AI conference papers

AI conference paper library

Search papers from 30+ top conferences, including NeurIPS, ICLR, ICML, CVPR, ACL and AAAI. Filter by keyword, then open any paper for its abstract, authors and links.

ACL 2026

Context as a Tool: Context Management for Long-Horizon SWE-Agents

Shukai Liu, Bo Jiang, Jian Yang, Yizhi Li, Jinyang Guo, Xianglong Liu, Bryan Dai

Agents based on large language models have recently shown strong potential on real-world software engineering (SWE) tasks that require long-horizon interaction with repository-scale codebases. However, most existing agents rely on append-only context maintenance or passively triggered compression heuristics, which often lead to context explosion, semantic drift, and degraded reasoning in long-running interactions. We propose Cat, a new context management paradigm that elevates context maintenance to a callable tool integrated into the decision-making process of agents. Cat formalizes a structured context workspace consisting of stable task semantics, condensed long-term memory, and high-fidelity short-term interactions, and enables agents to proactively compress historical trajectories into actionable summaries at appropriate milestones. To support context management for SWE-agents, we propose a trajectory-level supervision framework, CaT-Generator, based on an offline data construction pipeline that injects context-management actions into complete interaction trajectories. Using this framework, we train a context-aware model, SWE-Compressor. Experiments on SWE-Bench-Verified demonstrate that SWE-Compressor reaches a 57.6% solved rate and significantly outperforms ReAct-based agents and static compression baselines, while maintaining stable and scalable long-horizon reasoning under a bounded context budget.

ACL 2026

EvoCoT: Overcoming the Exploration Bottleneck in Reinforcement Learning for LLMs

Huanyu Liu, Jia Li, Yihong Dong, Chang Yu, Taozhi Chen, Lecheng Wang, Yongding Tao, Bin Gu, Ge Li

Reinforcement learning with verifiable reward (RLVR) has become a promising paradigm for post-training large language models (LLMs) to improve their reasoning capability. However, when the rollout accuracy is low on hard problems, the reward becomes sparse, limiting learning efficiency and causing exploration bottlenecks. Existing approaches either rely on teacher models for distillation or filter out difficult problems, which limits scalability or restricts reasoning improvement through exploration.We propose EvoCoT, a self-evolving curriculum learning framework based on two-stage chain-of-thought (CoT) reasoning optimization. EvoCoT constrains the exploration space by self-generating and verifying CoT trajectories, then gradually shortens CoT steps to expand the space in a controlled way. The framework enables LLMs to stably learn from initially unsolved hard problems under sparse rewards. We apply EvoCoT to multiple LLM families, including Qwen, DeepSeek, and Llama. Experiments show that EvoCoT enables LLMs to solve previously unsolved problems, improves reasoning capability without external CoT supervision, and is compatible with various RL fine-tuning methods. We release the source code to support future research.

ACL 2026

CARE-CR: Context-Aware Routing and Expert Fusion for Multi-Preference Cognitive Restructuring

Hongzhi Qi, Liangcheng Wang, Yijing YU, Jianqiang Li, Bing Xiang Yang, Qing Zhao

While Large Language Models (LLMs) offer promising avenues for automated cognitive restructuring in mental health settings, current approaches predominantly focus on superficial positive reframing and lack the adaptability to balance conflicting therapeutic dimensions, such as empathy and rationality. To address these deficiencies, we propose CARE-CR, a context-aware framework that implements a decoupled optimization paradigm. We first train expert policies specialized for distinct therapeutic attributes rather than relying on a monolithic alignment strategy. To mitigate expert data scarcity, we introduce Dimension-Guided Hierarchical Monte Carlo Tree Search (DG-HMCTS) for data-efficient preference augmentation. At inference, a context-aware routing module dynamically predicts optimal preference weights to fuse expert outputs based on the user’s specific distress context. Extensive experiments demonstrate that CARE-CR achieves consistent improvements over strong baselines across multiple evaluation dimensions, including diagnostic accuracy, contextual appropriateness, task effectiveness, and overall helpfulness, while enabling controllable cognitive restructuring generation. The dataset and code are publicly available at https://github.com/HongzhiQ/CARE-CR.

