Mental health assessment is crucial for early intervention and effective treatment, yet traditional clinician-based approaches are limited by the shortage of qualified professionals. Recent advances in artificial intelligence have sparked growing interest in automated psychological assessment, yet most existing approaches are constrained by their reliance on static text analysis, limiting their ability to capture deeper and more informative insights that emerge through dynamic interaction and iterative questioning. Therefore, in this paper, we propose a multi-agent framework for mental health evaluation that simulates clinical doctor-patient dialogues, with specialized agents assigned to questioning, adequacy evaluation, scoring, and updating. In detail, we introduce an adaptive questioning mechanism in which an evaluation agent assesses the adequacy of user responses to determine the necessity of generating targeted follow-up queries to address ambiguity and missing information. Additionally, we employ a tree-structured memory in which the root node encodes the user's basic information, while child nodes (e.g., topic and statement) organize key information according to distinct symptom categories and interaction turns. This memory is dynamically updated throughout the interaction to reduce redundant questioning and enhance the information extraction and contextual tracking capabilities. Experimental results on the DAIC-WOZ dataset illustrate the effectiveness of our proposed method, which achieves better performance than existing approaches. Our code is released at \url{https://github.com/MindIntLab-HFUT/AgentMental}.

Gang Hu, Yating Chen, Haiyan Ding, Wang Gao, Huang Jiajia, Min Peng, Qianqian Xie, Kun Yue. Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). 2026.

Xueqing Peng, Lingfei Qian, Yan Wang, Ruoyu Xiang, Yueru He, Yang Ren, Mingyang Jiang, Vincent Jim Zhang, Yuqing Guo, Jeff Zhao, Huan He, Yi Han, Yun Feng, Yuechen Jiang, Yupeng Cao, Haohang Li, Yangyang Yu, Xiaoyu Wang, Penglei Gao, Shengyuan Lin, Keyi Wang, Shanshan Yang, Yilun Zhao, Zhiwei Liu, Peng Lu, Jerry Huang, Suyuchen Wang, Triantafillos Papadopoulos, Polydoros Giannouris, Efstathia Soufleri, Nuo Chen, Zhiyang Deng, Heming Fu, Yijia Zhao, Mingquan Lin, Meikang Qiu, Kaleb E Smith, Arman Cohan, Xiao-Yang Liu, Jimin Huang, Guojun Xiong, Alejandro Lopez-Lira, Xi Chen, Junichi Tsujii, Jian-Yun Nie, Sophia Ananiadou, Qianqian Xie. Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). 2026.

Standardized patients (SPs) are indispensable for clinical skills training but remain expensive and difficult to scale.Although large language model (LLM)-based virtual standardized patients (VSPs) have been proposed as an alternative, their behavior remains unstable and lacks rigorous comparison with human standardized patients.We propose EasyMED, a multi-agent VSP framework that separates case-grounded information disclosure from response generation to support stable, inquiry-conditioned patient behavior.We also introduce SPBench, a human-grounded benchmark with eight expertdefined criteria for interaction-level evaluation.Experiments show that EasyMED more closely matches human SP behavior than existing VSPs, particularly in case consistency and controlled disclosure.A four-week controlled study further demonstrates learning outcomes comparable to human SP training, with stronger early gains for novice learners and improved flexibility, psychological safety, and cost efficiency.

