China

Large language models (LLMs) commonly risk copyright infringement by reproducing protected content verbatim or with insufficient transformative modifications, posing significant ethical, legal, and practical concerns. Current inference-time safeguards predominantly rely on restrictive refusal-based filters, often compromising the practical utility of these models. To address this, we collaborated closely with intellectual property experts to develop FUA-LLM (Fair Use Aligned Language Models), a legally-grounded framework explicitly designed to align LLM outputs with fair-use doctrine. Central to our method is FairUseDB, a carefully constructed dataset containing 18,000 expert-validated examples covering nine realistic infringement scenarios. Leveraging this dataset, we apply Direct Preference Optimization (DPO) to fine-tune open-source LLMs, encouraging them to produce legally compliant and practically useful alternatives rather than resorting to blunt refusal. Recognizing the shortcomings of traditional evaluation metrics, we propose new measures: Weighted Penalty Utility and Compliance Aware Harmonic Mean (CAH) to balance infringement risk against response utility. Extensive quantitative experiments coupled with expert evaluations confirm that FUA-LLM substantially reduces problematic outputs (up to 20%) compared to state-of-the-art approaches, while preserving real-world usability.

EMNLP 2025

Nine Ways to Break Copyright Law and Why Our LLM Won’t: A Fair Use Aligned Generation Framework

poster

## Welcome!
"I am excited to welcome you to this year’s edition of the Conference on Empirical Methods in Natural Language Processing! Importantly, it marks the 30th edition of EMNLP. With over 8,000 submissions, more than 3,000 accepted papers, and thousands of attendees, we have come a long way from that first
workshop, which had 14 accepted papers. As the field looks ahead, Suzhou is the fitting location for celebrating this milestone: rooted in a long literary tradition, yet modern and forward-looking, and home to a large share of the NLP community."<br>

*Message from the General Chair, Dirk Hovy*

[**Link to Conference Handbook**](https://drive.google.com/file/d/1johU5QqVVYO4RfH7QcIORr7qrVBdzdwC/view?usp=sharing)





<br>

Celebrate 30 Years of EMNLP! 
EMNLP 2025 will be held in Suzhou, China from November 5th to November 9th, 2025.

Text-to-SQL oriented table acquisition suffers from heterogeneous semantic gap. To address the issue, we propose a Reverse Engineering (RE) based optimization approach. Instead of forward table search using questions as queries, RE reversely generates potentially-matched question conditioned on table schemas, and promotes semantic consistency verification between homogeneous questions. We experiment on two benchmarks, including SpiderUnion and BirdUnion. The test results show that our approach yields substantial improvements compared to the Retrieval-Reranker (2R) baseline, and achieves competitive performance in both table acquisition and Text-to-SQL tasks.

Enhancing SQL Table Acquisition with Reverse Engineering for Text-to-SQL

Large Vision-Language Models (LVLMs) have recently shown promising results on various multimodal tasks, even achieving human-comparable performance in certain cases. Nevertheless, LVLMs remain prone to hallucinations--they often rely heavily on a single modality or memorize training data without properly grounding their outputs. To address this, we propose a training-free, tri-layer contrastive decoding with watermarking, which proceeds in three steps: (1) select a mature layer and an amateur layer among the decoding layers, (2) identify a pivot layer using a watermark-related question to assess whether the layer is visually well-grounded, and (3) apply tri-layer contrastive decoding to generate the final output. Experiments on public benchmarks such as POPE, MME and AMBER demonstrate that our method achieves state-of-the-art performance in reducing hallucinations in LVLMs and generates more visually grounded responses.

