Artificial Intelligence Institute of South Carolina

While several students from my group have not returned from AAAI or are dealing with the aftermath of high winds and rain in the morning, a small number of us have resumed our campus walk (which we do about twice a week) now that we have spring-like weather.

The photos show 16 of ~20 #AIISC researchers attending #AAAI2025 who could gather in one place at one time! This may be the largest participation at AAAI from any academic organization. We are engaged in many activities: giving tutorials/labs, organizing workshops, presenting papers, demos, posters,.... Activities during the first half of the conference: https://lnkd.in/gUkv2GVF Here are our upcoming activities Mon, 3 Mar QC+AI Workshop - Quantum Neural Networks – 2:45 pm - 2:58 pm: [Full] Erik Connerty, Ethan Evans, Gerasimos Angelatos and Vignesh Narayanan. Predicting Chaotic Systems with Quantum Echo-state Networks AI for Public Missions Workshop - Disseminating Authentic Public Messages using Chatbots - A Case Study with ElectionBot-SC to Understand and Compare Chatbot Behavior for Safe Election Information in South Carolina - Nitin Gupta, Vansh Nagpal, Biplav Srivastava. Tue, 4 Mar Defactify Workshop organized by Dr. Amitava Das Amit Sheth Aman Chadha Aishwarya Naresh Reganti Vikas Sharma Vinija Jain https://defactify.com/ Room 120A AAAI2025 Anomaly Detection in Scientific Domains Workshop: Time Series Foundational Models: Their Role in Anomaly Detection and Prediction Chathurangi Shyalika, Harleen Kaur Bagga, Ahan B., Renjith Prasad, Alaa AL Ghazo, Amit Sheth Generalization in Planning -? On Generalized Planning for Controlling Opinion Networks: Interpreting Human-AI Dialog States and Beliefs (12 PM - 12:30 PM) Bharath Chandra Muppasani, Protik Nag, Biplav Srivastava, Vignesh Narayanan Towards Learning Foundation Models for Heuristic Functions to Solve Pathfinding Problems [Poster (3:00 PM to 3:30 PM) and Contributed talk (2:50 PM - 3:00 PM)]? Vedant K., Amit Sheth, Forest Agostinelli PDDLFuse: A Tool for Generating Diverse Planning Domains [Poster (12:00 PM to 12:30 PM)] Vedant K., Amit Sheth, Forest Agostinelli

AIISC PhD students are organizing quite a few workshops this year. Here is the one Utkarshani Jaimini is organizing at #ESWC. https://lnkd.in/edFGhQHA Utkarshani also has extensive work in this area: https://lnkd.in/eDhExMkM Some more workshops will be listed at: https://lnkd.in/egqh-QpZ

Around twenty of us from the #AIISC are at AAAI25 in Philadelphia. Here is a partial list of our engagements (Part 1 of 2). ==== Tuesday, 25 Feb *AAAI Lab- 10:45am - 12:30pm, SOFAI Lab: A Hands-On Guide to Building Neurosymbolic Systems with Metacognitive Control [Room 115C] - Vishal Pallagani, Andrea Loreggia, Francesco Fabiano, Biplav Srivastava, Francesca Rossi, Lior Horesh *AAAI Lab- 4:15pm – 6:00pm, Developing explainable multimodal AI models with hands-on lab on the life-cycle of rare event prediction in manufacturing. [Room 118A]: Chathurangi Shyalika, Ruwan Wickramarachchi, Revathy Venkataramanan, Dhaval Patel, Amit Sheth.?https://lnkd.in/eEtNqF25 https://lnkd.in/eXjwBE5t *2:00 – 6:00 pm Tutorial: #Neurosymbolic AI for EGI. Deepa Tilwani, Seyed Ali Mohammadi, Edward Raff, Iman Azimi, Aman Chadha Chada, Manas Gaur. Room # 118C Wed, 26 Feb *Tutorial: Hallucinations in Large Multimodal Models. 8:30 – 10:15 am.? Vipula Rawte, Aman Chadha, Amit Sheth, Dr. Amitava Das Das Room # 113 https://lnkd.in/eChPctCe https://lnkd.in/eXDjMQTH *AAAAI 2025 Bridge on Constraint Programming and Machine Learning - 11:00 AM - 11:15 AM : A Neurosymbolic Fast and Slow Architecture for Graph Coloring (Vishal Pallagani, Vedant K., Biplav Srivastava, Francesca Rossi) *B3: Bridging Cognitive Science and AI to Bridge Neuro and Symbolic AI - 9 am-5 pm; [Room: 125] (Vishal Pallagani, Biplav Srivastava) *B4: Bridging Planning and Reasoning in Natural Languages with Foundational Models (PLAN-FM) - 9am-5pm ; [Room: 120C] Biplav Srivastava Thu, 27 Feb Blue Sky/ SM Presentation: "A Vision for Reinventing Credible Elections with Artificial Intelligence" - Biplav Srivastava, Time: 2:30-2:50pm Fri, 28 Feb *Main Paper Poster Presentation: Can LLMs Obfuscate Code? A Systematic Analysis of Large Language Models into Assembly Code Obfuscation. Seyedreza Mohseni, Seyedali Mohammadi, Deepa Tilwani, Yash Saxena, Gerald Ketu Ndawula, Sriram Vema, Edward Raff, Manas Gaur.? *Student Poster Presentation: Towards Enhancing Road Safety in South Carolina Using Insights from Traffic and Driver-Education Data - 12:30-2:30pm, Nitin Gupta, Bharath Chandra Muppasani, Biplav Srivastava Sat, 1 Mar AAAI Demo-Pic2Prep: A Multimodal Conversational? Agent for Cooking Assistance, Renjith Prasad, Chathurangi Shyalika, Revathy Venkataramanan, Darssan LE , Amit Sheth https://lnkd.in/enTU6SBJ Amit Sheth will be part of several events/activities involving AAAI Fellows. At the AAAI Fellows dinner: https://lnkd.in/eQZ6JJ9J More on this in the next post.