ACL 2026

A Dual-View Analysis of Multiple Languages in Colonial Newspapers

Zhan Su, Xiaoya Chen, Fengran Mo, Ida L. Vos, Prayag Tiwari, Yazhou Zhang, Qian Zheng, Natália da Silva Perez

Historical newspapers from the colonial period offer valuable evidence of how racializing language evolved over time. However, there are challenges in studying this type of historical data: 1) Data scarcity: acquiring large, annotated historical datasets is difficult, hindering the possibility of analyzing racialization comprehensively; 2) Digitized materials frequently contain Optical Character Recognition (OCR) errors and other types of noise that complicate text extraction and computational analysis; 3) Colonial newspapers are often multilingual and written in archaic prose, hindering the effectiveness of NLP tools developed for modern, single language texts. This paper addresses these challenges by conducting a dual-view, jointly studying multilingual event extraction and temporal semantic shift tasks. Specifically, we introduce a contextual question answering (CQA) and a visual question answering (VQA) derived from eighteenth- and nineteenth-century colonial newspapers. Content-wise, we focus on how enslaved people were described by enslavers as well as how they articulated their own condition through QA pairs of newspapers written in Dutch, English-French, and Spanish. Our results show that LLMs are still limited for low-resource VQA tasks. For temporal semantic change, we train temporal word embedding with a compass. The study concludes that racialization is a fluid process of linguistic recalibration where the decline of slavery merely shifted the language of control onto new categories of labor and identity.

ACL 2026

LinguaGame: A Linguistically Grounded Game-Theoretic Paradigm for Multi-Agent Dialogue Generation

Yuxiao Ye, Yiming Zhang, Yiran Rex Ma, Huiyuan Xie, Huining Zhu, Zhiyuan Liu

Large Language Models (LLMs) have enabled Multi-Agent Systems (MASs) where agents interact through natural language to solve complex tasks or simulate multi-party dialogues. Recent work on LLM-based MASs has mainly focused on architecture design, such as role assignment and workflow orchestration. In contrast, this paper targets the interaction process itself, aiming to improve agents’ communication efficiency by helping them convey their intended meaning more effectively through language. To this end, we propose LinguaGame, a linguistically-grounded game-theoretic paradigm for multi-agent dialogue generation. Our approach models dialogue as a signalling game over communicative intents and strategies, solved with a training-free equilibrium approximation algorithm for inference-time decision adjustment. Unlike prior game-theoretic MASs, whose game designs are often tightly coupled with task-specific objectives, our framework relies on linguistically informed reasoning with minimal task-specific coupling. Specifically, it treats dialogue as intentional and strategic communication, requiring agents to infer what others aim to achieve (intents) and how they pursue those goals (strategies). We evaluate our framework in simulated courtroom proceedings and debates, with human expert assessments showing significant gains in communication efficiency. We release our code and data on GitHub.

ACL 2026

Traces in the Brain: Neural Evidence for Syntactic Movement in English and Chinese

Yuhan Huang, Zhengwu Ma, Yuqi Jin, Beth Chan, Zheng Shen, Jackie Yan-Ki Lai, John T Hale, Jixing Li

Syntactic movement is a core concept in generative linguistics to account for word-order variation and long-distance dependencies, but its psychological and neurobiological status remains debated. Here, we test the neural reality of movement in English and Chinese by correlating brain activity during naturalistic listening with syntactic node counts, traces and word embeddings derived from X-bar style tree annotations. We find that deep structure significantly predicts neural responses in English but not in Chinese, providing partial support for movement-based accounts while revealing clear cross-linguistic differences.

