z LOCATION:Poster sessions and BOF track - Taj M G Road\, Bangalore\nBangalo re\, IN ORGANIZER;CN="Anthill Inside":MAILTO:no-reply@hasgeek.com URL:https://hasgeek.com/anthillinside/2019/schedule/model-interpretability -explainable-ai-and-the-right-to-information-B986KGjGMnGc38PSHCC3yr BEGIN:VALARM ACTION:display DESCRIPTION:Poster session: Model interpretability\, explainable AI and th e Right to Information (RTI) in Poster sessions and BOF track in 5 minutes TRIGGER:-PT5M END:VALARM END:VEVENT BEGIN:VEVENT SUMMARY:How we applied sampling algorithms to extract meaning from data DTSTART;VALUE=DATE-TIME:20191123T052000Z DTEND;VALUE=DATE-TIME:20191123T055000Z DTSTAMP;VALUE=DATE-TIME:20210227T094116Z UID:session/6PKGqBS9ekuyka7NzhtFZr@hasgeek.com CATEGORIES:Full talk of 40 mins CREATED;VALUE=DATE-TIME:20191108T060506Z DESCRIPTION:Generative models => Basic idea of sampling algorithms to infe rence parameters => Simple example using Gibbs sampling => Application to a more complex problem of resume understanding at Belong.co\n\n### Speaker bio\n\nCurrently CTO at Belong.co\, Vinodh Kumar is one of the top indust ry leaders with more than a decade of hands-on experience in search\, rank ing and machine learning. Prior to Belong\, Vinodh used to be CTO/M.D of B loomreach driving their e-commerce search engine efforts. Earlier Vinodh s pent more than 6 years at Google leading the Google News team and building the ranking algorithms that power Google News. He did his masters in comp uter science from the Indian Institute of Science after securing the All I ndia Rank #1 in Graduate Engineering Entrance Exam (GATE '99) in computer science. He has more than 10 patents to his name.\n GEO:12.973321659788686;77.61947496794164 LAST-MODIFIED;VALUE=DATE-TIME:20200619T062515Z LOCATION:Ballroom - Taj M G Road\, Bangalore\nBangalore\, IN ORGANIZER;CN="Anthill Inside":MAILTO:no-reply@hasgeek.com URL:https://hasgeek.com/anthillinside/2019/schedule/how-we-applied-samplin g-algorithms-to-extract-meaning-from-data-belong-co-6PKGqBS9ekuyka7NzhtFZr BEGIN:VALARM ACTION:display DESCRIPTION:How we applied sampling algorithms to extract meaning from dat a in Ballroom in 5 minutes TRIGGER:-PT5M END:VALARM END:VEVENT BEGIN:VEVENT SUMMARY:Morning beverage break DTSTART;VALUE=DATE-TIME:20191123T055000Z DTEND;VALUE=DATE-TIME:20191123T062000Z DTSTAMP;VALUE=DATE-TIME:20210227T094116Z UID:session/W4rYhjqq5ShMRQyasiCDnL@hasgeek.com CREATED;VALUE=DATE-TIME:20190806T093621Z DESCRIPTION:\n GEO:12.973321659788686;77.61947496794164 LAST-MODIFIED;VALUE=DATE-TIME:20191108T062216Z LOCATION:Ballroom - Taj M G Road\, Bangalore\nBangalore\, IN ORGANIZER;CN="Anthill Inside":MAILTO:no-reply@hasgeek.com BEGIN:VALARM ACTION:display DESCRIPTION:Morning beverage break in Ballroom in 5 minutes TRIGGER:-PT5M END:VALARM END:VEVENT BEGIN:VEVENT SUMMARY:Morning beverage break DTSTART;VALUE=DATE-TIME:20191123T055000Z DTEND;VALUE=DATE-TIME:20191123T062000Z DTSTAMP;VALUE=DATE-TIME:20210227T094116Z UID:session/Ciey1ABEGSSV774988skH4@hasgeek.com CREATED;VALUE=DATE-TIME:20191108T062213Z DESCRIPTION:\n GEO:12.973321659788686;77.61947496794164 LAST-MODIFIED;VALUE=DATE-TIME:20191108T062220Z LOCATION:Poster sessions and BOF track - Taj M G Road\, Bangalore\nBangalo re\, IN ORGANIZER;CN="Anthill Inside":MAILTO:no-reply@hasgeek.com BEGIN:VALARM ACTION:display DESCRIPTION:Morning beverage break in Poster sessions and BOF track in 5 m inutes TRIGGER:-PT5M END:VALARM END:VEVENT BEGIN:VEVENT SUMMARY:Introduction to Mira Abboud's talk on automated investments with A rtificial Intelligence DTSTART;VALUE=DATE-TIME:20191123T062000Z DTEND;VALUE=DATE-TIME:20191123T062500Z DTSTAMP;VALUE=DATE-TIME:20210227T094116Z UID:session/5Gbyb97dK6Z6SmyhYd79Bt@hasgeek.com CREATED;VALUE=DATE-TIME:20191122T061126Z DESCRIPTION:\n GEO:12.973321659788686;77.61947496794164 LAST-MODIFIED;VALUE=DATE-TIME:20191122T061147Z LOCATION:Ballroom - Taj M G Road\, Bangalore\nBangalore\, IN ORGANIZER;CN="Anthill Inside":MAILTO:no-reply@hasgeek.com BEGIN:VALARM ACTION:display DESCRIPTION:Introduction to Mira Abboud's talk on automated investments wi th Artificial Intelligence in Ballroom in 5 minutes TRIGGER:-PT5M END:VALARM END:VEVENT BEGIN:VEVENT SUMMARY:Artificial Intelligence for automated investment DTSTART;VALUE=DATE-TIME:20191123T062500Z DTEND;VALUE=DATE-TIME:20191123T070500Z DTSTAMP;VALUE=DATE-TIME:20210227T094116Z UID:session/8ZESyev9B324VvquhnCvB1@hasgeek.com CATEGORIES:Full talk,Advanced,Lecture CREATED;VALUE=DATE-TIME:20191014T064002Z DESCRIPTION:The talk will cover the following areas:\n- AI in finance vs AI in other fields.