China

Table question answering (TQA) requires accurate retrieval and reasoning over tabular data. Existing approaches attempt to retrieve relevant content before leveraging large language models (LLMs) for reasoning over long tables. However, these methods often fail to accurately retrieve contextually relevant data, resulting in information loss. They may also suffer from excessive encoding overhead and hallucination issues. In this paper, we propose TALON, a multi-agent framework designed for question answering over long tables. The framework features a planning agent that iteratively invokes a tool agent to access and manipulate tabular data based on intermediate feedback, progressively collecting the information necessary for answer generation, while a critic agent ensures accuracy and efficiency in tool usage and planning. To comprehensively assess the effectiveness of our framework, we introduce two benchmarks derived from the WikiTableQuestion and BIRD-SQL datasets, containing tables ranging from 50 to over 10,000 rows. Experimental results demonstrate that TALON achieves average accuracy improvements of 7.5% and 12.0% across all language models, establishing a new state-of-the-art in long-table question answering. Our code is publicly available at:https://anonymous.4open.science/r/TALON-uouxy.

EMNLP 2025

TALON: A Multi-Agent Framework for Long-Table Exploration and Question Answering

table qa

multi-agents

large language models

### Welcome!
The 62nd Annual Meeting of the Association for Computational Linguistics (ACL 2024) will take place in Bangkok, Thailand from August 11th to 16th, 2024. Our Virtual Poster Sessions will take place online Thursday, August 22, 2024.

You are required to register for this event. **Please register [here](https://2024.aclweb.org/registration). **

If you have already registered, please check your inbox for an email from Underline granting you access to ACL 2024 content.

Please register!

ACL 2024

The 62nd Annual Meeting of the Association for Computational Linguistics (ACL 2024) will take place in Bangkok, Thailand from August 11th to 16th, 2024. More information will be announced soon.

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." 

*Message from the General Chair, Dirk Hovy*

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

Large language models have many beneficial applications, but can they also be used to attack content-filtering algorithms in social media platforms? We investigate the challenge of generating adversarial examples to test the robustness of text classification algorithms detecting low-credibility content, including propaganda, false claims, rumours and hyperpartisan news. We focus on simulation of content moderation by setting realistic limits on the number of queries an attacker is allowed to attempt. Within our solution (TREPAT), initial rephrasings are generated by large language models with prompts inspired by meaning-preserving NLP tasks, such as text simplification and style transfer. Subsequently, these modifications are decomposed into small changes, applied through beam search procedure, until the victim classifier changes its decision. We perform (1) quantitative evaluation using various prompts and models, (2) targeted manual assessment of the generated text and (3) qualitative linguistic analysis. The results confirm the superiority of our approach in the constrained scenario, especially in case of long input text (news articles), where exhaustive search is not feasible.

Attacking Misinformation Detection Using Adversarial Examples Generated by Language Models

Table retrieval, essential for accessing information through tabular data, is less explored compared to text retrieval. The row/column structure and distinct fields of tables (including titles, headers, and cells) present unique challenges. For example, different table fields have varying matching preferences: cells may favor finer-grained (word/phrase level) matching over broader (sentence/passage level) matching due to their fragmented and detailed nature, unlike titles. This necessitates a table-specific retriever to accommodate the various matching needs of each table field. Therefore, we introduce a Table-tailored HYbrid Matching rEtriever (THYME), which approaches table retrieval from a field-aware hybrid matching perspective. Empirical results on two table retrieval benchmarks, NQ-TABLES and OTT-QA, show that THYME significantly outperforms state-of-the-art baselines. Comprehensive analyses have confirmed the differing matching preferences across table fields and validated the efficacy of THYME.