ACL 2026

Distributional Alignment for Large Language Models under Domain Shift

Viet-Thanh Pham, Lizhen Qu, Zhuang Li, Gholamreza Haffari

Distributional alignment enables large language models (LLMs) to predict how a target population distributes its responses across answer options, rather than collapsing disagreement into a single consensus answer. However, existing LLM-based distribution prediction is often unstable and degrades under cultural and domain shift. Token score-based estimates can change with minor option wording or formatting, response sampling-based estimates are expensive and sensitive to prompts and decoding settings, and directly generated distributions are frequently miscalibrated.We propose Evi-DA, an evidence-based alignment technique that improves the fidelity and robustness of LLM-based distribution estimation under domain and cultural shift. Given a target country and a multiple-choice question, Evi-DA retrieves related World Values Survey items and their answer distributions, predicts a coarse Welzel value signature for each option, and infers the country-conditioned answer distribution in a structured format. We train the LLMs using a two-stage pipeline, where reinforcement learning optimizes survey-derived rewards that encourage accurate intermediate value predictions, faithful final distributions, well-formed structured outputs, and reduced cultural bias. Across in-domain and out-of-domain benchmarks and multiple open-source backbones, Evi-DA reduces Jensen-Shannon divergence between predicted and gold distributions relative to strong baselines, with average relative improvements of up to 44%.

ACL 2026

Targeting the Needle, Ignoring the Haystack: Anchoring Crucial Cues for Evolving Scam Call Detection via an LLM-Assisted Classifier

Tong Wu, Qinliang Su, Jianxing Yu, Bo Liang, Minhua Huang

Automatic detection of fraudulent voice calls is essential for online service platforms but faces significant challenges due to the scarcity of labeled data and the continuous evolution of conversational contexts. Standard supervised methods often fail to generalize, as they tend to overfit to variable background narratives rather than capturing the core deceptive intent. In this paper, we propose a lightweight framework that anchors detection on Semantic Primitives, a set of stable, interpretable evidentiary cues derived from expert knowledge. Our approach decomposes the fraud detection task into two distinct stages: identifying the presence of these predefined semantic signals within the transcript, and deriving a final verdict through a logical combination of the detected cues. By explicitly prioritizing stable evidence over diverse conversational noise, this framework ensures that decisions are based on verifiable fraud tactics rather than spurious correlations. Experimental results demonstrate that our method achieves superior robustness and efficiency compared to traditional baselines, particularly in scenarios with shifting service contexts.

ACL 2026

CLewR: Curriculum Learning with Restarts for Machine Translation Preference Learning

Alexandra Dragomir, Florin Brad, Radu Tudor Ionescu

Large language models (LLMs) have demonstrated competitive performance in zero-shot multilingual machine translation (MT). Some follow-up works further improved MT performance via preference optimization, but they leave a key aspect largely underexplored: the order in which data samples are given during training. We address this topic by integrating curriculum learning into various state-of-the-art preference optimization algorithms to boost MT performance. We introduce a novel curriculum learning strategy with restarts (CLewR), which reiterates easy-to-hard curriculum multiple times during training to effectively mitigate the catastrophic forgetting of easy examples. We demonstrate consistent gains across several model families (Gemma2, Qwen2.5, Llama3.1) and preference optimization techniques. We publicly release our code at https://github.com/alexandra-dragomir/CLewR.

ACL 2026

Interactive Semantic Parsing with Reinforcement Learning for Knowledge Graph Reasoning

Yurun Song, Xiangqing Shen, Jianfei Yu, Rui Xia

While large language models (LLMs) have achieved remarkable success, their reliability in knowledge-intensive tasks is often compromised by factual hallucinations. Integrating Knowledge Graphs (KGs) addresses this issue; however, existing approaches typically rely on simple graph traversal.This paradigm decouples topological navigation from logical operations (e.g., temporal filtering, aggregation), leading to imprecise retrieval and heavy post-processing burdens.Although semantic parsing offers a solution by grounding reasoning in logical forms, it traditionally suffers from a dependency on scarce supervised annotations.To bridge this gap, we propose Interactive Semantic Parsing, a framework that formulates reasoning as the sequential generation of executable logical clauses. This design allows logical constraints to be dynamically interleaved with graph search, while optimizing via reinforcement learning with only final answer feedback eliminates the need for gold program annotations.To tackle the sparse reward challenge in the vast symbolic space, we introduce a distance-aware process reward to evaluate intermediate steps based on their topological proximity to the answer.Experimental results on WebQSP and CWQ demonstrate that our method achieves state-of-the-art performance, particularly on complex queries, validating the effectiveness of our dense reward signal in enabling robust reasoning without supervision.Our code is available at https://github.com/NUSTM/ISP-KGR.