Watermarking for Factuality: Guiding Vision-Language Models Toward Truth via Tri-layer Contrastive Decoding

Encoder models offer efficiency for specific tasks, but their performance depend on data availability. While Large Language Models (LLMs) excel at few-shot learning, their direct application in real-world scenarios is often hindered by their high computational cost. To address this challenge, we propose a simple yet effective approach that uses LLMs for data generation and scoring to improve encoder only model performance. We evaluate this framework on few-shot Multiple Choice Question Answering (MCQA), an important task where acquiring labeled data is costly. Our approach utilizes LLMs to create MCQA questions and choices (exploring both direct JSON and decomposed generation methods) and assigns probability scores to these choices. This generated data and the LLM scores are then used to fine-tune smaller and more efficient DeBERTa-v3-base using distillation loss. Extensive experiments on the MMLU benchmark demonstrate that our method improves accuracy from 28.9% to 39.3%, representing a gain of over 10% compared to a baseline finetuned directly on 5-shot examples. This shows the effectiveness of LLM-driven data generation and knowledge distillation for few-shot MCQA.

LLM Distillation for Efficient Few-Shot Multiple Choice Question Answering

Graphical User Interface (GUI) interaction, which aims to develop an intelligent GUI agent that executes user instructions to perform tasks such as installing applications by controlling digital devices, has gained significant attention due to its practical value. Although current advanced multimodal large language models (LLMs) provide GUI agents with robust perception and reasoning capabilities, they often struggle with the precise localization of small elements. To tackle this problem, we propose InReAct, a multimodal GUI agent framework that unifies observing, thinking, and acting for precise and interpretable decision-making. It is trained via a two-stage process: curriculum learning to progressively build perception, grounding, and reasoning abilities, followed by reinforcement learning to refine pixel-level grounding with an outcome-based reward. We introduce a rule-based reward function that jointly optimizes action-type selection and pixel-level localization accuracy. Experimental results on multiple datasets demonstrate the superiority of InReAct in both grounding and navigation tasks.

INREACT: An Inspire-Then-Reinforce Training Framework For Multimodal GUI Agent

Conceptual spaces represent entities and concepts using cognitively meaningful dimensions, typically referring to perceptual features. Such representations are widely used in cognitive science and have the potential to serve as a cornerstone for explainable AI. Unfortunately, they have proven notoriously difficult to learn, although recent LLMs appear to capture the required perceptual features to a remarkable extent. Nonetheless, practical methods for extracting the corresponding conceptual spaces are currently still lacking. While various methods exist for extracting embeddings from LLMs, extracting conceptual spaces also requires us to encode the underlying features. In this paper, we propose a strategy in which features (e.g. sweetness) are encoded by embedding the description of a corresponding prototype (e.g. a very sweet food). To improve this strategy, we fine-tune the LLM to align the prototype embeddings with the corresponding conceptual space dimensions. Our empirical analysis finds this approach to be highly effective.

Extracting Conceptual Spaces from LLMs Using Prototype Embeddings

Filtering data, particularly data scraped from the internet, has long been recognised as a means to improve model performance. Recent studies have shown that effective filters can be created by utilising Large Language Models (LLMs) to synthetically label data, which is then used to train smaller neural models for filtering purposes. However, this approach has been tested mainly in English. Our paper extends this approach to languages beyond English, including languages not officially supported by the LLM. We validate our results on the downstream task of NMT and demonstrate that our approach is effective at both filtering parallel text for translation quality and filtering for domain specificity. For training the filtering model, we experiment with two different objectives for finetuning pre-trained transformers, as well as an efficient approach based on *n*-gram language models.

Multilingual Data Filtering using Synthetic Data from Large Language Models

World models achieve remarkable success in predicting future states and planning in complex environments and Large Language Models (LLMs) serve as promising foundation to build general world models. However, their performances are usually constrained by the limited external knowledge to specific environments. Existing research attempts to enhance LLM-based world models through prompting or fine-tuning approaches, which are either requiring human knowledge or computationally extensive. Therefore, we introduce Retrieval-Augmented World Models (RAWM), a novel framework that leverages retrieval-augmented generation to efficiently integrate the external knowledge to LLM-based world models. Our main contributions are threefold: (i) We introduce a memory system and design an embedding model to retrieve relevant experiences as the in-context examples to improve the world model’s predictive accuracy. (ii) We develop a reinforcement learning (RL) training pipeline that fine-tunes a small MLP head on the pre-trained embedding model using Proximal Policy Optimization (PPO), further enhancing prediction performance. (iii) We conduct extensive experiments across three diverse environments, i.e., Game24, BlocksWorld, and BabyAI, demonstrating that RAWM consistently outperforms baseline models and exhibits strong generalizability. By leveraging the retrieval-augmented generation and the efficient RL training pipeline, RAWM dynamically utilizes relevant historical experiences and equips LLMs with environment-specific external knowledge without retraining, enabling more accurate and generalizable predictions.