Tailoring Table Retrieval from a Field-aware Hybrid Matching Perspective

Knowledge graph question answering (KGQA) presents significant challenges due to the structural and semantic variations across input graphs. Existing works rely on Large Language Model (LLM) agents for graph traversal and retrieval; an approach that is sensitive to traversal initialization, as it is prone to entity linking errors and may not generalize well to custom (``bring-your-own'') KGs. We introduce BYOKG-RAG, a framework that enhances KGQA by synergistically combining LLMs with specialized graph retrieval tools. In BYOKG-RAG, LLMs generate critical graph artifacts (question entities, candidate answers, reasoning paths, and OpenCypher queries), and graph tools link these artifacts to the KG and retrieve relevant graph context. The retrieved context enables the LLM to iteratively refine its graph linking and retrieval, before final answer generation. By retrieving context from different graph tools, BYOKG-RAG offers a more general and robust solution for QA over custom KGs. Through experiments on five benchmarks spanning diverse KG types, we demonstrate that BYOKG-RAG outperforms the second-best graph retrieval method by 4.5% points while showing better generalization to custom KGs. We will open-source our framework as a multi-strategy toolkit for graph linking and retrieval.

BYOKG-RAG: Multi-Strategy Graph Retrieval for Knowledge Graph Question Answering

Recent progress in large language models (LLMs) has enabled the automated processing of lengthy documents even without supervised training on a task-specific dataset. Yet, their zero-shot performance in complex tasks as opposed to straightforward information extraction tasks remains suboptimal. One feasible approach for tasks with lengthy, complex input is to first summarize the document and then apply supervised fine-tuning to the summary. However, the summarization process inevitably results in some loss of information. In this study we present a method for processing the summaries of long documents aimed to capture different important aspects of the original document. We hypothesize that LLM summaries generated with different aspect-oriented prompts contain different information signals, and we propose methods to measure these differences. We introduce approaches to effectively integrate signals from these different summaries for supervised training of transformer models. We validate our hypotheses on a high-impact task—30-day readmission prediction from a psychiatric discharge—using real-world data from four hospitals, and show that our proposed method increases the prediction performance for the complex task of predicting patient outcome.

Aspect-Oriented Summarization for Psychiatric Short-Term Readmission Prediction

Robust supervised fine-tuned small Language Models (sLMs) often show high reliability but tend to undercorrect. They achieve high precision at the cost of low recall. Conversely, Large Language Models (LLMs) often show the opposite tendency, making excessive overcorrection, leading to low precision. To effectively harness the strengths of LLMs to address the recall challenges in sLMs, we propose Post-Correction via Overcorrection (PoCO), a novel approach that strategically balances recall and precision. PoCO first intentionally triggers overcorrection via LLM to maximize recall by allowing comprehensive revisions, then applies a targeted post-correction step via fine-tuning smaller models to identify and refine erroneous outputs. We aim to harmonize both aspects by leveraging the generative power of LLMs while preserving the reliability of smaller supervised models. Our extensive experiments demonstrate that PoCO effectively balances GEC performance by increasing recall with competitive precision, ultimately improving the overall quality of grammatical error correction.

Leveraging What's Overfixed: Post-Correction via LLM Grammatical Error Overcorrection

Implicit hate speech detection is challenging due to its subtlety and reliance on contextual interpretation rather than explicit offensive words. Current approaches rely on contrastive learning, which are shown to be effective on distinguishing hate and non-hate sentences. Humans, however, detect implicit hate speech by first identifying specific targets within the text and subsequently interpreting how these target relate to their surrounding context. Motivated by this reasoning process, we propose AmpleHate, a novel approach designed to mirror human inference for implicit hate detection. AmpleHate identifies explicit target using a pretrained Named Entity Recognition model and capture implicit target information via [CLS] tokens. It computes attention-based relationships between explicit, implicit targets and sentence context and then, directly injects these relational vectors into the final sentence representation. This amplifies the critical signals of target-context relations for determining implicit hate. Experiments demonstrate that AmpleHate achieves state-of-the-art performance, outperforming contrastive learning baselines by an average of 82.14% and achieve faster convergence. Qualitative analyses further reveal that attention patterns produced by AmpleHate closely align with human judgement, underscoring its interpretability and robustness.

