Alternate text

Weekly Media Update

Week of July 10, 2026

Dolores Albarracín Appointed New Director of Annenberg Public Policy Center

Dolores Albarracín is the new director of the Annenberg Public Policy Center, succeeding Kathleen Hall Jamieson, who will continue her work as director emerita. “Dolores is an exceptional scholar whose work on communication, behavior, and public health is especially vital in this moment,” says Dean Sarah Banet-Weiser of the Annenberg School for Communication.

How Personalized AI Tutors Can Help Students Learn

For many students, asking ChatGPT for homework help has replaced raising a hand in class or going to office hours. As AI becomes a default classroom tutor, educators are grappling with a new question: How do you design these tools so they actually support learning? New research from Hamsa Bastani and collaborators reveals how small design changes can make AI tutoring more effective by emulating the most effective practices of human instructors.

Machine Text Detectors Are Membership Inference Attacks
7/10: MemFM @ ICML 2026 | Chris Callison-Burch

AI as a Tool for Mathematics, Computer Science, and Machine Learning
7/10: ICML 2026 Workshop | Edgar Dobriban

A Prompting Workflow and a Use Case in Mathematical Statistics
7/10: AI4Research @ ICML 2026 | Edgar Dobriban

Human-AI Collaborative Uncertainty Quantification
7/10: AI4GOOD @ ICML 2026 | Hamed Hassani

Multi-Objective Reinforcement Learning for Large-Scale Tote Allocation in Human-Robot Collaborative Fulfillment Centers
7/10: RLxF: Reinforcement Learning from World Feedback @ ICML 2026 | Michael Kearns and Aaron Roth

Contrastive Conceptor Activation Steering (COAST): Unlocking Vision-Language-Action Models Through Hidden States
7/10: Mechanistic Interpretability Workshop @ ICML 2026 | Lyle Ungar

New Frontiers in Game-Theoretic Learning
7/10: ICML 2026 | René Vidal

Leaders TalkX 12 - Co‑Creating Tomorrow: Digital and AI in Partnership
7/10: World Summit on the Information Society Forum 2026 | Christopher Yoo

Emergent Alignment via Competition
7/9: ICML 2026 | Surbhi Goel and Aaron Roth

Rethinking How We Measure Soccer Performance
7/8: Wharton Moneyball | Eric Bradlow

General Synthetic-Powered Inference
7/8: ICML 2026 | Edgar Dobriban

Statistical Early Stopping for Reasoning Models
7/8: ICML 2026 | Edgar Dobriban

Conformal Risk-Averse Decision Making With Action Conditional Guarantee
7/8: ICML 2026 | Hamed Hassani

Multi-Round Human–AI Collaboration With User-Specified Requirements
7/8: ICML 2026 | Hamed Hassani

A Constrained Optimization Perspective of Unrolled Transformers
7/8: ICML 2026 | Alejandro Ribeiro

Statistical Impossibility and Possibility of Aligning LLMs With Human Preferences: From Condorcet Paradox to Nash Equilibrium
7/8: ICML 2026 | Weijie Su

Beyond Test-Time Training: Learning to Reason via Hardware-Efficient Optimal Control
7/8: ICML 2026 | René Vidal

Kaleidoscope Academic Conference – AI and Frontier Technologies for Good
7/8: AI for Good Global Summit | Christopher Yoo

Position: AI Capabilities Are Not Increasing Exponentially
7/7: ICML 2026 | Hamsa Bastani

NSF-SciFy: Mining the NSF Awards Database for Scientific Claims
7/7: ACL 2026 | Chris Callison-Burch

Is Code Better Than Language for Algorithmic Reasoning?
7/7: ICML 2026 | Surbhi Goel

Temporal Difference Learning With Compressed Updates: Error-Feedback Meets Reinforcement Learning
7/7: ICML 2026 | Hamed Hassani

When to Trust the Cheap Check: Weak and Strong Verification for Reasoning
7/7: ICML 2026 | Hamed Hassani

REALISTA: Realistic Latent Adversarial Attacks That Elicit LLM Hallucinations
7/7: ICML 2026 | Hamed Hassani and René Vidal

Unlearning in Diffusion Models: A Unified Framework With KL Divergence and Likelihood Constraints
7/7: ICML 2026 | Alejandro Ribeiro

Correctness-Optimized Residual Activation Lens (CORAL): Transferrable and Calibration-Aware Inference-Time Steering
7/7: ICML 2026 | Lyle Ungar

AI in Education, Research and Skills Development Roundtable
7/7: AI for Good Global Summit | Christopher Yoo

The USMCA Countdown Is a Trap
7/6: Gad's Newsletter | Gad Allon

Online Conformal Prediction via Universal Portfolio Algorithms
7/6: ICML 2026 | Edgar Dobriban

Optimal Decision-Making Based on Prediction Sets
7/6: ICML 2026 | Edgar Dobriban

When Less Gives More: Bias From Small Dataset Can Speed up Training
7/6: ICML 2026 | Surbhi Goel

Human-AI Collaborative Uncertainty Quantification
7/6: ICML 2026 | Hamed Hassani

Formalizing the Binding Problem
7/6: ICML 2026 | Konrad Kording

Calibration, Decisions, and Collaboration in Learning
7/6: ICML 2026 | Aaron Roth

Stronger Benchmarks for Prediction as a Service With Constraints
7/6: ICML 2026 | Aaron Roth

A Single-Echo/Multi-Echo BOLD fMRI Test-Retest Dataset
7/6: Data in Brief | Ted Satterthwaite

Transformers Learn the Optimal DDPM Denoiser for Multi-Token GMMs
7/6: ICML 2026 | René Vidal

Uncertainty Quantification in LLM Agents: Foundations, Emerging Challenges, and Opportunities
7/5: ACL 2026 | Hamed Hassani

Knowing When Not to Answer: Lightweight KB-Aligned OOD Detection for Safe RAG
7/5: ACL 2026 | Lyle Ungar

Autorubric: A Unified Framework for Rubric-Based LLM Evaluation
7/4: RAG4Reports and GEM @ ACL 2026 | Chris Callison-Burch

ThinknCheck: Grounded Claim Verification With Compact, Reasoning-Driven, and Interpretable Models
7/4: Natural Language Processing and Information Systems | Chris Callison-Burch

What Do Claim Verification Datasets Actually Test? A Reasoning Trace Analysis
7/4: Natural Language Processing and Information Systems | Chris Callison-Burch

Testing Noise Assumptions of Learning Algorithms
7/3: COLT 2026 | Surbhi Goel

Predicting Variation in Emotional Valence and Arousal Over Time From Ecological Essays
7/3: SemEval-2026 | Lyle Ungar

A Sieve-Accelerated Quadrature Method for Exact Privacy Accounting in the 2020 U.S. Decennial Census
6/29: arXiv | Weijie Su

SGD at the Edge of Stability: Stochastic Stabilization With Large Learning Rates
6/29: arXiv | René Vidal

College Applications With Non-Homogeneous Application Costs
6/17: ACM Transactions on Economics and Computation | Sampath Kannan and Rakesh Vohra

 
XFacebook LinkedInBluesky
 

Website

Get Involved

Past Events

 

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

Manage your preferences | Opt Out using TrueRemove™
Got this as a forward? Sign up to receive our future emails.
View this email online.

Penn Engineering 107 Towne Building 220 South 33rd Street | Philadelphia, PA 19104 US

This email was sent to lhoot@seas.upenn.edu.
To continue receiving our emails, add us to your address book.

powered by emma