Composed Image Retrieval (CIR) enables users to search for images using multimodal queries that combine text and reference images. While metric learning methods have shown promise, they rely on deterministic point embeddings that fail to capture the inherent uncertainty in the input data, in which user intentions may be imprecisely specified or open to multiple interpretations. We address this challenge by reformulating CIR through our proposed Composed Probabilistic Embedding (CoPE) framework, which represents both queries and targets as Gaussian distributions in latent space rather than fixed points. Through careful design of probabilistic distance metrics and hierarchical learning objectives, CoPE explicitly captures uncertainty at both instance and feature levels, enabling more flexible, nuanced, and robust matching that can handle polysemy and ambiguity in search intentions. Extensive experiments across multiple benchmarks demonstrate that CoPE effectively quantifies both quality and semantic uncertainties within Composed Image Retrieval, achieving state-of-the-art performance on recall rate. Code: https://github.com/tanghme0w/ACL25-CoPE.

The rapid advancement of large language models has intensified public concerns about the potential misuse. Therefore, it is important to build trustworthy AI-generated text detection systems. Existing methods neglect stylistic modeling and mostly rely on static thresholds, which greatly limits the detection performance. In this paper, we propose the Mixture of Stylistic Experts (MoSEs) framework that enables stylistics-aware uncertainty quantification through conditional threshold estimation. MoSEs contain three core components, namely, the Stylistics Reference Repository (SRR), the Stylistics-Aware Router (SAR), and the Conditional Threshold Estimator (CTE). For input text, SRR can activate the appropriate reference data in SRR and provide them to CTE. Subsequently, CTE jointly models the linguistic statistical properties and semantic features to dynamically determine the optimal threshold. With a discrimination score, MoSEs yields prediction labels with the corresponding confidence level. Our framework achieves an average improvement 11.34% in detection performance compared to baselines. More inspiringly, MoSEs shows a more evident improvement 39.15% in the low-resource case. Our code is available at https://github.com/creator-xi/MoSEs.

The misuse of large language models (LLMs), such as academic plagiarism, has driven the development of detectors to identify LLM-generated texts. To bypass these detectors, paraphrase attacks have emerged to purposely rewrite these texts to evade detection. Despite the success, existing methods require substantial data and computational budgets to train a specialized paraphraser, and their attack efficacy greatly reduces when faced with advanced detection algorithms. To address this, we propose Contrastive Paraphrase Attack (CoPA), a training-free method that effectively deceives text detectors using off-the-shelf LLMs. The first step is to carefully craft instructions that encourage LLMs to produce more human-like texts. Nonetheless, we observe that the inherent statistical biases of LLMs can still result in some generated texts carrying certain machine-like attributes that can be captured by detectors. To overcome this, CoPA constructs an auxiliary machine-like word distribution as a contrast to the human-like distribution generated by the LLM. By subtracting the machine-like patterns from the human-like distribution during the decoding process, CoPA is able to produce sentences that are less discernible by text detectors. Our theoretical analysis suggests the superiority of the proposed attack. Extensive experiments validate the effectiveness of CoPA in fooling text detectors across various scenarios.

Large Language Models (LLMs) fine-tuned via Reinforcement Learning from Human Feedback (RLHF) and Reinforcement Learning with Verifiable Rewards (RLVR) significantly improve the alignment of human-AI values and further raise the upper bound of AI capabilities, particularly in reasoning-intensive, long-context Chain-of-Thought (long-CoT) tasks. However, existing RLHF (or RLVR) frameworks commonly face challenges such as inference bottlenecks and complexity barriers, restricting their accessibility for newcomers. To bridge this gap, we introduce OpenRLHF, a user-friendly, scalable, and easy-to-learn open-source RLHF framework built upon Ray, vLLM, DeepSpeed, and HuggingFace Transformers, featuring a simplified design, clear code structure, and comprehensive documentation to facilitate entry for researchers and practitioners. Experimental results show that OpenRLHF achieves superior training efficiency with speedups ranging from 1.22× to 1.68× across different model sizes compared to state-of-the-art frameworks, while requiring significantly fewer lines of code for implementation. OpenRLHF is publicly available at https://github.com/OpenRLHF/OpenRLHF, and has already been adopted by leading institutions to accelerate RLHF research and learning.

We present a Large Language Model (LLM) based system for question answering (QA) over tabular data that leverages multi-turn prompting to automatically generate executable Pandas functions. Our framework decomposes the problem into three key steps: (1) Answer Type Identification, where the system identifies the expected format of the response (e.g., boolean, number, category); (2) Pandas Function Generation, which generates a corresponding Pandas function using table metadata and in-context examples, and (3) Error Correction and Regeneration, where iteratively refining the function based on error feedback from executions. Evaluations on the SemEval-2025 Task 8 Tabular QA benchmark (Grijalba et al., 2024) demonstrate that our multi-turn approach significantly outperforms single-turn prompting models in exact match accuracy by 7.3%. The proposed system not only improves code generation robustness but also paves the way for enhanced and adaptability in table-QA reasoning tasks. Our implementation is available at https://github.com/Gyyz/Question_Answering-over-Tabular-Data.

