Papers by Charith Peris
Coordinated Replay Sample Selection for Continual Federated Learning (2023.emnlp-industry)
Copied to clipboard
Jack Good, Jimit Majmudar, Christophe Dupuy, Jixuan Wang, Charith Peris, Clement Chung, Richard Zemel, Rahul Gupta
| Challenge: | Continual Federated Learning (CFL) combines decentralized learning with continuous learning . ubiquity of personal devices with a network connection offers rich source of data for learning problems . |
| Approach: | They propose to combine decentralized learning with a continuous learning approach . they propose to coordinate gradient-based replay sample selection across clients . |
| Outcome: | The proposed method shows gains early in the low replay size regime, when the budget for storing past data is small. |
The steerability of large language models toward data-driven personas (2024.naacl-long)
Copied to clipboard
Junyi Li, Charith Peris, Ninareh Mehrabi, Palash Goyal, Kai-Wei Chang, Aram Galstyan, Richard Zemel, Rahul Gupta
| Challenge: | Large language models generate biased responses where opinions of certain groups and populations are underrepresented. |
| Approach: | They propose a data-driven notion of persona that allows for a more nuanced understanding of different (latent) social groups present in the population. |
| Outcome: | The proposed method improves model steerability by 57% over baselines. |
Evaluating Differentially Private Synthetic Data Generation in High-Stakes Domains (2024.findings-emnlp)
Copied to clipboard
| Challenge: | a lack of anonymization of sensitive text data hinders development of NLP tools . poorly anonymized sensitive data cannot be easily shared with annotators or external researchers . |
| Approach: | They propose to use synthetic data to generate differentially private language models in place of real data to facilitate NLP development without compromising privacy. |
| Outcome: | The proposed model can be used to train public models without compromising privacy. |
Attribute Controlled Fine-tuning for Large Language Models: A Case Study on Detoxification (2024.findings-emnlp)
Copied to clipboard
Tao Meng, Ninareh Mehrabi, Palash Goyal, Anil Ramakrishna, Aram Galstyan, Richard Zemel, Kai-Wei Chang, Rahul Gupta, Charith Peris
| Challenge: | Using a sequence-level constraint, we regularize the LLMtraining by penalizing the KL divergence between the desired output distribution and the LRM’s posterior. |
| Approach: | They propose a constraint learning schema forfine-tuning Large Language Models with attribute control by penalizing the KL divergence be-tween the desired output distribution and the LLM's posterior. |
| Outcome: | The proposed approach improves the performance of large language models while enhancing their utility and generation quality. |
Multi-Token Completion for Text Anonymization (2026.eacl-long)
Copied to clipboard
| Challenge: | Text anonymization is a critical task for enabling research and development in high-stakes domains containing private data. |
| Approach: | They propose a method for predicting replacements for sensitive spans with principled use-inspired evaluation criteria. |
| Outcome: | The proposed method produces more realistic text and preserves utility than alternative infilling methods and differentially private mechanisms across multiple domains without retraining. |
MASSIVE: A 1M-Example Multilingual Natural Language Understanding Dataset with 51 Typologically-Diverse Languages (2023.acl-long)
Copied to clipboard
Jack FitzGerald, Christopher Hench, Charith Peris, Scott Mackie, Kay Rottmann, Ana Sanchez, Aaron Nash, Liam Urbach, Vishesh Kakarala, Richa Singh, Swetha Ranganath, Laurie Crist, Misha Britan, Wouter Leeuwis, Gokhan Tur, Prem Natarajan
| Challenge: | We present the MASSIVE dataset–Multilingual Amazon Slu resource package (SLURP) for Slot-filling, Intent classification, and Virtual assistant evaluation. |
| Approach: | They present a 1M-example dataset of Amazon Slu utterances . they localize the dataset into 50 typologically diverse languages . |
| Outcome: | The proposed model includes exact match accuracy, intent classification accuracy, and slot-filling F1 score. |
ARES: Adaptive Red-Teaming and End-to-End Repair of Policy-Reward System (2026.acl-long)
Copied to clipboard
Jiacheng Liang, Yao Ma, Tharindu Kumarage, Satyapriya Krishna, Rahul Gupta, Kai-Wei Chang, Aram Galstyan, Charith Peris
| Challenge: | Existing red-teaming approaches focus on policy-level weaknesses, but they overlook systemic weaknesses . aRES exploits dual-targeting weaknesses in both the core LLM and the RM simultaneously. |