AmpleHate: Amplifying the Attention for Versatile Implicit Hate Detection

Graph Neural Networks (GNNs) have emerged as powerful models for learning from graph-structured data. However, GNNs lack the inherent semantic understanding capability of rich textual node attributes, limiting their effectiveness in applications. On the other hand, we empirically observe that for existing GNN models, no one can consistently outperforms others across diverse datasets. In this paper, we study whether LLMs can act as an ensembler for multi-GNNs and propose the LensGNN model. The model first aligns multiple GNNs, mapping the representations of different GNNs into the same space. Then, through LoRA fine-tuning, it aligns the space between the GNN and the LLM, injecting graph tokens and textual information into LLMs. This allows LensGNN to ensemble multiple GNNs and take advantage of the strengths of LLM, leading to a deeper understanding of both textual semantic information and graph structural information. The experimental results show that LensGNN outperforms existing models. This research advances text-attributed graph ensemble learning by providing a robust and superior solution for integrating semantic and structural information. We provide our code and data here: https://anonymous.4open.science/r/EnsemGNN-E267/.

Can Large Language Models Act as Ensembler for Multi-GNNs?

The integration of Large Language Models (LLMs) with retrieval systems has shown promising potential in retrieving documents (docs) or advertisements (ads) for a given query. Existing LLM-based retrieval methods generate numeric or content-based DocIDs to retrieve docs/ads. However, the one-to-few mapping between numeric IDs and docs, along with the time-consuming content extraction, leads to semantic inefficiency and limits the scalability of existing methods on large-scale corpora. In this paper, we propose the **R**eal-time **A**d **RE**trieval (RARE) framework, which leverages LLM-generated text called Commercial Intentions (CIs) as an intermediate semantic representation to directly retrieve ads for queries in real-time. These CIs are generated by a customized LLM injected with commercial knowledge, enhancing its domain relevance. Each CI corresponds to multiple ads, yielding a lightweight and scalable set of CIs. RARE has been implemented in a real-world online system, handling daily search volumes in billions. The online implementation has yielded significant benefits: a 5.04\% increase in consumption, a 6.37\% rise in Gross Merchandise Volume (GMV), a 1.28\% enhancement in click-through rate (CTR) and a 5.29\% increase in shallow conversions. Extensive offline experiments show RARE's superiority over ten competitive baselines in four major categories.

Real-time Ad Retrieval via LLM-generative Commercial Intention for Sponsored Search Advertising

Language models can be used to provide interactive, personalized student feedback in educational settings. However, real-world deployment faces three key challenges: privacy concerns, limited computational resources, and the need for pedagogically valid responses. These constraints require small, open-source models that can run locally and reliably ground their outputs in correct information. We introduce SCRIBE, a framework for multi-hop, tool-augmented reasoning designed to generate valid responses to student questions about feedback reports. SCRIBE combines domain-specific tools with a self-reflective inference pipeline that supports iterative reasoning, tool use, and error recovery. We distill these capabilities into 3B and 8B models via two-stage LoRA fine-tuning on synthetic GPT-4o-generated data. Evaluation using a human-aligned GPT-Judge and a user study with 108 students shows that SCRIBE matches or exceeds the perceived quality of much larger models, demonstrating its viability for low-resource, privacy-sensitive educational applications.

SCRIBE: Structured Chain Reasoning for Interactive Behaviour Explanations using Tool Calling

Large language models (LLMs) have gained increasing attention in recommender systems, but their inherent hallucination issues significantly compromise the accuracy and reliability of recommendation results. Existing LLM-based recommender systems predominantly rely on standard fine-tuning methodologies, often ignoring hallucination issues during the fine-tuning process. To address this challenge, we propose Logit Space Constraints Fine-Tuning (LCFT), a novel fine-tuning framework designed to mitigate hallucination in LLM-based recommenders. Specifically, LCFT takes as input semantically positive and negative instruction pairs and incorporates Kullback–Leibler (KL) divergence into the training objective to explicitly maximise their distributional disparity in the logit space. By conducting such logit space-constrained fine-tuning, LCFT encourages more distinguishable and semantically grounded representations, thereby reducing the model’s susceptibility to hallucination. Extensive experiments on two recommendation models with distinct LLM backbones and four real-world datasets demonstrate that LCFT consistently reduces hallucination and enhances recommendation performance.

Downloads

Next from EMNLP 2025

Attacking Misinformation Detection Using Adversarial Examples Generated by Language Models

Similar lecture

Is Table Retrieval a Solved Problem? Exploring Join-Aware Multi-Table Retrieval

Stay up to date with the latest Underline news!

PRESENTATIONS

CONFERENCES

COMPANY

RESOURCES