\n- Challenges faced while applying machine learning a lgorithms on stock market data (Daily data\, problems of Over/Under fittin g\, fat tails\, etc).\n- Limitations/problems of Supervised and Unsupervi sed learning\n- State of the art solutions.\n\n### Speaker bio\n\nAs a CT O and a data scientist at Neotic.ai SAL\, Dr. Mira Abboud leads Neotic’s long-term technology vision from supervising the technical team to improv ing the data science process development and is responsible for implementi ng new ideas after studying their feasibility. Mira is a Computer Sciences instructor at Lebanese University and researcher in the AI field. Holder of a Ph.D focused on AI and software architectures extraction\, from local & Nantes (France) university. Her Publications include "Towards Using KDD for an Interactive Software Architecture Extraction" and "KDD extension t ool for software architecture extraction".\n GEO:12.973321659788686;77.61947496794164 LAST-MODIFIED;VALUE=DATE-TIME:20200619T062515Z LOCATION:Ballroom - Taj M G Road\, Bangalore\nBangalore\, IN ORGANIZER;CN="Anthill Inside":MAILTO:no-reply@hasgeek.com URL:https://hasgeek.com/anthillinside/2019/schedule/artificial-intelligenc e-for-automated-investment-8ZESyev9B324VvquhnCvB1 BEGIN:VALARM ACTION:display DESCRIPTION:Artificial Intelligence for automated investment in Ballroom i n 5 minutes TRIGGER:-PT5M END:VALARM END:VEVENT BEGIN:VEVENT SUMMARY:Document digitization: rethinking with Deep Learning DTSTART;VALUE=DATE-TIME:20191123T070500Z DTEND;VALUE=DATE-TIME:20191123T074500Z DTSTAMP;VALUE=DATE-TIME:20210227T094116Z UID:session/RMnNsZnN6QzwMwufoBCgwg@hasgeek.com CATEGORIES:Full talk,Intermediate,Lecture CREATED;VALUE=DATE-TIME:20190827T090710Z DESCRIPTION:This talk will outline:\n* The problems and approaches we face d when building deep learning networks to solve problems in the informatio n extraction process.\n* Thought process on why and how we chose certain d eep learning strategies\n* The requirement for supervised learning\n* Limi tations of deep learning networks\n* Planning and executing research activ ities in short cycles\n* Evolution of team structures to support AI produc t building\n* Engineering practises required in building AI systems.\n\n## # Speaker bio\n\nNischal HP is currently the VP of Engineering and Data sc ience at Berlin based company omni:us\, which operates in the building of AI product for the insurance industry. \n\nPreviously\, he was a cofounder and data scientist at Unnati Data Labs\, where he worked towards building end-to-end data science systems in the fields of fintech\, marketing anal ytics\, event management and medical domain. Nischal is also a mentor for data science on Springboard. During his tenure at former companies like Re dmart and SAP\, he was involved in architecting and building software for ecommerce systems in catalog management\, recommendation engines\, sentime nt analyzers \, data crawling frameworks\, intention mining systems and ga mification of technical indicators for algorithmic trading platforms. \n\n Nischal has conducted workshops in the field of deep learning and has spok en at a number of data science conferences like Strata London 2019\, Qcon AI SF 2019\, Pycon Canda 2018\, Oreilly strata San jose 2017\, PyData Lond on 2016\, Pycon Czech Republic 2015\, Fifthelephant India (2015 and 2016)\ , Anthill\, Bangalore 2016. He is a strong believer of open source and lov es to architect big\, fast\, and reliable AI systems. In his free time\, h e enjoys traveling with his significant other\, music and groking the web. \n GEO:12.973321659788686;77.61947496794164 LAST-MODIFIED;VALUE=DATE-TIME:20200619T062515Z