Vertical: V2 – Artificial Intelligence and Machine Learning (AI & ML)

Notice Board -

Indian researchers/scientists/academician/Ph.D. students may send an email at vaibhav.aiml@iisc.ac.in, vaibhav.aiml@ai.iith.ac.in to attend session(s) of V2-AIML vertical with session ID as provided in Session Schedule. You will get a link to attend the session on your email after approval.

   Horizontals

V2H1 Foundations of AI

The science behind AI stands on the pillars of Statistical Learning Theory (SLT) and Mathematical Optimization (MO). The classical paradigms of AI, namely Learning with Supervision, Unsupervised Learning, and Reinforcement Learning, heavily draw upon both SLT and MO. The success of AI in recent times has much to do with the invention of Deep Learning (DL). Existing formalisms are often not adequate to explain the predictive power of DL and it also do poorly in understanding the design of DL models. As AI gets implemented in the field, one is confronted with new questions giving rise to many new areas such as (I) Adversarial Learning (2) Edge AI and (3) Fairness, Accountability, Transparency and Ethics (FATE). A broad question which confronts the area of AI is: “Are existing foundations adequate to deal with the rise of AI technologies?” To enable extensive discussion on this question, we have envisaged two panels where one panel will be exclusively devoted to FATE. The first panel named “Mathematical Foundations of AI” would engage in :

  • Identifying research opportunities in SLT and MO with a focus on identifying research opportunities in classical paradigms of Supervised Learning, Unsupervised Learning, and Reinforcement Learning.
  • Discussing foundational questions in new areas such as Edge AI, and Adversarial Machine learning. It should throw light on other emerging areas which may require foundational research.
  • Discussion of research schemes to build capacity in Indian Universities at the post-graduate level.

The second panel named “FATE in AI” would engage in

  • Discussing the state of the art in the Theory and practice of FATE in AI
  • Role of Industry in facilitating Data Access for research in FATE
  • Discussing Academic Curricula suited for FATE

Leading Universities around the Globe are formulating research initiatives which aim to examine and enhance the existing state of the art in Mathematical Foundations of AI. Apart from the traditional communities of SLT and MO, these initiatives are trying to bring together diverse expertise such as Statistical Mechanics, Theoretical Computer Science, Bayesian Non-parametrics, Numerical Analysis etc. FATE is an emerging area and it is believed that as AI technologies pervade the common citizens life the importance of FATE will only increase in the coming days. Universities have taken note of these challenges and are trying to create research opportunities in this area. India has strength in Mathematical Foundations of AI. Every year Indian Universities contribute several publications in leading journals and conferences on the Foundation areas of AI. FATE is a comparatively younger area and India’s contribution is at the nascent stage. This January the first workshop in India was held in IISc. It brought together researchers from several Indian Universities and also speakers from some of the leading universities in the world.


V2H2 AI and Signals

One of the horizontals/sub-themes identified in the AI/ML vertical is termed as “AI and Signals”, which in turn includes tracks on: (i) Computer Vision; (ii) Natural Language Processing; and (iii) Speech Understanding - each of which is a field by itself today, and will have a dedicated session in the summit, involving top researchers from India and of Indian origin living abroad. Over the last decade, since the advent of deep learning, these abovementioned areas have seen significant success in various domains including healthcare, education, energy, security, e-commerce, retail, finance, and many more. Many real-world applications in these domains demand the analysis of images/videos, text/documents and speech/music, and the growth of the industry in multinational corporations has relied on early adoption of advancements in the areas. Progress in these areas, along with the availability of large-scale compute and large-scale data, has also resulted in the creation of niche industry players in the high-end technology space (e.g. OpenAI, FiveAI, etc), and the exponential growth of startups in the use of these technologies for specialized domains such as pathology (e.g. PathAI), self-driving cars (e.g. Waymo) or virtual reality (e.g. Argo). Importantly, the areas of understanding images, text and speech are at a unique position where research in academia and industry is closely intertwined, leading to a direct impact of research on development of products, and thus, raising the quality of life of humanity at large. It is timely and imperative that India nurtures an ecosystem that allows these areas to flourish and build to a status of self-sufficiency and global leadership, over the next few years.

