Limitations of AI Systems wrt Explainability, Fairness, Ethics, and Causality

Limitations of AI Systems with respect to Explainability, Interpretability, Fairness, Ethics, and Causality -by Dr. Alok Aggarwal, CEO, Scry Analytics, Inc.

AGENDA:
Pre-register on Zoom:
6:45 Connection to Zoom chat with speaker and organizers
7:00 SFbayACM intro, upcoming events, introduce the speaker
7:10 presentation starts (~90 min with Q&A)

YouTube Live: https://youtu.be/fns8bq-khKs

ABSTRACT:

Today, there are more than fifty domains in which AI systems perform at least as well as humans, and there are hundreds more where they are helping humans in making better decisions. However, current AI systems suffer from several debilitating limitations. For example, even the well-trained deep learning networks are often not robust and can recognize random images having perturbed patterns with over 99% confidence (such as a king penguin or a starfish). Medical doctors make mistakes too but by and large, we trust them. Unfortunately, trust mainly develops over time, but transparency and predictability can help.

Whereas predictability usually requires that these systems do not falter with small perturbations, transparency requires openness, proper explanations, and useful interpretations. The latter requirements are particularly daunting since most AI algorithms act as “black-boxes” and provide no explanations. This lack of “explainability” can be a stumbling block in adopting these systems especially in the fields of health care, legal and criminal, security, defense and military, product liability, and financial services.

In this talk, we first discuss need for explainble AI (XAI) and application domains that generally require XAI. Since explainability may not be easy to come by for current AI systems, we discuss various research efforts for achieving interpretable, causal, fair, and ethical AI systems.

SPEAKER'S BIO:

Dr. Alok Aggarwal received his B. Tech. from the Indian Institute of Technology (IIT) Delhi in 1980 and his Ph. D. in 1984 from Johns Hopkins University in EE and Computer Science. He worked at IBM’s T. J. Watson Research Center in New York during 1984 and 2000. During 1988 and 1989, he took a sabbatical from IBM and to teach two courses at MIT and supervise two Ph.D. students.

During 1992 and 1995, along with other researchers, he built and sold a "Supply Chain Management Solution" for paper mills and steel mills, which was the first commercial AI-based solution of its kind. Aggarwal et. al published a seminal paper titled, “Cooperative Multi-objective Decision Support for the Paper Industry,” which won the Daniel H. Wagner prize for “Excellence in Operations Research Practice” from INFORMS in 1998.

In July 1997, he "founded" IBM’s India Research Laboratory, which he set-up inside the IIT Delhi. By July 2020, he had grown this Laboratory from "ground zero" to a 70-member team (with 35 PhDs and 35 Masters’ in Computer Science and related areas).

In December 2000, he “co-founded” Evalueserve (www.evalueserve.com), which is currently a 4,000-people company that provides various kinds of research and analytics services to clients in North America, Europe, and Asia Pacific from its five research centers in Delhi-Gurgaon, India; Shanghai, China; Cluj, Romania; Santiago-Valparaiso, Chile; and Raleigh, North Carolina.

In 2003, he pioneered the concept of “Knowledge Process Outsourcing (KPO)” and wrote the first article regarding KPO. Today, KPO is a well-known term in the outsourcing industry and is considered a sub-sector on its own with more than 300 KPO companies in India alone.

In February 2014, Dr. Aggarwal founded Scry Analytics (www.scryanalytics.com) a company that codifies different kinds of workflows in several industries and uses artificial intelligence, machine learning, data mining as well as statistical analysis to improve their efficiency with respect to timeliness, quality, revenue, cost, customer experience, compliance, and aggregated risks. This company provides four AI-based products and currently has three R&D centers (in San Jose, California; Gurgaon, India; and Hyderabad, India) with a fourth one to start in 2023 in Dallas, Texas.

During 1997-2000, he was a member of Executive Committee on Information Technology of the Confederation of the Indian Industry (CII) and the Telecom Committee of Federation of Indian Chamber of Commerce and Industry (FICCI). Since 2002, he has been a charter member of The Indus Entrepreneur (TiE) organization and was on its the executive board of its New York chapter during 2002 and 2005.

In 2008, Dr. Aggarwal received Distinguished Alumnus Award from IIT Delhi. He has published 115 research articles, has been granted 8 patents from the US Patents Office, has been an editor of several academic journals in Computer Science, and is writing a book titled, “Hundred Years of Artificial Intelligence: 1950 – 2049,” which will be published by December 2022.