ACL 2026

LifeSim: Long-Horizon User Life Simulator for Personalized Assistant Evaluation

Feiyu Duan, Xuanjing Huang, Zhongyu Wei

The rapid advancement of large language models (LLMs) has accelerated progress toward universal AI assistants. However, existing benchmarks for personalized assistants remain misaligned with real-world user-assistant interactions, failing to capture the complexity of external contexts and users’ cognitive states. To bridge this gap, we propose LifeSim, a user simulator that models user cognition through the Belief-Desire-Intention (BDI) model within physical environments for coherent life trajectories generation, and simulates intention-driven user interactive behaviors. Based on LifeSim, we introduce LifeSim-Eval, a comprehensive benchmark for multi-scenario, long-horizon personalized assistance. LifeSim-Eval covers 8 life domains and 1,200 diverse scenarios, and adopts a multi-turn interactive method to assess models’ abilities to complete explicit and implicit intentions, recover user profiles, and produce high-quality responses. Under both single-scenario and long-horizon settings, our experiments reveal that current LLMs face significant limitations in handling implicit intention and long-term user preference modeling.

ACL 2026

Fine-Grained Data Ordering Improves Fine-Tuning for Large Language Models

Xiaomeng Hu, Yixuan Tang, Haoze Li, Hao Chen, Qi Zhang, Zhanming Shen, Yiming Zhang, Haobo Wang, Junbo Zhao

With the rapid progress of large language models (LLMs), aligning a general-purpose model with downstream tasks through fine-tuning has become a central research focus. Selecting only high-quality examples for training has been shown to be one of the most effective ways to improve fine-tuning performance. However, prior work concentrates almost exclusively on data preprocessing: filtering and cleaning data before training begins. While the order and composition of training data during training have received little fine-grained attention. To fill this gap, our work proposed Fine-Grained Order Fine-Tuning, a fine-grained scheduling method of data order in epochs. Drawing on curriculum-learning principles, FOT defines data difficulty based on the relevance between the data and the model, and then performs dynamic scheduling of the training order in each epoch according to the difficulty. On both large-scale continued pre-training and small-scale supervised fine-tuning experiments, FOT has achieved an average 2.4% improvement over baselines. Our study offers a new perspective on data governance in the fine-tuning phase.

ACL 2026

MimicLM: Zero-Shot Voice Imitation through Autoregressive Modeling of Pseudo-Parallel Speech Corpora

Tao Feng, Yuxiang Wang, Yuancheng Wang, Xueyao Zhang, Dekun Chen, Chaoren Wang, Xun Guan, Zhizheng Wu

Voice imitation aims to transform source speech to match a reference speaker’s timbre and speaking style while preserving linguistic content. A straightforward approach is to train on triplets of (source, reference, target), where source and target share the same content but target matches the reference’s voice characteristics, yet such data is extremely scarce. Existing approaches either employ carefully designed disentanglement architectures to bypass this data scarcity or leverage external systems to synthesize pseudo-parallel training data. However, the former requires intricate model design, and the latter faces a quality ceiling when synthetic speech is used as training targets. To address these limitations, we propose MimicLM, which takes a novel approach by using synthetic speech as training sources while retaining real recordings as targets. This design enables the model to learn directly from real speech distributions, breaking the synthetic quality ceiling. Building on this data construction approach, we incorporate interleaved text-audio modeling to guide the generation of content-accurate speech and apply post-training with preference alignment to mitigate the inherent distributional mismatch when training on synthetic data. Experiments demonstrate that MimicLM achieves superior voice imitation quality with a simple yet effective architecture, significantly outperforming existing methods in naturalness while maintaining competitive similarity scores across speaker identity, accent, and emotion dimensions.