Owing to the success of contemporary AI/ML methods in understanding of images, speech and text, India has a strong presence, interest and demand for research in areas of computer vision, natural language processing and speech understanding - across academia, industry and government organizations. Many research groups at institutions around the country including IITs and IISc publish their research at top-tier venues in these areas such as CVPR, ICCV, ACL, EMNLP, ICASSP, InterSpeech, etc. Many of the newly formed Technology Innovation Hubs (TIH), as part of the National Mission on Interdisciplinary Cyber Physical Systems (NM-ICPS), are closely associated with research in this horizontal. Faculty and researchers working in these areas, also have close collaborations with the industry on problems leading to product development and deployment. Despite the presence of such groups, there is a gap between demand and supply in these areas. There is a strong demand for formal coursework/training in areas such as computer vision, natural language processing and speech understanding - both in academic institutions across the country, as well as working professionals that seek to upskill themselves. This is showcased in the registrations of NPTEL courses offered in these areas (of the order of 1000s), which easily outnumber registrations in other areas. Industry players (both multinational corporations) as well as startups are constantly looking for skilled practitioners in these fields, as well as guidance from experts to be competitive at the global level. Government research organizations such as DRDO and ISRO have also been leveraging technological advancements in these areas, but need constant exposure and collaboration at times, to be in touch with a field that has been growing rapidly, with newer algorithms and methods being developed each day. Addressing this gap between demand and supply in this theme would be a key requirement.


V2H3 AI and Robotics

The need for Robotics and Autonomous Systems has never been felt more than in the present Covid-19 Pandemic era. According to Tractica Research, the need to automate physical processes and merge it with the digital ecosystem is going to result in a $248.5 bn market by 2025 comprising Personal robots, Commercial robots, Industrial robots and Military robots. Autonomous/Semi-Autonomous mobility of vehicles in warehouses, mines, ports, campuses etc. has the potential of tremendous economic impact. These spaces are unstructured, yet restricted in terms of overall traffic and hence offer a possibility of development of robust solutions based on latest advances in Sensors and AI. In addition, air mobility of micro and mini-air vehicles opens up new societal and business applications, while it has its own different set of technological and regulatory challenges. Interaction involves machines manipulating their environment - e.g. harvesting or de-weeding in agriculture, picking and placing of objects in warehouses, co-working with a factory worker to provide tools and other support etc.

The need for Robotics and Autonomous Systems has never been felt more than in the present Covid-19 Pandemic era. According to Tractica Research, the need to automate physical processes and merge it with the digital ecosystem is going to result in a $248.5 bn market by 2025 comprising Personal robots, Commercial robots, Industrial robots and Military robots. Autonomous/Semi-Autonomous mobility of vehicles in warehouses, mines, ports, campuses etc. has the potential of tremendous economic impact. These spaces are unstructured, yet restricted in terms of overall traffic and hence offer a possibility of development of robust solutions based on latest advances in Sensors and AI. In addition, air mobility of micro and mini-air vehicles opens up new societal and business applications, while it has its own different set of technological and regulatory challenges. Interaction involves machines manipulating their environment - e.g. harvesting or deweeding in agriculture, picking and placing of objects in warehouses, co-working with a factory worker to provide tools and other support etc. Creating technologies that are safe and allow for creating a diverse set of interactions needing various degrees of skilling, is a key research challenge. Another perspective for robots comes from thinking about their physical dimensions. This can range from mega (electronically controlled tractors, trucks, cranes), to macro (AGVs, electronically controlled forklifts), to human (service robots, electronically controlled wheelchairs, Assistive Devices), mini (household robots, MAVs) and finally to micro scale (microbots). The solution approaches to mobility and interaction will be fundamentally different for micro-bots, compared to that at other scales. Micro bots offer promising applications in healthcare, especially for surgery, targeted drug delivery etc. Achieving cyber-physical autonomy for these machines in a safe, secure & reliable manner is the key scientific and engineering challenge, and the main intellectual and technological focus of our hub. Successful development of technologies in this domain, requires a truly interdisciplinary and collaborative approach, cutting across these diverse areas, backed by world class experimental facilities.

Robotics research in India is in a very early phase compared to the rest of the world. There are only a few academic researchers throughout the Indian Academic system - and these are spread across a large number of institutions - with no institution having a critical mass. There are no graduate academic programs in Robotics. We have one DRDO laboratory devoted to AI & Robotics - which is Centre for AI & Robotics. A part of the problem is that Robotics research requires significant capex - though this is much less than for a nano fab, that has been set up in multiple academic institutions. Another reason for the relative lack of development in Robotics in Indian Academia is that it is truly an interdisciplinary area - needing Mechanical, Electronics and Computer Science to come together and we have systemic issues in supporting such activities at scale. The Department of Science & Technology has taken cognizance of this fact and has given funding to set up two innovation hubs - one in IISc Bangalore and the other in IIT Delhi. On the industry front - things are a bit better - we do have quite a few startups now in this space - example include systemantics, rightbot etc. In summary - we need to do much more to systematically inculcate AI & Robotics related teaching, research, technology innovations and entrepreneurship.


V2H4 AI For Social Good  

AI has become a ubiquitous part of our daily lives, and this rise in AI applications has stimulated significant interest from the public, media and policy makers. This increasing attention is often focused on the negative consequences of AI, often overlooking the societal benefits that AI can deliver. This workshop on “AI and Social Good” will focus on that missing perspective: how AI can help solve difficult societal challenges, in assisting low resource communities, public health and welfare, public safety and security, conservation, and governance.