Knowledge Graph Question Answering (KGQA) aims to answer natural language questions by reasoning over structured knowledge graphs (KGs). While large language models (LLMs) have advanced KGQA through their strong reasoning capabilities, existing methods continue to struggle to fully exploit both the rich knowledge encoded in KGs and the reasoning capabilities of LLMs, particularly in complex scenarios. They often assume complete KG coverage and lack mechanisms to judge when external information is needed, and their reasoning remains locally myopic, failing to maintain coherent multi-step planning, leading to reasoning failures even when relevant knowledge exists. We propose Graph-RFT, a novel two-stage reinforcement fine-tuning KGQA framework with a ''plan–KGsearch–and–Websearch–during–think'' paradigm, that enables LLMs to perform autonomous planning and adaptive retrieval scheduling across KG and web sources under incomplete knowledge conditions. Graph-RFT introduces a chain-of-thought (CoT) fine-tuning method with a customized plan–retrieval dataset activates structured reasoning and resolves the GRPO cold-start problem. It then introduces a novel plan–retrieval guided reinforcement learning process integrates explicit planning and retrieval actions with a multi-reward design, enabling coverage-aware retrieval scheduling. It employs a Cartesian-inspired planning module to decompose complex questions into ordered sub-questions, and logical expression to guide tool invocation for globally consistent multi-step reasoning. This reasoning–retrieval process is optimized with a multi-reward combining outcome and retrieval-specific signals, enabling the model to learn when and how to combine KG and web retrieval effectively. Experiments on multiple KGQA benchmarks demonstrate that Graph-RFT achieves superior performance over strong baselines, even with smaller LLM backbones, and substantially improves complex question decomposition, factual coverage, and tool coordination.

Haohang Li, Yupeng Cao, Yangyang Yu, Shashidhar Reddy Javaji, Zhiyang Deng, Yueru He, Yuechen Jiang, Zining Zhu, K.p. Subbalakshmi, Jimin Huang, Lingfei Qian, Xueqing Peng, Jordan W. Suchow, Qianqian Xie. Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). 2025.

Misinformation is prevalent in various fields such as education, politics, health, etc., causing significant harm to society. However, current methods for cross-domain misinformation detection rely on effort- and resource-intensive fine-tuning and complex model structures. With the outstanding performance of LLMs, many studies have employed them for misinformation detection. Unfortunately, they focus on in-domain tasks and do not incorporate significant sentiment and emotion features (which we jointly call affect). In this paper, we propose RAEmoLLM, the first retrieval augmented (RAG) LLMs framework to address cross-domain misinformation detection using in-context learning based on affective information. RAEmoLLM includes three modules. (1) In the index construction module, we apply an emotional LLM to obtain affective embeddings from all domains to construct a retrieval database. (2) The retrieval module uses the database to recommend top K examples (text-label pairs) from source domain data for target domain contents. (3) These examples are adopted as few-shot demonstrations for the inference module to process the target domain content. The RAEmoLLM can effectively enhance the general performance of LLMs in cross-domain misinformation detection tasks through affect-based retrieval, without fine-tuning. We evaluate our framework on three misinformation benchmarks. Results show that RAEmoLLM achieves significant improvements compared to the other few-shot methods on three datasets, with the highest increases of 15.64%, 31.18%, and 15.73% respectively. This project is available at https://github.com/lzw108/RAEmoLLM.

General and legal domain LLMs have demonstrated strong performance in various tasks of LegalAI. However, their current evaluations lack alignment with the fundamental logic of legal reasoning, the legal syllogism. This hinders trust and understanding from legal experts. To bridge this gap, we introduce LAiW, the Chinese legal LLM benchmark structured around the legal syllogism. We evaluate legal LLMs across three levels of capability, each reflecting a progressively more complex stage of legal syllogism: fundamental information retrieval, legal principles inference, and advanced legal applications, and encompassing a wide range of tasks in different legal scenarios. Our automatic evaluation reveals that LLMs, despite their ability to answer complex legal questions, lack the inherent logical processes of the legal syllogism. This limitation poses a barrier to acceptance by legal professionals. Furthermore, manual evaluation with legal experts confirms this issue and highlights the importance of pre-training on legal text to enhance the legal syllogism of LLMs. Future research may prioritize addressing this gap to unlock the full potential of LLMs in legal applications.

We propose ELAINE (EngLish-jApanese-chINesE)-medLLM, a trilingual (English, Japanese, Chinese) large language model adapted for the bio-medical domain based on Llama-3-8B. The training dataset was carefully curated in terms of volume and diversity to adapt to the biomedical domain and endow trilingual capability while preserving the knowledge and abilities of the base model. The training follows 2-stage paths: continued pre-training and supervised fine-tuning (SFT). Our results demonstrate that ELAINE-medLLM exhibits superior trilingual capabilities compared to existing bilingual or multilingual medical LLMs without severely sacrificing the base model’s capability.