ACL 2026

Veri-R1: Toward Precise and Faithful Claim Verification via Online Reinforcement Learning

Qi He, Cheng Qian, Xiusi Chen, Bingxiang He, Yi R. Fung, Heng Ji

Claim verification with large language models (LLMs) has recently attracted growing attention, due to their strong reasoning capabilities and transparent verification processes compared to traditional answer-only judgments. However, existing approaches to online claim verification, which requires iterative evidence retrieval and reasoning, still mainly rely on prompt engineering or pre-designed reasoning workflows, without unified training to improve necessary skills. Therefore, we introduce Veri-R1, an online reinforcement learning (RL) framework that enables an LLM to interact with a search engine and to receive reward signals that explicitly shape its planning, retrieval, and reasoning behaviors. The dynamic interaction between models and retrieval systems more accurately reflects real-world verification scenarios and fosters comprehensive verification skills. Empirical results show that Veri-R1 improves joint accuracy by up to 30% and doubles evidence score, often surpassing its larger-scale model counterparts. Ablation studies further reveal the impact of reward components, and the link between output logits and label accuracy. Our results highlight the effectiveness of online RL for precise and faithful claim verification, and provide a foundation for future research.

ACL 2026

EmoMM: Benchmarking and Steering MLLM for Multimodal Emotion Recognition under Conflict and Missingness

Yueru Sun, Yimeng Zhang, Haoyu Gu, Nuo Chen, Dong She, Xianrong Yao, Yang Gao, Zhanpeng Jin

Multimodal Emotion Recognition (MER) is critical for interpreting real-world interactions. While Multimodal Large Language Models (MLLM) have shown promise in MER, their internal decision-making mechanisms under modality conflict and missingness remain largely underexplored. In this paper, to systematically investigate these behaviors, we introduce EmoMM, a comprehensive benchmark featuring modality-aligned, conflict, and missing subsets. Through extensive evaluation, we uncover a Video Contribution Collapse (VCC) phenomenon, where MLLM marginalize video evidence due to high token redundancy and modality preferences. To address this, we propose Conflict-aware Head-level Attention Steering (CHASE), a lightweight mechanism that detects modality conflicts and performs inference-time attention steering, effectively mitigating decision bias without retraining the backbone. Experimental results demonstrate that CHASE consistently improves performance across various settings, significantly enhancing the reliability of MLLM in complex affective scenarios.

ACL 2026

Modeling Multi-Dimensional Cognitive States in Large Language Models under Cognitive Crowding

Lin Zhong, Siyu Zhu, Zizhen Yuan, Jinhao Cui, Xinyang Zhao, Lingzhi Wang, Hao Chen, Qing Liao

Modeling human cognitive states is essential for advanced artificial intelligence. Existing Large Language Models (LLMs) mainly address isolated tasks such as emotion analysis or stance detection, and fail to capture interactions among cognitive dimensions defined in psychology, including emotion, thinking style, stance, and intention. To bridge this gap, we construct CognitiveBench, the first benchmark with unified annotations across the above four dimensions. Experiments on CognitiveBench show that although LLMs perform well on single dimension tasks, their performance drops sharply in joint multi-dimensional modeling. Using Gromov-hyperbolicity analysis, we find that CognitiveBench exhibits a strong hierarchical structure. We attribute the performance bottleneck to “Cognitive Crowding”, where hierarchical cognitive states require exponential representational space, while the Euclidean space of LLMs grows only polynomially, causing representation overlap and degraded performance. To address this mismatch, we propose HyCoLLM, which models cognitive states in hyperbolic space and aligns LLM representations via Hyperbolic Guided Alignment Tuning. Results show that HyCoLLM substantially improves multi-dimensional cognitive understanding, allowing 8B parameter model to outperform strong baselines, including GPT-4o. Our code is available at https://anonymous.4open.science/r/HycoLLM.

ACL 2026

Mitigating Hallucinations in VLMs: Enhancing Visual Attention via Head-Wise Perturbation

Zhenghua Wang, Yixin Wu, Feiran Zhang, Qi Qian, Changze Lv, Xuanjing Huang, Xiaoqing Zheng

Vision–Language Models (VLMs) have demonstrated strong capabilities in tasks that require joint understanding of text and images. However, as many VLMs are built upon pre-trained large language models, they often over-rely on linguistic priors at the expense of visual features, causing persistent hallucinations. We observe that these hallucinations stem not only from insufficient visual attention but also from imbalanced activation profiles across attention heads, while hallucinated samples tend to disproportionately activate heads that fail to capture visual cues. To promote a more balanced attention distribution, we propose HWP, a strategy that incorporates head-wise attention perturbation via continuous multiplicative noise, coupled with a visual-guided loss focused on vision-sensitive text tokens. Beyond simply strengthening visual grounding, this design encourages a broader set of attention heads to engage with visual signals, thereby alleviating information loss caused by activation concentration on a few dominant heads. Consistent gains across different architectures and scales on multiple benchmarks demonstrate the effectiveness and robustness of our approach in mitigating VLM hallucinations.

