Fordham-IBM Workshop on Generative AI

Fordham-IBM Workshop on Generative AI

Friday, Oct. 27, 2023, 10:30 am-3:00 pm

Walsh Family Library in Flom Auditorium (lower level), Fordham Rose Hill campus
2691 Southern Blvd., Bronx, NY 10458

For directions and a map of the Rose Hill campus.

Sponsored by the Fordham Office of Research

Generative AI presents many applications across a variety of disciplines and plenty of opportunities to transform industries, organizations, and individuals. But it does come with risks and threats. At the Workshop, scientists and practitioners from IBM and Fordham will present some GAI opportunities with cases and some best practices on how to mitigate those risks and threats.

We are looking forward to this significant, exciting, and timely workshop. Thank you. 

PROGRAM

10:30 am-10:40 am    Opening Remarks

Chair: Dr. George Hong (Chief Research Officer and Associate Provost for Research)

  1. Dr. Dennis Jacobs (Provost and Senior Vice President for Academic Affairs)
  2. Radha Ratnaparkhi (T. J. Watson Research Center, IBM)
  3. Dr. Frank Hsu (CIS Dept., Fordham University)

10:40-11:10 am          Keynote Speech Session 

Dr. J. R. Rao (CTO, IBM Security Research, T. J. Watson Research Center, IBM): “The Role of AI in Cybersecurity”

Chair: Dr. Frank Hsu

11:20 am – 12:20 pm             Panel Presentation (I) 

  1. Dr. Mohamed Rahouti (CIS Dept., Fordham University): “Enhancing AI/ML-based DoS Attack Detection Through Generative AI and Combinatorial Fusion Analysis”

Chair: Dr. Pinyu Chen (T.J. Watson Research Center, IBM)

      2. Dr. Aniket Kesari (School of Law, Fordham University): “Trustworthy AI in Law and Society”

Chair: Dr. Yingdong Lu (Mathematical Sciences, IBM Research, IBM)

12:20 pm- 1:20 pm    Lunch Break

1:20 pm- 2:50 pm      Panel Presentation (II) 

  1. Dinesh Verma (IBM Fellow, T. J. Watson Research Center, IBM): “Foundation Models and Generative AI with WatsonX”

Chair: Dr. Ying Mao (CIS Dept., Fordham)

  1. Dr. Navid Asgari (Gabelli School of Business, Fordham University): “Using LLMs for Strategy and Business Economics Research”

Chair: Dr. Juntao Chen (CIS Dept., Fordham)

  1. Dr. Nan Jiang (Laboratory of Informatics and Data Mining, Fordham University): “Generative AI and Multi-Property Drug Discovery Prediction”

Chair: Dr. Wenqi Wei (CIS Dept., Fordham)

Program Committee: 1) Radha Ratnaparkhi (T. J. Watson Research Center, IBM); 2) Frank Hsu (CIS Dept., Fordham University); 3) Divyesh Jadav (IBM Research-Almaden, IBM); 4) Ying Mao (CIS Dept., Fordham University)

Generative AI presents many applications across a variety of disciplines and plenty of opportunities to transform industries, organizations, and individuals. But it does come with risks and threats. At the Workshop, scientists and practitioners from IBM and Fordham will present some GAI opportunities with cases and some best practices on how to mitigate those risks and threats.


Presenters’ Abstracts and Bios

Keynote Speaker: Dr. J.R. Rao, “The Role of AI in Cybersecurity” 

Abstract: Today's Security Operations Center (SOC) analysts and cyber threat hunters invest a significant amount of time in investigating alerts to reconstruct and hunt for potential attacks that are underway. The large volume of alerts to be investigated, multiple reports in threat and vulnerability feeds and the lack of any automated support in processing alerts and consolidating threat intelligence makes the task of security practitioners very challenging. In this talk, we discuss how AI can be used at every stage of the threat management lifecycle (detect, investigate, and respond) to help ease the cognitive burden of security practitioners. Specifically, we show how AI can be used (a) to sift through large volumes of events and alerts to identify anomalous behaviors and accurately detect elusive threats, (b) to consolidate threat intelligence from multiple external security sources, and (c) to automatically corroborate findings by
correlating threat kinetics with threat intelligence, thereby demonstrating how AI can be used to provide much-needed support and raise the bar on security.

Bio: Dr. J.R. Rao is an IBM Fellow and CTO for the Security Research team at IBM. Based at the IBM Thomas. J. Watson Research Center, the global team comprises more than 200 researchers who work in the areas of Cybersecurity, Cloud and Systems Security, AI Security, Information Security and Cryptography. JR works closely with commercial and government customers, academic partners and IBM business units to drive new and innovative technologies into IBM's products and services and definitive industry standards. The goal of his research is to significantly raise the bar on the quality of security while simultaneously easing the overhead of
developing and deploying secure solutions. Dr. Rao has published widely in premier security conferences and workshops and holds numerous US and European patents. Dr. Rao obtained his doctorate degree from the University of Texas at Austin, a Master's degree in Computer Science from the State University of New York at Stony Brook, and a Bachelor of Technology degree
in Electrical Engineering from the Indian Institute of Technology, Kanpur.

 (In Alphabetical Order)

1) Dr. Navid Asgari: “Using LLMs for Strategy and Business Economics Research”

Abstract: I am using LLMs for measurement purposes in three papers. In the first paper, I use OpenAI’s Ada to determine the word embeddings of 310,000 patents of pharmaceutical firms. These word embeddings will be used to measure the textual similarity of the patents and, in turn, the degree of competition among firms. In the second paper, Chat GPT 3.5 is used to interpret scientific papers that report the outcome of clinical trials registered with the National Institute of Health (NIH). Finally, the third paper uses training data influence (TDI) algorithms applied to LLMs to measure organizational knowledge decay.

