PALM Lab at USF 🌴
Data-centric AI, AI Safety, Mechanistic Interpretability, Agentic AI
University of South Florida
PALM Lab at USF 🌴
Data-centric AI, AI Safety, Mechanistic Interpretability, Agentic AI
University of South Florida
Embedded Files
Data-centric AI, AI Safety, Mechanistic Interpretability, Agentic AI
University of South Florida
The Pioneering Advancements in Learning Methods (PALM 🌴) Lab is a team of researchers at the University of South Florida who work on holistically improving current and next-generation of Artificial Intelligence models (e.g. Large Language Models) with respect to multiple facets of performance, such as utility, privacy, robustness, safety, and security. The team is especially interested in developing data-centric learning approaches that can benefit a large class of AI models in multiple domains such as Natural Language Processing and Computer Vision. Beyond core AI/ML problems, the team also works to translate these ideas to impact real-world models deployed in production systems (such as social media AIs), which by default are closed-source. More recently, we have also been working on problems related to Multimodal Generative AI.
Join us! The PALM Lab is always interested in recruiting talented and self-motivated PhD students and remote interns who are interesting in working on cutting-edge AI problems. If interested, please read our recent works and reach out to Prof. Anshuman Chhabra (PI) over email with your specific interests to inquire about the opportunities available within the lab.Â
The Pioneering Advancements in Learning Methods (PALM 🌴) Lab is a team of researchers at the University of South Florida who work on holistically improving current and next-generation of Artificial Intelligence models (e.g. Large Language Models) with respect to multiple facets of performance, such as utility, privacy, robustness, safety, and security. The team is especially interested in developing data-centric learning approaches that can benefit a large class of AI models in multiple domains such as Natural Language Processing and Computer Vision. Beyond core AI/ML problems, the team also works to translate these ideas to impact real-world models deployed in production systems (such as social media AIs), which by default are closed-source. More recently, we have also been working on problems related to Multimodal Generative AI.
Join us! The PALM Lab is always interested in recruiting talented and self-motivated PhD students and remote interns who are interesting in working on cutting-edge AI problems. If interested, please read our recent works and reach out to Prof. Anshuman Chhabra (PI) over email with your specific interests to inquire about the opportunities available within the lab.Â
You can find our existing and past group members here. More details on our recent research can be found here. We are located in sunny and beautiful Tampa Bay, FL.Â
You can find our existing and past group members here. More details on our recent research can be found here. We are located in sunny and beautiful Tampa Bay, FL.Â
News:
News:
6/1/2026: Our paper "Curvature-Weighted Capacity Allocation: A Minimum Description Length Framework for Layer-Adaptive LLM Optimization" was accepted to the UAI 2026 Main Conference [pdf]
4/30/2026: Our paper "Less Diverse, Less Safe: The Indirect But Pervasive Risk of Test-Time Scaling in Large Language Models" was accepted to the ICML 2026 Main Conference [pdf]
4/6/2026: Our paper "Bridging the Culture Gap: A Framework for LLM-Driven Socio-Cultural Localization of Math Word Problems in Low-Resource Languages" was accepted to the ACL 2026 Conference (Findings) [pdf] [code]
3/15/2026: Our paper "User-Side Interventions Reduce Harmful Content Exposure in Algorithmic Feeds" was accepted to the ICWSM 2026 Main Conference [pdf]
1/26/2026: Our paper "First is Not Really Better Than Last: Evaluating Layer Choice and Aggregation Strategies in Language Model Data Influence Estimation" was accepted to ICLR 2026 Main Conference [pdf]
9/18/2025: Our paper "LayerIF: Estimating Layer Quality for Large Language Models using Influence Functions" was accepted to NeurIPS 2025 Main Conference [pdf]
5/15/2025: Our paper "Re-ranking Using Large Language Models for Mitigating Exposure to Harmful Content on Social Media Platforms" was accepted to ACL 2025 Main Conference [pdf][code]
5/1/2025: Our paper "Outlier Gradient Analysis: Efficiently Identifying Detrimental Training Samples for Deep Learning Models" was accepted as an oral talk (top 1% of papers) at the ICML 2025 Main Conference [pdf][code]
1/22/2025: Our paper "Watching the AI Watchdogs: A Fairness and Robustness Analysis of AI Safety Moderation Classifiers" was accepted at the NAACL 2025 Main Conference [pdf][code]
1/22/2025: Our paper "Assessing LLMs for Zero-shot Abstractive Summarization Through the Lens of Relevance Paraphrasing" was accepted at the NAACL 2025 Conference (Findings) [pdf][code]
8/12/2024: Our paper "Incentivizing News Consumption on Social Media Platforms Using Large Language Models and Realistic Bot Accounts" was accepted for publication in PNAS Nexus [pdf]
3/13/2024: Our paper "Revisiting Zero-Shot Abstractive Summarization in the Era of Large Language Models from the Perspective of Position Bias" was accepted at the NAACL 2024 Main Conference as an oral talk [pdf] [code]
1/16/2024: Our paper "What Data Benefits My Classifier? Enhancing Model Performance and Interpretability Through Influence-Based Data Selection" was accepted as an oral talk (top 1.2% of papers) at ICLR 2024 [pdf] [code]
11/29/2023: Our paper "Towards Fair Video Summarization" has been accepted for publication in Transactions on Machine Learning Research (TMLR) [pdf] [code]
9/21/2023: Our paper "Auditing YouTube’s Recommendation System for Ideologically Congenial, Extreme, and Problematic Recommendations" was accepted for publication in PNAS [pdf] [code]
1/20/2023: Our paper "Robust Fair Clustering: A Novel Fairness Attack and Defense Framework" was accepted at the ICLR 2023 Main Conference [pdf] [code] [poster]
9/14/2022: Our paper "On the Robustness of Deep Clustering Models: Adversarial Attacks and Defenses" was accepted at the NeurIPS 2022 Main Conference [pdf] [supplementary] [code] [poster]
Page updated
Google Sites
Report abuse