LOCATION:Ballroom - Taj M G Road\, Bangalore\nBangalore\, IN ORGANIZER;CN="Anthill Inside":MAILTO:no-reply@hasgeek.com URL:https://hasgeek.com/anthillinside/2019/schedule/document-digitization- rethinking-it-with-deep-learning-RMnNsZnN6QzwMwufoBCgwg BEGIN:VALARM ACTION:display DESCRIPTION:Document digitization: rethinking with Deep Learning in Ballro om in 5 minutes TRIGGER:-PT5M END:VALARM END:VEVENT BEGIN:VEVENT SUMMARY:Birds of Feather (BOF) session: Myths and realities of data labeli ng for Deep Learning DTSTART;VALUE=DATE-TIME:20191123T070500Z DTEND;VALUE=DATE-TIME:20191123T080500Z DTSTAMP;VALUE=DATE-TIME:20210227T094116Z UID:session/T6quRXcrsUZukXCBiJLasM@hasgeek.com CATEGORIES:Birds Of Feather (BOF) session ,Intermediate,Discussion CREATED;VALUE=DATE-TIME:20190823T050603Z DESCRIPTION:- setting the context : data labeling for NLP and CV\n- how to define a data labeling task : novice vs expert\n- does crowd sourcing of data labeling really work : adv vs disadv.\n- how to manage in house data labeling teams : adv vs disadv\n- what is the criticality of the correctne ss of data labels\n- what is the experience and expertise expectation of d ata labelers\n- how to ensure correctness of data labels : manual vs autom ated checks\n- how to resolve labeling conflicts\n- how does an engineer k now if she has enough labeled data\n- what are the time\, cost\, correctne ss trade-offs\n- how to ensure and execute class balanced data labeling\n- how to plan and execute weakly supervised data labeling\n- how to train m odels on small set of labeled data and generate 'soft tags' for the rest o f the unlabeled data\n- how does one know if a model is performing well in practice on unseen and real-time inputs\n- how does feedback loop work wh en some of the unseen and real-time inputs are labeled to fine-tune the mo dels\n\n### Speaker bio\n\nVijay is the co-founder and CTO of Infilect Tec hnologies\, a Computer Vision and Deep Learning start-up\, builidng B2B Sa aS products for global retail industry. Vijay has a PhD in CSE\, from IIT Bombay. Vijay has worked as research scientist in IBM Research Labs.\n GEO:12.973321659788686;77.61947496794164 LAST-MODIFIED;VALUE=DATE-TIME:20191122T061150Z LOCATION:Poster sessions and BOF track - Taj M G Road\, Bangalore\nBangalo re\, IN ORGANIZER;CN="Anthill Inside":MAILTO:no-reply@hasgeek.com URL:https://hasgeek.com/anthillinside/2019/schedule/myths-and-realities-of -data-labeling-for-deep-learning-T6quRXcrsUZukXCBiJLasM BEGIN:VALARM ACTION:display DESCRIPTION:Birds of Feather (BOF) session: Myths and realities of data la beling for Deep Learning in Poster sessions and BOF track in 5 minutes TRIGGER:-PT5M END:VALARM END:VEVENT BEGIN:VEVENT SUMMARY:Lunch break DTSTART;VALUE=DATE-TIME:20191123T074500Z DTEND;VALUE=DATE-TIME:20191123T083500Z DTSTAMP;VALUE=DATE-TIME:20210227T094116Z UID:session/6ftLEGrBoybXE9pcJ6x1oB@hasgeek.com CREATED;VALUE=DATE-TIME:20190806T093928Z DESCRIPTION:\n GEO:12.973321659788686;77.61947496794164 LAST-MODIFIED;VALUE=DATE-TIME:20191122T061155Z LOCATION:Ballroom - Taj M G Road\, Bangalore\nBangalore\, IN ORGANIZER;CN="Anthill Inside":MAILTO:no-reply@hasgeek.com BEGIN:VALARM ACTION:display DESCRIPTION:Lunch break in Ballroom in 5 minutes TRIGGER:-PT5M END:VALARM END:VEVENT BEGIN:VEVENT SUMMARY:Lunch break DTSTART;VALUE=DATE-TIME:20191123T080500Z DTEND;VALUE=DATE-TIME:20191123T083500Z DTSTAMP;VALUE=DATE-TIME:20210227T094116Z UID:session/4Buzo2sFVQRX2RHbnXohH9@hasgeek.com CREATED;VALUE=DATE-TIME:20191108T062231Z DESCRIPTION:\n GEO:12.973321659788686;77.61947496794164 LAST-MODIFIED;VALUE=DATE-TIME:20191122T061157Z LOCATION:Poster sessions and BOF track - Taj M G Road\, Bangalore\nBangalo re\, IN ORGANIZER;CN="Anthill Inside":MAILTO:no-reply@hasgeek.com BEGIN:VALARM ACTION:display DESCRIPTION:Lunch break in Poster sessions and BOF track in 5 minutes TRIGGER:-PT5M END:VALARM END:VEVENT BEGIN:VEVENT SUMMARY:Why smart-device based virtual assistants are incapable of assisti ng with gender based violence concerns in India. DTSTART;VALUE=DATE-TIME:20191123T083500Z DTEND;VALUE=DATE-TIME:20191123T090500Z DTSTAMP;VALUE=DATE-TIME:20210227T094116Z UID:session/8PUjkrJ5LYLd21ey5CtnTE@hasgeek.com CATEGORIES:Full talk of 40 mins,Full talk,Intermediate,Lecture CREATED;VALUE=DATE-TIME:20191108T061238Z DESCRIPTION:Part 1. Introduction to Gendered Biases\nThe talk will begin w ith a brief introduction to fairness and gendered bias concerns in Artific ial Intelligence technologies with relevant examples. \n\nPart 2. Are Sma rt-Device Based Virtual Assistants Capable of Assisting with Gender Based Violence\nConcerns in India?