KnowDis Machine Learning Day 2021 – 3rd Edition on December 3, 2021

New Delhi (India), December 03: KnowDis and School of AI, IIT Delhi, have jointly organized the third edition of the online conference that looks to bring together leading experts from all over the world for a meeting of the minds. Hearing these discussions will stimulate the minds of young researchers to push the frontiers of AI even further and help India to become a leader in this field, a small step from our end to live up to the dream of our PM Shri Narendra Modi for Digital India.

 Highlights of the Conference

The conference started with Invoking blessings of Ganesh and Saraswati, shlok rendition by Shrinvias Joshi and Viraj Joshi, Son and grandson of the legendary Pt. Bhimsen Joshi.

 The KnowDis Machine Learning Award has been presented to Prof. Partha Talukdar of IISc Bangalore, and he spoke about: Recent advances in Multilingual Representation Learning an opportunity to transfer knowledge and supervision from high web-resource languages to languages with lower web resources.

The Keynote address was delivered by Dr Shailesh Kumar, Chief Data Scientist at Reliance Jio, and he shared insights about Ashtaang  AI- an eight-layer stack to form the building blocks of an “AI Architecture” for Complex ecosystems of the future – Telecom, Retail, Manufacturing, Smart Cities

Dr. Manish Gupta is the head of Google AI India, and he talked about how AI begins to impact various aspects of our lives. There is a need to combine it with a better understanding of the drivers of human behavior.

Mr. Saurabh Singal, CEO of KnowDis talked about the use of AI in Drug Discovery. He shared insights that there are many rare diseases that impact less than one person in one million. There are over 5,000 of such diseases, and because of this rareness, even a very experienced physician might not have seen a single patient in his lifetime. AI-based applications can not only retrieve such symptoms but also perform meaningful diagnostics. He also described the role of AI in the discovery of new drug molecules.

Mr. Mark Moraes, Head of Engineering, D. E. Shaw Research, described their massively parallel supercomputer Anton 3,  which achieves simulation speeds at least 100-fold faster than the fastest general-purpose supercomputers. This supercomputer enables research in drug discovery.

 Dr. Vikram Mulligan, Research Scientist at the Flatiron Institute, New York, talked about Practical Quantum Computing for Peptide Drug Discovery

Dr Avantika Lal, Senior Scientist at Insitro talked about how Deep learning models can be used to learn patterns in epigenomic data from human cells in order to prioritize sequence variants and propose biological mechanisms linking these variants to an individual’s risk of disease.

 There was a special Panel Discussion on AI for chess. This Panel discussion featured GM Parimarjan Negi, Chess Master and PhD student at Computer Science & AI Lab, MIT.

Along with WIM, Natasha Regan, an award-winning chess author, is best known for her book Game Changer.  This session was moderated by Saurabh Singal@Knowdis. It discussed how Reinforcement Learning based chess engines are proving to be more powerful than traditional chess engines based on alpha-beta search.

KnowDis leverages Artificial Intelligence to build applications for E-commerce, Advertising, Fashion Industry and Pharma.

KnowDis announced Business Ready AI-based APIs which can be customized for business-specific requirements:

1.   KnowDis Translator is currently available in 7 indian languages

Have the advantage of communicating with everyone. Translate your business catalogues/chats/anything from English to any Indian Language* and vice-a-versa.

2. Text to Image Search

Get rid of tag-based image search, and design your own search engine that instantly searches images to the most delicate details.

3. Reverse Image Search

– Because “An image is worth a thousand words”.

– One express what you are searching for through Image as a Query.

– ImSim fetches you all the similar images.

AI solutions for Fashion Industry

1. Robe: Each new design needs a model, and photoshoots are time-consuming and costly. Wouldn’t it be great if you could use your favourite previous models without having to invite them to your studio? Robe makes this possible
2. MICA: Leverage the power of image classification with Mica, pass your images and categories, let Mica do its magic.
3. FIBre: Remove background using AI – Fast and Highly Accurate

AI Solutions for E-commerce Industry

1. Product Search & Recommendation Engine
2. Query Correction and Understanding Module