Weekly Media Update

Week of April 25, 2025

Chief AI Scientist at Meta Visits Penn for AI Month Fireside Chat

Yann LeCun, the Chief AI Scientist at Meta, visited Penn last week for a fireside chat hosted by Michael Kearns. Kearns worked alongside LeCun in the 1990s at what was once AT&T’s Bell Laboratories. At the event, Kearns and LeCun spoke about their experience at Bell Labs, which turns 100 this year, and specifically the importance of sharing scientific research.

How AI Vaporizes Long-Term Learning

The use of artificial intelligence (AI) chatbots in the classroom has sparked much debate in the education community. A 2024 study from Hamsa Bastani and colleagues, featured on Edutopia, revealed AI tools like ChatGPT could boost test scores—but ultimately undermined students’ learning and retention. 

Breaking Boundaries, Advancing Rights: A Celebration of Anita Allen's Scholarship and Leadership

4/25: Penn Carey Law | Anita Allen

Beyond the Imitation Game: Quantifying and Extrapolating the Capabilities of Language Models

4/25: ICLR 2025 | Chris Callison-Burch

Progressive Distillation Induces an Implicit Curriculum

4/25: ICLR 2025 | Surbhi Goel

Learning Efficient Positional Encodings With Graph Neural Networks

4/25: ICLR 2025 | Alejandro Ribeiro

Fine-Tuning Attention Modules Only: Enhancing Weight Disentanglement in Task Arithmetic

4/25: ICLR 2025 | Weijie Su

InCoDe: Interpretable Compressed Descriptions for Image Generation

4/25: ICLR 2025 | René Vidal

Negative Stepsizes Make Gradient-Descent-Ascent Converge

4/24: UPenn Optimization Seminar | Jason Altschuler

Scientific AI

4/24: Penn School of Arts and Sciences | Maria Cuellar 

TSVD: Bridging Theory and Practice in Continual Learning With Pre-Trained Models

4/24: ICLR 2025 | Alejandro Ribeiro and René Vidal

Conformal Language Model Reasoning with Coherent Factuality

4/24: ICLR 2025 | Aaron Roth and Surbhi Goel

Magnetic Preference Optimization: Achieving Last-Iterate Convergence for Language Model Alignment

4/24: ICLR 2025 | Weijie Su

Ashley Casovan: From Privacy Practice to AI Governance

4/24: The Road to Accountable AI | Kevin Werbach 

Meet Jean, Penn Engineering’s AI Tutor, Available 24/7

4/23: Penn Engineering AI | Chris Callison-Burch

Approaching Rate-Distortion Limits in Neural Compression With Lattice Transform Coding

4/23: ICLR 2025 Spotlight | Hamed Hassani and Shirin Saeedi Bidokhti

Fake News, Echo Chambers, and Algorithms: A Data Science Perspective

4/23: ASSET Center Seminar | Duncan Watts

REAL ID and Unreal Lines: TSA’s New Travel Adventure

4/21: Gad's Newsletter | Gad Allon

Confirmation Bias: The Original Error

4/21: Community for Rigor | Konrad Kording

The Impact of Loper Bright on US Administrative Law: Power Balance and Future Prospects

4/19: Forward Pathway | Cary Coglianese 

Probabilistic Stability Guarantees for Feature Attributions

4/18: arXiv | Surbhi Goel  

Graph-Based Path Planning With Dynamic Obstacle Avoidance for Autonomous Parking

4/17: arXiv | Nikolai Matni

Resolving the Reference Class Problem At Scale in Machine Learning

4/14: Philosophy of Science | Aaron Roth 

Did Loper Bright Also Overturn Notice-and-Comment Rulemaking Procedure?

4/11: Yale Journal on Regulation | Cary Coglianese

Editor’s Note: The Process Is the Punishment

4/10: Sustainable Views | Cary Coglianese 

Reconstructing Cell Lineage Trees from Phenotypic Features with Metric Learning

3/18: arXiv | Junhyong Kim 

Website

Get Involved

Past Events

107 Towne Building
220 South 33rd Street
Philadelphia, Pennsylvania 19104
warrenct@seas.upenn.edu