\nI will present my research which critically examines the responses of five Virtual Assistants in India -- Siri\, Googl e Now\, Bixby\, Cortana\, and Alexa -- to a standardized set of concerns r elated to Gender-Based Violence (GBV). A set of concerns regarding Sexual Violence and Cyber Violence were posed in the Virtual Assistant’s natur al language\, English. Non-crisis concerns were asked to set a baseline. A ll crisis responses by the Virtual Assistants were characterized based on the ability to (1) recognize the crisis\, (2) respond with respectful lang uage\, and (3) refer to an appropriate helpline\, or other resources. The findings of my study indicate missed opportunities to leverage technology to improve referrals to crisis support services in response to gender-base d violence.\nRead my paper here: https://itforchange.net/e-vaw/wp-content/ uploads/2018/01/Are-Smart-Device-Based-Virtual-Assistants-Capable-of-Assis ting-with-Gender-Based-Violence-Concerns-in-India-1.pdf\n\nPart 3. Feminis t Perspectives on the Social Media Construction of Artificial Intelligence \nI will analyse how Microsoft’s Twitter bot Tay went from tweeting “c an i just say that im stoked to meet u? humans are super cool” to “I . ... hate feminists and they should all die and burn in hell” and how we can avoid designing such biased AI technologies for the future.\nRead my w ork here: https://gendermediacultureblog.wordpress.com/2018/12/24/feminist -perspectives-on-the-social-media-construction-of-artificial-intelligence/ \n\n### Speaker bio\n\nI currently work as a Programme Officer at the Cent re for Internet and Society (CIS)\, New Delhi\, researching the intersecti ons of gender and emerging technologies such as Artificial Intelligence. P reviously\, I worked with the Internet Governance Forum of the United Nati ons as a Consultant on Gender and Access (2018)\, and with the Association of Progressive Communications (APC) (2017) on gender and technology. I ha ve a Master’s degree in Women’s Studies from the Tata Institute of Soc ial Sciences (TISS)\, Mumbai\, and a Bachelor’s degree in Computer Scien ce Engineering from M.S. Ramaiah Institute of Technology\, Bengaluru. Outs ide of work\, you will find me tweeting about feminism\, writing on Medium \, and engaging with grassroots political activism.\nTwitter: @so_radhikal \nLinkedIn: https://www.linkedin.com/in/radhika-radhakrishnan/ \nMedium: h ttps://medium.com/radhika-radhakrishnan\n GEO:12.973321659788686;77.61947496794164 LAST-MODIFIED;VALUE=DATE-TIME:20191122T061237Z LOCATION:Ballroom - Taj M G Road\, Bangalore\nBangalore\, IN ORGANIZER;CN="Anthill Inside":MAILTO:no-reply@hasgeek.com URL:https://hasgeek.com/anthillinside/2019/schedule/gendered-biases-in-art ificial-intelligence-8PUjkrJ5LYLd21ey5CtnTE BEGIN:VALARM ACTION:display DESCRIPTION:Why smart-device based virtual assistants are incapable of ass isting with gender based violence concerns in India. in Ballroom in 5 minu tes TRIGGER:-PT5M END:VALARM END:VEVENT BEGIN:VEVENT SUMMARY:Birds of a Feather (BOF) session: On Interpretability DTSTART;VALUE=DATE-TIME:20191123T083500Z DTEND;VALUE=DATE-TIME:20191123T093500Z DTSTAMP;VALUE=DATE-TIME:20210227T094116Z UID:session/YcPoT6Q3C1DaCgkDx63grZ@hasgeek.com CATEGORIES:Birds Of Feather (BOF) session ,Intermediate,Discussion CREATED;VALUE=DATE-TIME:20190813T063957Z DESCRIPTION:- Why is model interpretability important?\n- Trade off betwee n accuracy and interpretability.\n- Developments in explainable AI.\n- Int erpret black box models\, global and local interpretation.