ACL 2026

DRIFT: Transferring Reasoning Priors for Efficient MLLM Fine-Tuning

Chao Huang, Zeliang Zhang, Jiang Liu, Ximeng Sun, Jialian Wu, Xiaodong Yu, Ze Wang, Chenliang Xu, Emad Barsoum, Zicheng Liu

Multimodal large language models (MLLMs) have made rapid progress, yet their reasoning ability often lags behind strong text-only LLMs. Bridging this gap typically requires large-scale multimodal reasoning data or reinforcement learning, incurring substantial cost. An appealing alternative is parameter-space model merging between reasoning-enhanced LLMs and MLLMs, but we show that naive merging is fragile: its effectiveness varies widely across model families and can significantly degrade performance (e.g., for Qwen-based MLLMs). We propose Directional Reasoning Injection for Fine-Tuning (DRIFT), a lightweight method that transfers reasoning knowledge in the gradient space while preserving multimodal alignment. DRIFT precomputes a reasoning prior from the parameter differences between text-only reasoning experts and multimodal models, and uses it to bias gradients during supervised fine-tuning. This design retains the simplicity of standard SFT pipelines while enabling efficient and stable reasoning transfer. Experiments on multimodal reasoning benchmarks, including MathVista and MathVerse, show that DRIFT consistently outperforms naive merging and standard SFT, and matches or surpasses training-intensive methods with substantially lower data and compute.

ACL 2026

FACT-E: Causality-Inspired Evaluation for Trustworthy Chain-of-Thought Reasoning

Yuxi Sun, Aoqi Zuo, Haotian Xie, Wei Gao, Mingming Gong, Jing Ma

Chain-of-Thought (CoT) prompting has improved LLM reasoning, but models often generate explanations that appear coherent while containing unfaithful intermediate steps. Existing self-evaluation approaches are prone to inherent biases: the model may confidently endorse coherence even when the step-to-step implication is not valid, leading to unreliable faithfulness evaluation. We propose FACT-E, a causality-inspired framework for evaluating CoT quality. FACT-E uses controlled perturbations as an instrumental signal to separate genuine step-to-step dependence from bias-driven artifacts, producing more reliable faithfulness estimates (intra-chain faithfulness). To select trustworthy trajectories, FACT-E jointly considers intra-chain faithfulness and CoT-to-answer consistency, ensuring that selected chains are both faithful internally and supportive of the correct final answer. Experiments on GSM8K, MATH, and CommonsenseQA show that FACT-E improves reasoning-trajectory selection and yields stronger in-context learning exemplars. FACT-E also reliably detects flawed reasoning under noisy conditions, providing a robust metric for trustworthy LLM reasoning.

ACL 2026

Breaking the "Provable Security": Detecting Finite-Precision Artifacts in LLM-based Steganography via Low-Probability Vanishing

Wenzhao Cao, Yaofei Wang, Donghui Hu

Recent advances in Large Language Models have fostered a new class of generative linguistic steganography, claim “provably secure” by theoretically aligning the steganographic distribution with the language model’s natural distribution. We challenge this premise by exposing Low-Probability Vanishing (LPV), an inevitable vulnerability arising from finite-precision arithmetic. To exploit this, we propose RRNs-HT, a novel steganalysis framework based on Representative Random Numbers and Hypothesis Testing, which transforms the detection task from semantic classification to a statistical audit of the sampling mechanism. Crucially, unlike previous work that contrasts machine text against human text, we validate our method in a rigorous homologous setting to strictly isolate sampling artifacts. Experiments demonstrate that RRNs-HT effectively breaks the security of AC and Meteor with high detection accuracy, whereas state-of-the-art semantic steganalyzers degrade to random guessing. Our findings prove that theoretical security is unattainable in practice without addressing finite-precision leakage.

Keep browsing

More papers from top AI conferences

30+ top conferences, by conference, year, and topic.