C^3AN: Custom, Compact and Composite AI Systems - A NeuroSymbolic Approach: 4th-Generation Evolution of Intelligent Systems. Check out our latest thoughts (and ongoing work) on what is next for hashtag #AI to support mission-critical, high-value hashtag #EnterpriseAI applications: 1: from #LLMs to #custom focused/ #targeted, #Agile and #neurosymbolic AI models that rely not on just lots of data but extensive knowledge and human expertise and can be trained at a fraction of the cost, 2: from AI models to #composite_AI systems, 3. do much more to make AI systems #intelligent, #robust and #trustworthy by better support for fourteen foundation elements: #reliability, #consistency, #alignment, #causality, #analogy, #abstraction, #reasoning, #planning, #attribution, #interpretability, #explainability, #safety. 4. Address this: A month or two ago, Gartner and, separately, Daron Acemoglou (2024 Nobel Prize in Economics) said- while the current (Gen)AI addresses 5-10% of the problems, 50% of the AI will be "Enterprise AI" (with little role for LLMs) by 2027. Amit Sheth Kaushik Roy Revathy Venkataramanan @Venkatesh N and our Group at?#AIISC https://lnkd.in/eqvbkDMj

Our upcoming participation (tutorial, demo) at the #AAMAS25 (A*) on the timely and vital topic of multiagent coPilots in Industrial Applications (#SmartManufacturing). Several members of our team and collaborators are expected to attend/present: Chathurangi Shyalika Renjith Prasad Fadi El Kalach (Clemson) Cory Henson (Bosch)

Congratulations to Vipula Rawte Dr. Amitava Das, and all coauthors of this innovative and impactful work https://lnkd.in/eBa86tjy (Swagata Chakraborty Agnibh Pathak Anubhav Sarkar S.M Towhidul Islam Tonmoy Aman Chadha Amit Sheth) for this coverage in the #Nature: "Just how bad is the hallucination problem? Researchers have developed a variety of metrics to track the issue. Vipula Rawte, who is doing her PhD in hallucinatory AI behaviours at the University of South Carolina in Columbia, for example, has helped to create a Hallucination Vulnerability Index, which sorts hallucinations into six categories and three degrees of severity.?" Earlier, this work was at the center of a piece in #WaPo: https://lnkd.in/eMYUWRY8 .

Led by Revathy Venkataramanan, our collaboration with Hewlett Packard Enterprise labs resulted in this exciting paper. It is significant because it uses a comprehensive metadata knowledge graph for improving end-to-end pipelines by recommending contextually relevant AI pipelines, which can act as a seed for experimentation.

Our Utkarshani Jaimini is the lead organizer of this very exciting workshop on Causal AI. Also, check out a seminar Utkarshani gave on "Industrial Applications of Causal AI" that the audience loved: https://lnkd.in/eFizMc67