Bio: Navid Asgari, Ph.D., is an associate professor and the Grose Family Endowed Chair at the Strategy Area of the Gabelli School of Business. Dr. Asgari also serves as the associate director of the Global Healthcare Innovation Management Center at Fordham University. In his research, he examines firms’ innovation strategies in the context of the healthcare industry. Some of his teaching courses include “Machine Learning for Strategic Decision Making,” “Innovation and Change Management,” and “Strategy.” As a Fordham-IBM Fellow, he is working on using LLMs to extract information from scientific corpora. 

2) Dr. Nan Jiang, “Generative AI and Multi-Property Drug Discovery Prediction”

Abstract: Various generative AI (GAI) models/platforms such as GCNNs, GANs, DRL, GENTRL, REINVENT, and Torch Drug have been used since 2018 in multi-property drug discovery prediction at industrial scale. However, these existing models often lack robustness and extensibility. In this talk, we focus on Combinatorial Fusion Analysis (CFA) which provides a GAI and multi-property platform for pretraining, generating, evaluating, and optimizing new and diverse models. CFA-optimized generative models show superior performance compared to most of the state-of-the-art individual ADMET-property prediction models on the Therapeutics Data Commons (TDC) (https://tdcommons.ai/) leaderboard w.r.t. the 22 ADMET benchmark datasets.

Bio: Dr. Nan Jiang is an affiliated member of the Laboratory of Informatics and Data Mining (LIDM) at Fordham University and a quantitative analyst at Obex group LLC. His research interests are in financial and health informatics using computational/statistical methods. In particular, he has focused on the application of AI/ML and model fusion to multi-property optimization, equity ranking, and drug discovery. Nan received his M.S. from Valparaiso University and Ph.D. from Fordham University, both in Economics. 

3) Dr. Aniket Kesari, “Trustworthy AI in Law and Society”

Abstract: What role will the law play in promoting trustworthy AI? Research on “fairness in machine learning” often uses legal definitions to measure bias and fairness. Regulators play an important role in promoting the adoption of privacy-enhancing technologies. Federal, state, and local governments maintain vast troves of data and are among the biggest customers of AI tools. Generative AI promises to transform law and the legal profession, with applications such as making consumer contracts more readable, summarizing court opinions, and even giving legal advice. But AI also comes with risks – entrenching racial biases in credit lending, denying people vital Medicaid benefits, and degrading privacy. How should the law adapt to AI technology, and ensure that AI achieves its promise while avoiding its potential pitfalls?  

Bio: Dr. Aniket Kesari is an Associate Professor at Fordham Law School. His research focuses on law & technology, data science, and public policy. He uses techniques drawn from causal inference, machine learning, and natural language processing to investigate questions in law & tech, and he is also interested in integrating data science into empirical legal studies more broadly. Some of his recent scholarship covers data breach notification laws, mandatory cybersecurity risk disclosures, privacy and algorithmic fairness, trademark search engines, and online hate speech. In ongoing work, he is conducting computational audiovisual analysis on courtroom videos, using large language models to summarize legal documents, and developing toolkits for effective AI explanations. 

4) Dr. Mohamed Rahouti, “Enhancing AI/ML-based DoS Attack Detection Through Generative AI and Combinatorial Fusion Analysis”

Abstract: Mitigating Denial-of-Service (DoS) attacks is vital for online service security and availability. While AI and ML methods have been used for DoS attack detection, new strategies are needed to enhance their performance w. r. t. generalizability and robustness. Combinatorial Fusion Analysis (CFA) combines multiple pre-trained AI/ML models using advanced model fusion algorithms. On the other hand, Generative AI (GAI) models like GANs, VAEs, and GPTs can augment data to its limited or imbalanced list and generate new features to enrich data representation. The union of GAI/CFA would provide a robust and sustainable platform for the DoS mitigation system. This talk will present recent advances in intrusion detection using GAI/CFA synergy.

Bio: Dr. Mohamed Rahouti received an M.S. degree in Mathematics (Statistics Concentration) and a Ph.D. degree in Electrical and Computer Engineering, both from the University of South Florida (Tampa, FL). He is currently an Assistant Professor in the Department of Computer and Information Science at Fordham University in New York City. His research interest focuses on blockchain technology, computer networking, machine learning, and network security with applications to smart cities. Dr. Rahouti has authored/co-authored over 40 peer-reviewed journals/conference papers and is a member of the IEEE Computer and Communications Societies.

5) Dinesh Verma, “Foundation Models and Generative AI with WatsonX”

Abstract: The talk will cover the value of Generative AI to Enterprises and also highlight some of the ethical challenges it raises.

Bio:  Dinesh C. Verma is an IEEE Fellow, IBM Fellow, AAIA Fellow and Fellow of UK Royal Academy of Engineering. He has authored 11 books, 200+ technical papers and 200+ U.S. patents. He has chaired/vice-chaired IEEE technical committee on computer communications and vice-chaired IEEE Internet technical committee. He has led several multi-national multi-organizational research programs. He has contributed to several IBM products and service offerings including significant contributions to server networking stack, network management products and cellular network analytics. He is currently leading exploratory science projects in the area of futuristic networks and distributed systems at IBM Research.

#################################################