| Approach: | a new framework uncovers weaknesses in both the core and the reward models simultaneously . a "Safety Mentor" generates semantically coherent adversarial prompts . |
| Outcome: | ARES uncovers weaknesses in both the core LLM and the RM simultaneously . it fine-tunes the LM to detect harmful content, then optimizes the core model . |
Knowledge Distillation Transfer Sets and their Impact on Downstream NLU Tasks (2022.emnlp-industry)
Copied to clipboard
| Challenge: | Domain Classification (DC) and Intent Classification/Named Entity Recognition (ICNER) are the most common methods for reducing teacher-student knowledge into manageable sizes for low-latency downstream applications. |
| Approach: | They investigate whether distillation from a generic LM benefits downstream tasks . a domain classification and a task-specific data set are used to fine tune the model . |
| Outcome: | The proposed model improves across tasks and test sets when only task-specific data is used. |
Towards Safety Reasoning in LLMs: AI-agentic Deliberation for Policy-embedded CoT Data Creation (2025.findings-acl)
Copied to clipboard
Tharindu Kumarage, Ninareh Mehrabi, Anil Ramakrishna, Xinyan Zhao, Richard Zemel, Kai-Wei Chang, Aram Galstyan, Rahul Gupta, Charith Peris
| Challenge: | Safety reasoning paradigms require high-quality policy-embedded chain-of-thought datasets . generating such data through human annotations is prohibitively expensive . |
| Approach: | They propose AIDSAFE: Agentic Iterative Deliberation for Safety Reasoning . AIDS AFE leverages multi-agent deliberation to iteratively expand reasoning on safety policies . |
| Outcome: | The proposed model improves policy adherence and reasoning quality while maintaining acceptable utility and over-refusal accuracy. |
Controlling the Extraction of Memorized Data from Large Language Models via Prompt-Tuning (2023.acl-short)
Copied to clipboard
Mustafa Ozdayi, Charith Peris, Jack FitzGerald, Christophe Dupuy, Jimit Majmudar, Haidar Khan, Rahil Parikh, Rahul Gupta
| Challenge: | Large Language Models memorize significant portions of training data, which poses privacy risk. |
| Approach: | They propose a prompt-tuning approach to control the extraction rates of memorized content in large language models. |
| Outcome: | The proposed techniques yield 9.3% increase in extraction rate compared to baseline model . the proposed defense achieves 97.7% reduction with a perplexity increase of 16.9% . |
SWAN: Semantic Watermarking with Abstract Meaning Representation (2026.acl-long)
Copied to clipboard
Ziping Ye, Gourab Dey, Christos Christodoulopoulos, Charith Peris, Anil Ramakrishna, Weitong Ruan, Aram Galstyan, Kai-Wei Chang, Rahul Gupta, Ninareh Mehrabi
| Challenge: | Existing methods to embed signatures by adjusting token selection preferences during text generation are highly sensitive to paraphrasing and synonyms. |
| Approach: | They propose a framework that embeds watermark signatures into the semantic structure of a sentence using Abstract Meaning Representation (AMR). |
| Outcome: | Empirical evaluation shows SWAN matches state-of-the-art detection performance on unaltered watermarked text while improving robustness against paraphrasing. |
Tree-of-Traversals: A Zero-Shot Reasoning Algorithm for Augmenting Black-box Language Models with Knowledge Graphs (2024.acl-long)
Copied to clipboard
Elan Markowitz, Anil Ramakrishna, Jwala Dhamala, Ninareh Mehrabi, Charith Peris, Rahul Gupta, Kai-Wei Chang, Aram Galstyan
| Challenge: | Knowledge graphs (KGs) complement Large Language Models (LLMs) by providing reliable, structured, domain-specific, and up-to-date external knowledge. |
| Approach: | They propose a zero-shot reasoning algorithm that augments black-box LLMs with one or more KGs. |
| Outcome: | The proposed algorithm significantly improves performance on question answering and KG question answering tasks. |
Defenses Against Prompt Attacks Learn Surface Heuristics (2026.acl-long)
Copied to clipboard
| Challenge: | Large language models (LLMs) are increasingly deployed in security-sensitive applications . recent defenses rely on supervised fine-tuning with benign and malicious labels . position bias arises when benign content placed later in a prompt is rejected at much higher rates . |
| Approach: | They analyze three recurring shortcut behaviors induced by supervised fine-tuning . position bias arises when benign content placed later in a prompt is rejected . token trigger bias occurs when strings common in attack data raise rejection probability . |
| Outcome: | The proposed model overrides intended logic when adversarial instructions appear . the proposed model has low rejection rates but narrow correlations in defense data . |