AI can hugely benefit a range of areas such as policy formation, effective implementation of Government programs, urban transportation, agriculture, faster delivery of justice, prevention of crime and cross-border terrorism, prevention and avoidance of epidemics, disaster mitigation and management, better management of healthcare, prevention of child labour, poverty alleviation, protection of critical national resources, etc.

Clearly, AI has the potential to transform various aspects of businesses, life, and governance, and to offer solutions to societal problems. This has spurred many countries to invest massively in AI research and development. China is a frontrunner in AI investment. Its investments in AI are reported to be 25 times that of India. In the United States, many leading universities such as MIT and UC-Berkeley have launched major initiatives in AI and many of these initiatives have societal problems at the core of their agenda. The number of papers in leading AIML conferences on topics related to AI for social good has dramatically increased. India is yet to catch up on this front.


Session Schedule

Horizontal

Session ID

Name

University/Organisation

Country

Foundation of AIML

V2H1S1

Mathematical Foundations

5/10/20

6:30 - 9 PM(IST)

Dr. Pradeep Ravikumar

CMU

United States of America

Dr. Shipra Agarwal

Columbia University

United States of America

Dr. Sanjoy Dasgupta

UCSD

United States of America

Prof. Svetha Venkatesh

Deakin University

Australia

Prof. DevdattDubhashi

Chalmers University

Sweden

Prof. Arindam Banerjee

Univ. of Minnesota

United States of America

Prof. Prasad Tadepalli

Oregon State University

United States of America

Mr. Ujjwal Das Gupta

Apple

United States of America

Dr. Shikhar Vashishth

CMU

United States of America

Prof. Chiranjib Bhattacharyya

IISc

India

Prof. Sunita Sarawagi

IITB

India

Prof. Ravindran Balaraman

IIT Madras

India

Dr. Shivaram Kalyanakrishnan

IIT Bombay

India

Dr. Prateek Jain

Microsooft Research

India

Prof. Aditya Gopalan

IISc

India

Prof. A. Dukkipatti

IISc

India

Prof. S. Bhatnagar

IISc

India

Dr. Harish Guruprasad

IIT Madras

India

V2H1S2

FATE (Fairness, Accountability, Transparency, and Ethics

17-10-2020

6:30-8:30pm (IST)

Dr. Krishna Gummadi

Max Planck Institute

Germany

Prof. Sampath Kannan

U Penn

United States of America

Dr. HimabinduLakkaraju

Harvard Univ

United States of America

Dr. Pradeep Ravikumar

CMU

United States of America

Prof. Arnab Bhattacharyya

NUS

Singapore

Prof. DevdattDubhashi

Chalmers

Sweden

Dr. Shubro Das

IBM

United States of America

Prof. Chiranjib Bhattacharyya

IISc

India

Ms. Srijoni Sen

National Law School

India

Prof. Siddharth Barman

IISc

India

Prof. NiloyGanguly

IITKGP

India

Prof. Vineeth Balasubramanian

IITH

India

Dr. Amit Deshpande

MSRI

India

Mr. Amit Sharma

MSRI

India

AIML and Signals

V2H2S1

Speech Understanding

17-10-2020

8-11pm (IST)

Prof. Raj Reddy

CMU

United States of America

Dr. Tara Sainath

Google

United States of America

Dr. Sanjeev Khudanpur

Johns Hopkins Univ

United States of America

Prof. Shri Narayanan

USC

United States of America

Dr. Mathew Magimai Doss

IDIAP

Switzerland

Dr. BhuvanaRamabhadran

Google

United States of America

Dr. Prem Natarajan

Amazon

United States of America

Dr. Kishore Prahallad

Apple

United States of America

Prof. Yegna

IIIT Hyderabad

India

Prof. S. Umesh

IITM

India

Prof. Rajesh Hegde

IITK

India

Dr. S. R. Mahadeva Prasanna

IIT Dharwad

India

Dr. Preethi Jyothi

IIT Bombay

India

Dr. K Sri Rama Murty

IIT Hyderabad

India

Prof. Hema Murthy

IIT Madras

India

Dr. Sunayana Sitaram

Microsoft Research India

India

Prof. Thippur V. Sreenivas

IISc

India

Prof. V. Ramasubramanian

IIIT-Bangalore (IIITB)

India

Prof. Chandra Sekhar Seelamantula

IISc

India

Prof. Sriram Ganapathy

IISc

India

Prof. Prasanta Kumar Ghosh

IISc

India

V2H2S2

Computer Vision

14/10/2020

8 - 11 PM (IST)