\n\n### Speaker bio\n\nLed by Jacob Joseph\, Namrata Hanspal. Discussants: Nishant Sinha a nd Madhu Gopinathan\n GEO:12.973321659788686;77.61947496794164 LAST-MODIFIED;VALUE=DATE-TIME:20191122T061200Z LOCATION:Poster sessions and BOF track - Taj M G Road\, Bangalore\nBangalo re\, IN ORGANIZER;CN="Anthill Inside":MAILTO:no-reply@hasgeek.com URL:https://hasgeek.com/anthillinside/2019/schedule/birds-of-a-feather-on- interpretability-YcPoT6Q3C1DaCgkDx63grZ BEGIN:VALARM ACTION:display DESCRIPTION:Birds of a Feather (BOF) session: On Interpretability in Poste r sessions and BOF track in 5 minutes TRIGGER:-PT5M END:VALARM END:VEVENT BEGIN:VEVENT SUMMARY:ML application lifecycle: what is important at each stage DTSTART;VALUE=DATE-TIME:20191123T090500Z DTEND;VALUE=DATE-TIME:20191123T094500Z DTSTAMP;VALUE=DATE-TIME:20210227T094116Z UID:session/KGWPdgsMtZENo8MscTWkud@hasgeek.com CATEGORIES:Full talk,Intermediate,Lecture CREATED;VALUE=DATE-TIME:20191101T103137Z DESCRIPTION:Building good ML systems is not very unlike developing good so ftware. Just as developing good software requires mastering not only progr amming theory\, tools\, and design patterns\, but also the process of soft ware development itself\, building a good ML system entails familiarity wi th the ML application lifecycle. In this talk\, we will discuss the vario us stages of ML application life cycle - problem formulation\, data defini tions\, modeling\, production system design &implementation\, testing\, deployment & maintenance\, online evaluation & evolution\, and some key l earnings that are relevant for each of these stages.\n\n### Speaker bio\n\ nSrujana is an independent machine learning researcher and consultant with over 15 years of experience. Till recently\, she was the chief scientist of CuspEra\, a software marketplace startup. Prior to that\, she was a pri ncipal data scientist at Flipkart (Bangalore) and a volunteer for Ekstep\, an education startup. She has been employed with the machine learning gro ups at Amazon (Bangalore)\, IBM Research (Bangalore/New Delhi/Almaden/York town Hts)\, and Yahoo Research (Santa Clara). Srujana has published her wo rk in several top-tier conferences and journals on data mining/machine lea rning and is the recipient of multiple best paper awards. She received her M.S. and Ph.D. from the University of Texas at Austin and her B. Tech. de gree from IIT Madras.\n GEO:12.973321659788686;77.61947496794164 LAST-MODIFIED;VALUE=DATE-TIME:20200619T062515Z LOCATION:Ballroom - Taj M G Road\, Bangalore\nBangalore\, IN ORGANIZER;CN="Anthill Inside":MAILTO:no-reply@hasgeek.com URL:https://hasgeek.com/anthillinside/2019/schedule/ml-application-lifecyc le-recommendations-for-each-stage-KGWPdgsMtZENo8MscTWkud BEGIN:VALARM ACTION:display DESCRIPTION:ML application lifecycle: what is important at each stage in B allroom in 5 minutes TRIGGER:-PT5M END:VALARM END:VEVENT BEGIN:VEVENT SUMMARY:Poster session: Open source tools and archive for tackling misinfo rmation on chat apps in India DTSTART;VALUE=DATE-TIME:20191123T094500Z DTEND;VALUE=DATE-TIME:20191123T101500Z DTSTAMP;VALUE=DATE-TIME:20210227T094116Z UID:session/BgdEbU74n2oEReUJtjzkjx@hasgeek.com CATEGORIES:Short talk of 20 mins CREATED;VALUE=DATE-TIME:20191114T111148Z DESCRIPTION:* Motivation and Goals of the Project\n * How does it aim to affect the misinformation challenge in India\n* Data Collection \n * Ways of collecting media from Chat Apps\n * Collecting media from allied sour ces (fact checking websites)\n* Data Processing (Tools to navigate the arc hive)\n * Duplicate Detection\n * Approximate Search\n * Semantic Searc h\n * Use of embeddings over hashing\n* Ethical Considerations in this wo rk\n * Consent frameworks for data collection\n * Managing access and us e\n * Managing violent and pornographic content\n\n### Speaker bio\n\nKes hav Joshi is a data scientist @Tattle working to bring together an archive of misinformation and keep developing the data science stack. Keshav has several years of experience as a data scientist/researcher/lecturer\, with two Masters in Physics & CS from Georgia Tech.