From Teacher to Student: Tracking Memorization Through Model Distillation
机构 * Indian Institute of Technology, Roorkee（印度理工学院拉胡尔学院）
Comments 5 pages, in-proceedings L2M2 @ ACL 2025
期刊&会议
ACL
Annual Meeting of the Association for Computational Linguistics · 会议 · Natural Language Processing
SECQUE: A Benchmark for Evaluating Real-World Financial Analysis Capabilities
机构 * Microsoft Industry AI（微软产业人工智能）
Comments Benchmark available at: https://huggingface.co/datasets/nogabenyoash/SecQue
Journal ref n Proceedings of the Fourth Workshop on Generation, Evaluation and Metrics, Association for Computational Linguistics (2025) https://aclanthology.org/2025.gem-1.16/
Measuring Social Biases in Masked Language Models by Proxy of Prediction Quality
机构 * New York University（纽约大学）
Journal ref Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), 2025, pages 1337--1361
Rationalizing Transformer Predictions via End-To-End Differentiable Self-Training
机构 * Computational Linguistics, Department of Linguistics Bielefeld University（语言学计算系，语言学系比勒菲尔德大学）
Journal ref Proceedings of the 2024 Conference on Empirical Methods in Natural Language Processing, pages 11894-11907, Miami, Florida, USA. Association for Computational Linguistics
Model Interpretability and Rationale Extraction by Input Mask Optimization
机构 * Bielefeld University Faculty for Linguistics and Literary Studies（比勒菲尔德大学语言学与文学研究系）
Journal ref Findings of the Association for Computational Linguistics: ACL 2023, pages 13722-13744, Toronto, Canada. Association for Computational Linguistics
PersonaTwin: A Multi-Tier Prompt Conditioning Framework for Generating and Evaluating Personalized Digital Twins
机构 * Viterbi School of Engineering, University of Southern California（南加州大学维特比工程学院） ； Human-centered Analytics Lab, University of Notre Dame（诺特大学人本分析实验室） ； Department of IT, Analytics, and Operations, University of Notre Dame（诺特大学信息技术、分析与运营部门） ； Department of Information Systems, Business Statistics and Operations Management, HKUST（香港科技大学信息系统、商业统计与运营管理部门）
Comments Presented at the Generation, Evaluation & Metrics (GEM) Workshop at ACL 2025
ImpliHateVid: A Benchmark Dataset and Two-stage Contrastive Learning Framework for Implicit Hate Speech Detection in Videos
机构 * Indian Institute of Technology Indore（印度理工学院印地尔分校） ； Chaitanya Bharathi Institute of Technology（恰伊坦亚·巴哈蒂技术学院）
Comments Published in ACL 2025
S2WTM: Spherical Sliced-Wasserstein Autoencoder for Topic Modeling
机构 * Indian Association for the Cultivation of Science（印度科学培养协会）
Comments Accepted as a long paper for ACL 2025 main conference
Rethinking the Role of Prompting Strategies in LLM Test-Time Scaling: A Perspective of Probability Theory
机构 * MAIS, Institute of Automation, Chinese Academy of Sciences（自动化研究所，中国科学院） ； School of Artificial Intelligence, University of Chinese Academy of Sciences（中国科学院大学人工智能学院） ； University of California, Santa Barbara（加州大学圣巴巴拉分校） ； Beijing Wenge Technology Co., Ltd（北京文格科技有限公司）
Comments ACL 2025 Outstanding Paper Award, 33 pages, 51 figures
Journal ref ACL.Volume 1: Long Papers (2025) 27962-27994
A Dual-Perspective NLG Meta-Evaluation Framework with Automatic Benchmark and Better Interpretability
机构 * Wangxuan Institute of Computer Technology（计算机技术研究院）