Dr. Ramesh Raskar

MIT

United States of America

Dr. Pulkit Agarwal

MIT

United States of America

Prof. Mohan Kankanahalli

NUS

Singapore

Dr. KarteekAlahari

INRIA

France

Dr. Vinay Namboodiri

Univ of Bath

United Kingdom

Dr. VenuGovindaraju

SUNY Buffalo

United States of America

Dr. Dinesh Jayaraman

U Penn

United States of America

Prof. Rama Chellappa

Johns Hopkins Univ

United States of America

Prof. Narendra Ahuja

UIUC

United States of America

Prof Amit Roy Chowdhury

Univ of California Riverside

United States of America

Dr. Santanu Chaudhury

IIT-Jodhpur

India

Prof. Subhashis Chaudhuri

IIT-Bombay

India

Prof. P J Narayanan

IIIT-Hyderabad

India

Dr. Venkatesh Babu

IISc

India

Dr. Chetan Arora

IIT-Delhi

India

Prof. Bhabhatosh Chanda

ISI, Kolkata

India

Dr. C V Jawahar

IIIT-Hyderabad

India

Prof Parag Chaudhuri

IIT-Bombay

India

Prof Uma Mudengudi

KLE University, Karnataka

India

Dr. Vineeth N Balasubramanian

IIT-Hyderabad

India

V2H2S3

Natural Language Processing

19/10/2020

8 - 11 PM (IST)

Dr. Avirup Sil

IBM Research

United States of America

Dr. Mohit Bansal

University of North Carolina Chapel Hill

United States of America

Dr. Abhijit Mishra

Apple inc.,

United States of America

Dr. Siva Reddy

McGill University

Canada

Dr. Dipanjan Das

Google

United States of America

Dr. Srinivas Bangalore

Interactions LLC.

United States of America

Dr. Aditya Joshi

CSIRO

Australia

Dr. Sameer Pradhan

LDC@UPenn and cemantix

United States of America

Dr. Mitesh Khapra

IITM

India

Dr. Anoop Kunchukuttan

Microsoft Hyderabad

India

Prof. Dipti Mishra

IIITH

India

Dr. Karthik Sankaranarayanan

IBM

India

Mr. Girish Palshikar

TCS

India

Dr. Srijith P K

IITH

India

Prof. Sudeshna Sarkar

IIT Kharagpur

India

Prof. Pushpak Bhattacharyya

IITB

India

Dr. Partha Talukdar

IISc

India

Dr. Asif Eqbal

IIT Patna

India

Dr. MonojitChoudharyy

Microsoft Research

India

Dr. MaunendraDesarkar

IITH

India

Dr. Mausam

IITD

India

AI and Robotics

V2H3S1

AI and Robotics

22/10/2020

8-11pm (IST)

Prof. Gaurav Sukhatme

University of Southern California

United States of America

Dr. Dileep George

Vicarious

United States of America

Mr. Ujjwal Das Gupta

Apple

United States of America

Dr. Pulkit Agarwal

MIT

United States of America

Dr. Amit Kumar Pandey

Hanson Robotics

France

Dr. Shishir N K

IISc

India

Dr. Suresh Sundaram

IISc

India

Prof. Bharadwaj Amrutur

IISc

India

Dr. Madhav Rao

IIIT-Bangalore (IIITB)

India

Dr. G. Athithan

DRDO-CAIR

India

AI for Social Good

V2H4S1

AI for Social Good

9/10/20

8-11pm (IST)

Prof. Milind Tambe

Harvard University

United States of America

Prof. Madhav Marathe

Univ of Virginia

USA

Prof. Anil Vullikanti

Univ of Virginia

USA

Prof. Viktor Prasanna

Univ of Southern California

USA

Dr. Arnab Bhattacharyya

NUS

Singapore

Prof. Ashish Goel

Stanford Univ

USA

Prof. Jay Sethuraman

Columbia Univ

USA

Prof. Ramayya Krishnan

CMU

USA

Dr. Amulya Yadav

Penn State

USA

Dr. Gita Sukhthakar

Univ of Central Florida

USA

Dr. Kaustubh Patil

FZ JUELICH

Germany

Dr. ArshaNagrani

Univ of Oxford

UK

Dr. Nithya Sambasivan

Google

USA

Dr. Ram Sriram

National Institute of Standards and Technology (NIST)

United States of America

Prof. Y. Narahari

IISc

India

Prof. P.P. Chakrabarti

IIT Kharagpur

India

Dr. Shweta Jain

IIT Ropar

India

Prof. Vijay Chandru

IISc

India

Prof. P.J. Narayanan

IIIT Hyderabad

India

Dr. Rajendra Kumar

MeitY, Govt. of India

India

Dr. B.K. Murthy

MeitY, Govt. of India

India

Prof. T. K. Srikant

IIIT-Bangalore

India

Prof. Anirban Dasgupta

IIT Gandhinagar

India