\n GEO:12.973321659788686;77.61947496794164 LAST-MODIFIED;VALUE=DATE-TIME:20191122T111136Z LOCATION:Poster sessions and BOF track - Taj M G Road\, Bangalore\nBangalo re\, IN ORGANIZER;CN="Anthill Inside":MAILTO:no-reply@hasgeek.com URL:https://hasgeek.com/anthillinside/2019/schedule/open-source-tools-and- archive-for-tackling-misinformation-on-chatapps-in-india-BgdEbU74n2oEReUJt jzkjx BEGIN:VALARM ACTION:display DESCRIPTION:Poster session: Open source tools and archive for tackling mis information on chat apps in India in Poster sessions and BOF track in 5 mi nutes TRIGGER:-PT5M END:VALARM END:VEVENT BEGIN:VEVENT SUMMARY:Why you need an enterprise grade data labelling pipeline to scale your ML/AI pipelines DTSTART;VALUE=DATE-TIME:20191123T094500Z DTEND;VALUE=DATE-TIME:20191123T102500Z DTSTAMP;VALUE=DATE-TIME:20210227T094116Z UID:session/9uzSTVnNZ9F9EfinvbcJRo@hasgeek.com CATEGORIES:Full talk of 40 mins CREATED;VALUE=DATE-TIME:20191108T061028Z DESCRIPTION:In Software 2.0\, Data is code. A mindful approach to your dat a annotation pipeline and practices is critical to the outcomes of your ML algorithms. If not done right\, your ability to scale this pipeline can o ften prove to be a major blocker to productionization.\n\nIn this talk we focus on why and how to build your data labeling pipeline to be enterprise grade. We will describe the considerations and insights that go into maki ng your data pipeline a mindful part of your development pipeline\, so tha t you can follow the journey from PoC to production. We describe best prac tices and provide pointers to designing a high quality\, iterative\, and s calable data annotation practice.\n\nA pipeline designed for human judgeme nt and incremental training on edge cases\, can provide that last mile of acceptability to roll out a machine learning solution in production. We wi ll describe successful examples of this approach.\n\n### Speaker bio\n\nBi kram is broadly interested in the role of AI in addressing challenges and improving outcomes in education\, skill development and employability. At iMerit\, Bikram heads innovation with a primary focus on designing tech-en abled learning experiences that enable iMerit’s workforce – sourced la rgely from impact communities - develop digital economy skills and deliver high quality data annotation services to AI organizations. Bikram is also CTO of Anudip Foundation\, iMerit’s not-for-profit sister organization that trains youth from underserved communities in new age digital skills a nd helps them find gainful employment. Prior to joining iMerit and Anudip\ , Bikram spent 15 years at IBM Research\, establishing and leading global R&D programs in educational technologies\, service delivery\, and software engineering. Bikram holds MS and PhD degrees in Computer Science from the State University of New York\, Stony Brook. He has co-authored more than 50 scientific publications in international journals and conferences\, and holds several US patents.\n GEO:12.973321659788686;77.61947496794164 LAST-MODIFIED;VALUE=DATE-TIME:20200619T062515Z LOCATION:Ballroom - Taj M G Road\, Bangalore\nBangalore\, IN ORGANIZER;CN="Anthill Inside":MAILTO:no-reply@hasgeek.com URL:https://hasgeek.com/anthillinside/2019/schedule/build-an-enterprise-gr ade-data-labelling-pipeline-to-scale-your-ml-ai-pipelines-9uzSTVnNZ9F9Efin vbcJRo BEGIN:VALARM ACTION:display DESCRIPTION:Why you need an enterprise grade data labelling pipeline to sc ale your ML/AI pipelines in Ballroom in 5 minutes TRIGGER:-PT5M END:VALARM END:VEVENT BEGIN:VEVENT SUMMARY:Evening beverage break DTSTART;VALUE=DATE-TIME:20191123T102500Z DTEND;VALUE=DATE-TIME:20191123T105000Z DTSTAMP;VALUE=DATE-TIME:20210227T094116Z UID:session/EK48ZgdpnHRWotHQD8yhsA@hasgeek.com CREATED;VALUE=DATE-TIME:20190806T094544Z DESCRIPTION:\n GEO:12.973321659788686;77.61947496794164 LAST-MODIFIED;VALUE=DATE-TIME:20191122T061209Z LOCATION:Ballroom - Taj M G Road\, Bangalore\nBangalore\, IN ORGANIZER;CN="Anthill Inside":MAILTO:no-reply@hasgeek.com BEGIN:VALARM ACTION:display DESCRIPTION:Evening beverage break in Ballroom in 5 minutes TRIGGER:-PT5M END:VALARM END:VEVENT BEGIN:VEVENT SUMMARY:Poster session: Accountable behavioural change detection (VEDAR) u sing ML