AGH Auditorium

“Can Generative AI Deepen Our Own Thinking? Supporting Appropriate Reliance, Human Agency, and Beyond”

Jenn Wortman Vaughan (GEng’06, Gr’09)
Senior Principal Research Manager
Microsoft Research

AGH Auditorium

AGH Lobby

10:30 to 10:45 a.m.: A Fast, Reliable, and Secure Programming Language for LLM Agents
Osbert Bastani, Computer and Information Science (Penn Engineering)

10:45 to 11 a.m.: Physics-Aware AI: Building Foundation Models for Scientific Simulation
Paris Perdikaris, Mechanical Engineering and Applied Mechanics (Penn Engineering)

11 to 11:15 a.m.: What Information Do Robot Learners Need?
Dinesh Jayaraman, Computer and Information Science (Penn Engineering)

11:15 to 11:30 a.m.: MAD Games: Multi-Agent Dynamic Games for Collaborative Agents in Adversarial Competitions
Rahul Mangharam, Electrical and Systems Engineering / Computer and Information Science (Penn Engineering)

11:30 to 11:45 a.m.: Cracking “Undruggable” Proteins with Cryptic Pockets
Greg Bowman, Biochemistry and Biophysics (PSOM) / Bioengineering (Penn Engineering)

11:45 a.m. to 12 p.m.: Conversational Agents for Improving Health Behaviors
Sharath Chandra Guntuku, Computer and Information Science (Penn Engineering)

AGH Auditorium

Davids Brown, Penn Engineering
Benchmarking Mitigations Against Covert Misuse

Zihao Chen, Penn Engineering
Your Contrastive Learning Problem Is a Secretly Alignment Problem

Seewon Choi, Penn Engineering
CTSketch: Compositional Tensor Sketching for Scalable Neurosymbolic Learning

Raghav Garimella, Penn Engineering
TBD

Helen Jin, Penn Engineering
TBD

Mayank Keoliya, Penn Engineering
Stable Prediction of Adverse Events in Medical Time-Series Data

Avishree Khare, Penn Engineering
Confidence Scores for Temporal Properties over Sequences of Predictions

Chaehyeon Kim, Penn Engineering
Decomposing Feature Attributions for Temporally-Aware Explainability in Sepsis

Bryan Li, Penn Engineering
Multilingual Retrieval Augmented Generation for Culturally Sensitive Tasks

Stephen Mell, Penn Engineering
A Fast, Reliable, and Secure Programming Language for LLM Agents with Code Actions

Casey Mogilevsky, Penn Engineering
SoftAlign-MSA: Scalable Multiple Sequence Alignment via Learned Continuous Representations

Nayan Patel, Penn Engineering
Spectral Asteroid Classification with Convolutional Neural Networks

Neil Sehgal, Penn Engineering
Opportunities for AI Chatbots in Health Persuasion: Results from Two RCTs

Guruprerana Shabadi, Penn Engineering
Composing Agents to Minimize Worst-Case Risk

Alaia Solko-Breslin, Penn Engineering
CTSketch: Compositional Tensor Sketching for Scalable Neurosymbolic Learning

Adam Stein, Penn Engineering
Once Upon an Input: Reasoning via Per-Instance Program Synthesis

Darshan Thaker, Penn Engineering
Frequency Guided Posterior Sampling for Diffusion-Based Image Restoration

Jiayi Xin, Penn Engineering
Interpretable Multimodal Interaction-Aware Mixture-of-Experts

Oscar Xu, Penn Engineering
Delta Activations: A Representation for Finetuned Large Language Models

Weiqiu You, Penn Engineering
Probabilistic Soundness Guarantees in LLM Reasoning Chains

AGH Lobby & AGH 105

2 to 2:15 p.m.: Evidence-Driven Conversational AI for Early Pregnancy Care: Design and Deployment of the CIRCA Platform
Anurag Venma, Penn Medicine

2:15 to 2:30 p.m.: AI Literacy at Penn Libraries
Jaj Karajgikar, Research Data and Digital Scholarship (Penn Libraries)

2:30 to 2:45 p.m.: Reconstructing Cell Lineage Trees from RNA Using Weakly Supervised Metric Learning
Junhyong Kim, Biology (School of of Arts & Sciences)

2:45 to 3 p.m.: Valid Forecasting of Heat Waves Two Weeks in Advance
Richard Berk, Criminology (SAS) / Statistics and Data Science (Wharton)

3 to 3:15 p.m.: Two Possibilities: Real World AI Governance, or What AI Risk Management Can Learn from Bank Supervision
Kevin Werbach, Legal Studies and Business Ethics (Wharton)

3:15 to 3:30 p.m.: Generative AI for Creativity and Innovation at Mack Institute
Valery Yakubovich, Mack Institute for Innovation Management (Wharton)

3:30 to 3:45 p.m.: Consumer Agents
Rory Van Loo, Legal Studies and Business Ethics (Wharton)

3:45 to 4 p.m.: Towards Improving the Reliability of AI
Edgar Dobriban, Statistics and Data Science (Wharton)

AGH Auditorium

Raghu Arghal, Penn Engineering
Algorithmic Information Mediators: Controlled Social Learning

Fabian Baumann, School of Arts and Sciences
Modeling Social Welfare of Generative AI

Mekides Belie, School of Arts and Sciences
RAG for Scientific Documents

Vicente Bosca, School of Arts and Sciences
Neural Networks as Local-to-Global Computations

Xuyang Chen, School of Arts and Sciences
Watermark in the Classroom: A Conformal Framework for Adaptive AI Usage Detection

Sourav Dey, School of Arts and Sciences
TBD

Daniel Herrera Esposito, School of Arts and Sciences
Supervised Quadratic Feature Analysis: Information Geometry for Dimensionality Reduction

Nayoon Lee, School of Arts and Sciences
Title TBD

Viet-Anh Le, Penn Engineering
Learning to Optimize and Adapt in Model Predictive Control

Xiang Li, Penn Medicine
Evaluating the Unseen Capabilities: How Much Do LLMs Actually Know?

Guiqiu Liao, Penn Medicine
FORLA: Federated Object-Centric Representation Learning with Slot Attention

Vivian Lin, Penn Engineering
Title TBD

Sydney Pugh, Penn Medicine
WATCH-SS: A Trustworthy and Explainable Modular Framework for Detecting Cognitive Impairment from Spontaneous Speech

Benjamin Shaffer, Penn Engineering
Structure Preserving Machine Learning for Robotics

Alok Shah, School of Arts and Sciences
Title TBD

Pavel Shapturenka, Penn Engineering
Soft-AE: Accelerating Conductive Polymer Development Through Accessible Autonomous Experimentation

Nandan Tumu, Penn Engineering
Social Influence Games: Modeling Adversarial Persuasion in Opinion Networks

Jie Wang, Penn Engineering
RoboArena: Distributed Real-World Evaluation of Generalist Robot Policies

Yangxinyu Xie, Wharton
Title TBD

Ziqing Xu, Wharton
Understanding the Learning Dynamics of LoRA: A Gradient Flow Perspective on Low-Rank Adaptation in Matrix Factorization

AGH Lobby & AGH 105