Comments Accepted by ACL 2025
CLEAR: Character Unlearning in Textual and Visual Modalities
机构 * AIRI ； HSE University（俄罗斯高等经济大学） ； Skoltech（斯克里普契克技术大学） ； Sber AI（俄罗斯储蓄银行人工智能实验室） ； MTUCI（莫斯科国立大学） ； MIPT（米哈伊尔戈尔斯基理工学院） ； ISP RAS Research Center for Trusted AI（俄罗斯科学院信息与系统问题研究所可信人工智能研究中心）
Journal ref https://aclanthology.org/2025.findings-acl.1058/
One Size Fits None: Rethinking Fairness in Medical AI
机构 * DFKI ； Friedrich-Alexander-Universität Erlangen-Nürnberg（弗里德里希-亚历山大-埃朗根-纽伦堡大学） ； Charité - Universitätsmedizin Berlin（Charité - 大学医学中心柏林） ； DNC Information Management GmbH（DNC信息管理有限公司） ； University of Regensburg（雷根萨大学） ； University Hospital Regensburg（雷根萨大学医院） ； Asklepios Klinikum Harburg（阿斯克勒庇奥斯克linikum哈弗堡） ； TU Berlin（柏林技术大学）
Comments Accepted at the 6th Workshop on Gender Bias in Natural Language Processing at ACL 2025
Journal ref Proceedings of the 6th Workshop on Gender Bias in Natural Language Processing (GeBNLP), 2025, pp. 282-289
DNB-AI-Project at SemEval-2025 Task 5: An LLM-Ensemble Approach for Automated Subject Indexing
机构 * Deutsche Nationalbibliothek（德国国家图书馆）
Comments 11 pages, 4 figures, submitted to SemEval-2025 workshop Task 5: LLMs4Subjects
Journal ref In Proceedings of the 19th International Workshop on Semantic Evaluation (SemEval-2025), pages 1118-1128, Vienna, Austria. Association for Computational Linguistics
LED-Merging: Mitigating Safety-Utility Conflicts in Model Merging with Location-Election-Disjoint
机构 * Shanghai Jiao Tong University（上海交通大学） ； Shanghai AI Laboratory（上海人工智能实验室） ； East China Normal University（华东师范大学）
Comments Accepted by ACL2025 main conference
LACA: Improving Cross-lingual Aspect-Based Sentiment Analysis with LLM Data Augmentation
机构 * Department of Computer Science and Engineering（计算机科学与工程系） ； NTIS – New Technologies for the Information Society University of West Bohemia in Pilsen, Faculty of Applied Sciences（新信息技术社会大学西波希米亚大学皮尔森分校应用科学学院）
Comments Published in Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics; Volume 1: Long Papers (ACL 2025). Official version: https://aclanthology.org/2025.acl-long.41/
DP-BART for Privatized Text Rewriting under Local Differential Privacy
Comments Accepted at ACL Findings 2023
How reparametrization trick broke differentially-private text representation learning
Comments Pre-print version; Accepted at ACL 2022; (v2: fixing title typo)
Fine-grained Video Dubbing Duration Alignment with Segment Supervised Preference Optimization
机构 * Alibaba Digital Media and Entertainment Group（阿里巴巴数字媒体与娱乐集团） ； School of Software Engineering, Huazhong University of Science and Technology（华中科技大学软件学院） ； Beijing Key Laboratory of Traffic Data Mining and Embodied Intelligence, Beijing Jiaotong University（北京交通大学交通数据挖掘与具身智能重点实验室）
Comments This paper is accepted by ACL2025 (Main)
Journal ref Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). 2025: 4524-4546
ChatBench: From Static Benchmarks to Human-AI Evaluation
机构 * Microsoft Research（微软研究院） ； University of California, Berkeley（加州大学伯克利分校） ； University of Toronto（多伦多大学）
Comments ACL 2025 (main)
Sleepless Nights, Sugary Days: Creating Synthetic Users with Health Conditions for Realistic Coaching Agent Interactions
机构 * Google DeepMind（谷歌DeepMind） ； Verily Life Sciences（Verily生命科学） ； Google（谷歌）