This study presents an approach for detecting connective language—defined as language that facilitates engagement, understanding, and conversation—from social media discussions. We developed and evaluated two types of classifiers: BERT and GPT-3.5 turbo. Our results demonstrate that the BERT classifier significantly outperforms GPT-3.5 turbo in detecting connective language. Furthermore, our analysis confirms that connective language is distinct from related concepts measuring discourse qualities, such as politeness and toxicity. We also explore the potential of BERT-based classifiers for platform-agnostic tools. This research advances our understanding of the linguistic dimensions of online communication and proposes practical tools for detecting connective language across diverse digital environments.

Temporal knowledge graph (TKG) reasoning aims to predict missing facts based on a given history. Most of the existing methods unifiedly model the evolution process of different events and ignore their inherent asynchronous characteristics, resulting in suboptimal performance. To tackle this challenge, we propose a Natural Evolution-based Dual-level Aggregation framework (NEDA) for TKG reasoning. Specifically, we design a natural division strategy to group TKGs into different patches according to the occurrence of a given target entity. Then, we present a dual-level aggregation scheme to extract local representations from information within patches and then aggregate these representations with adaptive weights as the final entity representations. By assigning varying weights to different patches, this aggregation scheme can incorporate the asynchronous characteristics of event evolution for representation computation, thus enhancing prediction performance. Extensive experiments demonstrate the significant improvement of our proposed model.

Large language models have demonstrated exceptional capability in natural language understanding and generation. However, their generation speed is limited by the inherently sequential nature of their decoding process, posing challenges for real-time applications. This paper introduces Lexical Unit Decoding (LUD), a novel decoding methodology implemented in a data-driven manner, accelerating the decoding process without sacrificing output quality. The core of our approach is the observation that a pre-trained language model can confidently predict multiple contiguous tokens, forming the basis for a lexical unit, in which these contiguous tokens could be decoded in parallel. Extensive experiments validate that our method substantially reduces decoding time while maintaining generation quality, i.e., 33% speed up on natural language generation with no quality loss, and 30% speed up on code generation with a negligible quality loss of 3%. Distinctively, LUD requires no auxiliary models and does not require changes to existing architectures. It can also be integrated with other decoding acceleration methods, thus achieving an even more pronounced inference efficiency boost. We posit that the foundational principles of LUD could define a new decoding paradigm for future language models, enhancing their applicability for a broader spectrum of applications. All codes are be publicly available at https://github.com/tjunlp-lab/Lexical-Unit-Decoding-LUD-.

Emotion Recognition in Conversations (ERC) is a well-studied task with numerous potential real-world applications. However, existing ERC models trained on the MELD dataset derived from TV series, struggle when applied to daily conversation datasets. A closer examination of the datasets unveils the prevalence of linguistic artifacts such as repetitions and interjections in TV scripts, which ERC models may exploit when making predictions. To address this issue, we explore two techniques aimed at reducing the reliance of ERC models on these artifacts: 1) using contrastive learning to prioritize emotional features over dataset-specific linguistic style and 2) refining emotion predictions with pseudo-emotion intensity score. Our experiment results show that reducing reliance on the linguistic style found in TV transcripts could enhance model’s robustness and accuracy in diverse conversational contexts.

Text-to-SQL translates user queries into SQL statements that can retrieve relevant answers from relational databases. Recent approaches to Text-to-SQL rely on pre-trained language models that are computationally expensive and technically challenging to deploy in real-world applications that require real-time or on-device processing capabilities. In this paper, we perform a focused study on the feasibility of applying recent model compression techniques to sketch-based and sequence-to-sequence Text-to-SQL models. Our results reveal that sketch-based Text-to-SQL models generally have higher inference efficiency and respond better to model compression than sequence-to-sequence models, making them ideal for real-world deployments, especially in use cases with simple SQL statements.

The success of ChatGPT has ignited an AI race, with researchers striving to develop new large language models (LLMs) that can match or surpass the language understanding and generation abilities of commercial ones. In recent times, a number of models have emerged, claiming performance near that of GPT-3.5 or GPT-4 through various instruction-tuning methods. As practitioners of Text-to-SQL parsing, we are grateful for their valuable contributions to open-source research. However, it is important to approach these claims with a sense of scrutiny and ascertain the actual effectiveness of these models. Therefore, we pit six popular large language models against each other, systematically evaluating their Text-to-SQL parsing capability on nine benchmark datasets with five different prompting strategies, covering both zero-shot and few-shot scenarios. Regrettably, the open-sourced models fell significantly short of the performance achieved by closed-source models like GPT-3.5, highlighting the need for further work to bridge the performance gap between these models.