DTSTART;VALUE=DATE-TIME:20191123T105000Z DTEND;VALUE=DATE-TIME:20191123T112000Z DTSTAMP;VALUE=DATE-TIME:20210227T094116Z UID:session/CRYS2k7VgKKUQ5Jqiu2skP@hasgeek.com CATEGORIES:Advanced,Potentially Anthill Inside talk ,Awaiting review on em ail to decide on whether to shortlist or reject,This proposal is NOT a rej ect yet.,Short talk of 20 mins,Under review,Has potential. Mentor proposer . ,Respond to comments asap,Accepted as poster session CREATED;VALUE=DATE-TIME:20191021T104609Z DESCRIPTION:This talk mainly covers VEDAR algorithem in detail and benchma rks comparison with other streamingly anomoly detection. More details in the https://arxiv.org/abs/1902.06663\n\n### Speaker bio\n\nAravilli Srini vasa Rao working as Sr. Engineering Manager in Cisco CTO group and leading innovation & incubation of ML and AI projects. As a speaker presented in following conferences/workshops\n1) Presented about Cisco's ML/AI Applicat ions in PDPC/CIPL workshop in Singapore. As a panelist shared experiences and thoughts on Accountable and Responsible AI. 2 ) Presented in IoT and AI Sumit organized by CII ’s in India about IoT and ML applications and related platforms in IoT space. 3) Presented about “Streaming Anomaly D etection” in Cisco’s Data Science Summit in Prague \n\nHe has a patent in Software recommendations uisng Reinforcement Learning.\n GEO:12.973321659788686;77.61947496794164 LAST-MODIFIED;VALUE=DATE-TIME:20191122T061211Z LOCATION:Poster sessions and BOF track - Taj M G Road\, Bangalore\nBangalo re\, IN ORGANIZER;CN="Anthill Inside":MAILTO:no-reply@hasgeek.com URL:https://hasgeek.com/anthillinside/2019/schedule/accountable-behavioura l-change-detection-vedar-using-machine-learning-CRYS2k7VgKKUQ5Jqiu2skP BEGIN:VALARM ACTION:display DESCRIPTION:Poster session: Accountable behavioural change detection (VEDA R) using ML in Poster sessions and BOF track in 5 minutes TRIGGER:-PT5M END:VALARM END:VEVENT BEGIN:VEVENT SUMMARY:The shape of U DTSTART;VALUE=DATE-TIME:20191123T105000Z DTEND;VALUE=DATE-TIME:20191123T112000Z DTSTAMP;VALUE=DATE-TIME:20210227T094116Z UID:session/PWdVmj78ALxvP8SPeZbpUe@hasgeek.com CATEGORIES:Crisp talk,Advanced,Lecture CREATED;VALUE=DATE-TIME:20191027T113617Z DESCRIPTION:In this talk\, we will showcase our efforts at OffNote Labs to improve the developer experience when programming with tensors. In partic ular\, we will discuss:\n\n1. The idea of naming dimensions of tensors and how named shapes can make tensor programming dramatically less painful.\n 2. The tsalib library\, which allows used named dimensions in Python 3.x p rograms with multiple backend libraries (numpy\, tensorflow\, pytorch\, …).\n3. The tsanley library\, which builds on tsalib\, and helps catch t ricky tensor shape errors at runtime and annotate existing programs with n amed shapes.\n\n### Speaker bio\n\nNishant Sinha is an independent researc her and consultant at OffNote Labs\, with broad experience in building dee p learning systems (across text\, vision and speech domains) and symbolic reasoning systems. Nishant helps companies understand and maneuver through the evolving deep learning/AI space and build IP\, in-house teams and sol utions that enable market leadership. He is also passionate about making c utting-edge research consumable and building tools that improve developer experience.\n\nHe received his Ph.D. from Carnegie Mellon University and B . Tech. in Computer Science from IIT Kharagpur.\n GEO:12.973321659788686;77.61947496794164 LAST-MODIFIED;VALUE=DATE-TIME:20200619T062515Z LOCATION:Ballroom - Taj M G Road\, Bangalore\nBangalore\, IN ORGANIZER;CN="Anthill Inside":MAILTO:no-reply@hasgeek.com URL:https://hasgeek.com/anthillinside/2019/schedule/the-shape-of-u-PWdVmj7 8ALxvP8SPeZbpUe BEGIN:VALARM ACTION:display DESCRIPTION:The shape of U in Ballroom in 5 minutes TRIGGER:-PT5M END:VALARM END:VEVENT BEGIN:VEVENT SUMMARY:Poster session: Tensorboard: almost a one-stop shop for ML develop ment DTSTART;VALUE=DATE-TIME:20191123T112000Z DTEND;VALUE=DATE-TIME:20191123T115000Z DTSTAMP;VALUE=DATE-TIME:20210227T094116Z UID:session/NaZyDfogs5LutodcbAbQcJ@hasgeek.com CATEGORIES:Full talk,Beginner,Lecture,Submit revised slides within 7 days, None,Accepted as poster session CREATED;VALUE=DATE-TIME:20191021T104951Z DESCRIPTION:1. Addressing the problems faced while developing machine lear ning models using just terminal for interface.