Efficient Integration of External Knowledge to LLM-based World Models via Retrieval-Augmented Generation and Reinforcement Learning

Humour, as a complex and often esoteric language form, is derived from myriad aspects of life, whilst existing work on computational humour has focussed almost exclusively on short, self-contained pun-based jokes. In this work, we investigate whether the ability of Large Language Models (LLMs) to explain humour is dependent on the particular humour form. Specifically, we aim to compare models on simple pun-based jokes and more complex contemporary topical humour that requires knowledge of real-world entities and events. In doing so, we curate a dataset of 600 jokes, evenly split across 4 joke types, and manually write high-quality textual explanations, including heterographic and homographic puns, as well as contemporary internet humour and topical jokes where adequate understanding is reliant on reasoning beyond "common sense", rooted instead in world knowledge regarding news events, politics, pop culture, and more. Using this dataset, we compare the zero-shot abilities of a range of open- and closed-source LLMs to accurately and comprehensively explain jokes of different types, identifying key research gaps in the task of humour explanation. We find that none of the tested models are capable of reliably generating adequate explanations of all joke forms (inc. reasoning models), further highlighting the narrow focus of most works in computational humour processing on overly simple humour forms.

Comparing Apples to Oranges: A Dataset & Analysis of LLM Humour Understanding from Traditional Puns to Topical Jokes

While Large Language Models (LLMs) have demonstrated impressive performance across natural language generation tasks, their ability to generate truly creative content—characterized by novelty, diversity, surprise, and quality—remains limited. Existing methods for enhancing LLM creativity often focus narrowly on diversity or specific tasks, failing to address creativity’s multifaceted nature in a generalizable way. In this work, we propose Creative Preference Optimization (CrPO), a novel alignment method that injects signals from multiple creativity dimensions into the preference optimization objective in a modular fashion. We train and evaluate creativity-augmented versions of several models using CrPO and MuCE, a new large-scale human preference dataset spanning over 200,000 human-generated responses and ratings from more than 30 psychological creativity assessments. Our models outperform strong baselines, including GPT-4o, on both automated and human evaluations, producing more novel, diverse, and surprising generations while maintaining high output quality. Additional evaluations on NoveltyBench further confirm the generalizability of our approach. Together, our results demonstrate that directly optimizing for creativity within preference frameworks is a promising direction for advancing the creative capabilities of LLMs without compromising output quality.

Creative Preference Optimization

Virtual Teaching Assistants (VTAs) can reduce the workload of teaching teams in Asynchronous Learning Environments (ALEs) where timely, personalized support is often limited. As VTA systems grow more capable, rigorous and pedagogically sound evaluation becomes essential. Existing assessments often rely on surface-level metrics and lack sufficient grounding in educational theory, making it difficult to meaningfully compare the pedagogical effectiveness of VTA systems. To bridge this gap, we propose a pedagogically-oriented evaluation framework that is rooted in learning sciences and tailored to asynchronous forum discussions, a common VTA deployment context in ALE. We construct classifiers using expert annotations of VTA responses on a diverse set of forum posts. We evaluate the effectiveness of our classifiers, identifying approaches that improve accuracy as well as challenges that hinder generalization. Our work establishes a foundation for theory-driven evaluation of VTA systems, paving the way for more pedagogically effective AI in education.

Downloads

Next from EMNLP 2025

Enhancing SQL Table Acquisition with Reverse Engineering for Text-to-SQL

Stay up to date with the latest Underline news!

PRESENTATIONS

CONFERENCES

COMPANY

RESOURCES