Comments Published in Findings of the Association for Computational Linguistics: ACL 2025
Journal ref Findings of the Association for Computational Linguistics: ACL 2025, pages 14159-14181, Vienna, Austria. Association for Computational Linguistics
Bridging the Visual Gap: Fine-Tuning Multimodal Models with Knowledge-Adapted Captions
机构 * Tel Aviv University（特拉维夫大学） ； Google Research（谷歌研究）
Comments Accepted to NAACL 2025
Journal ref Proceedings of the 2025 Conference of the North American Chapter of the Association for Computational Linguistics, Human Language Technologies, Long Papers, pp. 10497-10518
ICC: Quantifying Image Caption Concreteness for Multimodal Dataset Curation
机构 * Tel-Aviv University（特拉维夫大学）
Comments Accepted to ACL 2024 (Finding). For Project webpage, see https://moranyanuka.github.io/icc/
Journal ref Findings of the Association for Computational Linguistics: ACL 2024, pages 11048-11064, Bangkok, Thailand, August 2024
Learning to Insert [PAUSE] Tokens for Better Reasoning
机构 * KAIST AI（韩国科学技术院人工智能）
Comments 18 pages, 5 figures, ACL findings
Journal ref https://aclanthology.org/2025.findings-acl.1217/
Predicting Depression in Screening Interviews from Interactive Multi-Theme Collaboration
机构 * Xidian University（西安电子科技大学） ； Harbin Institute of Technology（哈尔滨工业大学）
Comments Findings of ACL2025
LLMs can be easily Confused by Instructional Distractions
机构 * IPAI, Seoul National University（IPAI，首尔国立大学） ； LG AI Research（LG人工智能研究所） ； Dept. of ECE, Seoul National University（电子工程系，首尔国立大学） ； SNU-LG AI Research Center（SNU-LG人工智能研究所）
Comments 8 pages
Journal ref Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), pages 19483-19496, Vienna, Austria, July 2025
Fairness through Difference Awareness: Measuring Desired Group Discrimination in LLMs
机构 * Stanford University（斯坦福大学） ； Cornell Tech（康奈尔科技）
Comments Best Paper award at ACL 2025; dataset available at https://github.com/Angelina-Wang/difference_awareness
Auto-TA: Towards Scalable Automated Thematic Analysis (TA) via Multi-Agent Large Language Models with Reinforcement Learning
Comments Presented at ACL 2025 SRW
ACCESS DENIED INC: The First Benchmark Environment for Sensitivity Awareness
机构 * L3S Research Center, Leibniz University Hannover（莱比锡大学汉诺威分校L3S研究中心） ； E.ON Grid Solutions（E.ON电网解决方案）
Comments 20 pages, 4 figures, 8 tables, ACL 2025 (Findings), Figure 3 was corrected and is now aligned with the values of Table 2, Project Page: https://drenfazlija.github.io/AccessDeniedInc/
Journal ref Findings of the Association for Computational Linguistics: ACL 2025, pages 13221-13240, Vienna, Austria, Association for Computational Linguistics, 2025
Embracing Imperfection: Simulating Students with Diverse Cognitive Levels Using LLM-based Agents
机构 * Zhejiang University（浙江大学） ； Hong Kong University of Science and Technology（香港科学与技术大学）
Comments ACL 2025
Planning with Diffusion Models for Target-Oriented Dialogue Systems
机构 * Department of Computer Science and Engineering, The Ohio State University, USA（计算机科学与工程系，俄亥俄州立大学） ； Department of Biomedical Informatics, The Ohio State University, USA（生物医学信息学系，俄亥俄州立大学） ； Translational Data Analytics Institute, The Ohio State University, USA（转化数据分析研究所，俄亥俄州立大学）
Comments Accepted to ACL 2025 Main Conference