\n2. How some companies have leveraged this pain into making a paid service for monitoring model train ing.\n3. Why tensorboard is better than any other paid service.\n4. Showca se of in-house tools built at Infilect.\n\n### Speaker bio\n\nTushar Pawar is Machine Learning Engineer at Infilect. He has around 3 years of experi ence in the field of Deep Learning. Has worked with several computer visio n problems such as image classification\, object detection\, image generat ion etc.\n GEO:12.973321659788686;77.61947496794164 LAST-MODIFIED;VALUE=DATE-TIME:20191122T061308Z LOCATION:Poster sessions and BOF track - Taj M G Road\, Bangalore\nBangalo re\, IN ORGANIZER;CN="Anthill Inside":MAILTO:no-reply@hasgeek.com URL:https://hasgeek.com/anthillinside/2019/schedule/tensorboard-almost-a-o ne-stop-shop-for-machine-learning-development-NaZyDfogs5LutodcbAbQcJ BEGIN:VALARM ACTION:display DESCRIPTION:Poster session: Tensorboard: almost a one-stop shop for ML dev elopment in Poster sessions and BOF track in 5 minutes TRIGGER:-PT5M END:VALARM END:VEVENT BEGIN:VEVENT SUMMARY:Closing talk: Feast - feature store for Machine Learning DTSTART;VALUE=DATE-TIME:20191123T112000Z DTEND;VALUE=DATE-TIME:20191123T120000Z DTSTAMP;VALUE=DATE-TIME:20210227T094116Z UID:session/DrCW1PvEZ5Q6DYvh27zPR7@hasgeek.com CATEGORIES:Full talk,Beginner,Lecture CREATED;VALUE=DATE-TIME:20190823T050153Z DESCRIPTION:GOJEK\, Indonesia’s first billion-dollar startup\, has seen an explosive growth in both users and data over the past three years. Toda y\, it uses big data-powered machine learning to inform decision making in its ride-hailing\, lifestyle\, logistics\, food delivery\, and payment pr oducts\, from selecting the right driver to dispatch to dynamically settin g prices to serving food recommendations to forecasting real-world events. Hundreds of millions of orders per month\, across 18 products\, are all d riven by machine learning.\n\nFeatures are at the heart of what makes thes e machine learning systems effective. However\, many challenges still exis t in the feature lifecycle. Developing features from big data is often an engineering heavy task\, with challenges in both the scaling of data proce sses and the serving of features in production systems. Teams also face ch allenges in enabling discovery\, reducing duplication\, improving understa nding\, and providing standardization of features throughout organizations .\n\nWillem will explain the need for features at organizations like GOJEK and discuss the challenges faced in creating\, managing\, and serving the m in production. He'll describe how in partnership with Google\, they desi gned and built a feature store called Feast to address these challenges an d explore their motivations\, the lessons they learned along the way\, and the impact the feature store had on GOJEK. Finally\, he will talk about t he open source plans for Feast and their roadmap going forward.\n\n### Spe aker bio\n\nWillem Pienaar leads the data science platform team at GOJEK\, working on the GOJEK ML platform\, which supports a wide variety of model s and handles over 100 million orders every month. His main focus areas ar e building data and ML platforms\, allowing organizations to scale machine learning and drive decision making. In a previous life\, Willem founded a nd sold a networking startup and was a software engineer in industrial con trol systems.\n GEO:12.973321659788686;77.61947496794164 LAST-MODIFIED;VALUE=DATE-TIME:20200619T062515Z LOCATION:Ballroom - Taj M G Road\, Bangalore\nBangalore\, IN ORGANIZER;CN="Anthill Inside":MAILTO:no-reply@hasgeek.com URL:https://hasgeek.com/anthillinside/2019/schedule/feast-feature-store-fo r-machine-learning-DrCW1PvEZ5Q6DYvh27zPR7 BEGIN:VALARM ACTION:display DESCRIPTION:Closing talk: Feast - feature store for Machine Learning in Ba llroom in 5 minutes TRIGGER:-PT5M END:VALARM END:VEVENT END:VCALENDAR