Xueqing Peng, Triantafillos Papadopoulos, Efstathia Soufleri, Polydoros Giannouris, Ruoyu Xiang, Yan Wang, Lingfei Qian, Jimin Huang, Qianqian Xie, Sophia Ananiadou. Proceedings of the 2025 Conference on Empirical Methods in Natural Language Processing. 2025.

Recent advancements in LLM alignment leverage token-level supervisions to perform finegrained preference optimization.However, existing token-level alignment methods either optimize on all available tokens, which can be noisy and inefficient, or perform selective training with complex and expensive key token selection strategies.In this work, we propose Selective Preference Optimization (SePO), a novel selective alignment strategy that centers on efficient key token selection without requiring strong, fine-grained supervision signals.We prove the feasibility of Direct Preference Optimization (DPO) as token-level reward function estimators, which applies to any existing alignment datasets and enables costefficient token selection with small-scale model sizes and training data.We then train an oracle model with DPO on the target data and utilize the estimated reward function to score all tokens within the target dataset, where only the key tokens are selected to supervise the target policy model with a contrastive objective function.Extensive experiments on three public evaluation benchmarks show that SePO significantly outperforms competitive baseline methods by only optimizing on 30% key tokens with up to 60% reduction in GPU training hours.We also explore SePO as a new paradigm for weakto-strong generalization, showing that weak oracle models effectively supervise strong policy models with up to 16.8× more parameters.SePO also selects useful supervision signals from out-of-distribution data, alleviating the over-optimization problem.The project is open-sourced here.

With the advent of large language models (LLMs), managing scientific literature via LLMs has become a promising direction of research. However, existing approaches often overlook the rich structural and semantic relevance among scientific literature, limiting their ability to discern the relationships between pieces of scientific knowledge, and suffer from various types of hallucinations. These methods also focus narrowly on individual downstream tasks, limiting their applicability across use cases. We propose LitFM, the first literature foundation model designed for a wide variety of practical downstream tasks on domain-specific literature, with a focus on citation information. At its core, LitFM contains a novel graph retriever that can provide accurate and diverse recommendations for LLM to integrate graph structure information and relevant literature. LitFM also leverages a knowledge-infused LLM, fine-tuned through a well-developed instruction paradigm. It enables LitFM to extract domain-specific knowledge from literature and reason relationships among them. By integrating citation graphs during both training and inference, LitFM can generalize to unseen papers and accurately assess their relevance within existing literature. Additionally, we introduce new large-scale literature citation benchmark datasets on three academic fields, featuring sentence-level citation information and local context. Extensive experiments validate the superiority of LitFM, achieving 28.1% improvement on retrieval task in precision, and an average improvement of 7.52% over state-of-the-art across six downstream literature-related tasks.

Graph foundation models (GFMs) have emerged as a promising paradigm for learning transferable knowledge across diverse graph-structured data. The inherent heterogeneity in features and graph structures poses significant challenges for building scalable and generalizable GFMs. Existing research has employed mixture-of-experts (MoE) models to handle the challenges, assigning the most suitable expert to each graph. Despite this, the underlying mechanisms of MoE within the context of GFMs remain insufficiently explored. In this work, we conduct an in-depth experimental study on an MoE-based GFM and uncover an intriguing finding: the experts ranked second and third assigned by the router perform better than the top-ranked expert. This insight motivates us to investigate the potential of leveraging knowledge embedded across multiple experts. However, directly ensembling the outputs of multiple experts would incur substantial computational overhead, while applying a standard expert merging strategy risks suboptimal performance. To address these challenges, we introduce two enhanced expert merging strategies that retain the computational efficiency of expert merging, while improving performance to approach the effectiveness of expert ensembling. Specifically, we propose (i) a knowledge distillation-inspired expert merging method that aligns the behavior of parameter-fused experts with expert ensembles, and (ii) a theoretical parameter proximity approach that leverages the similarity of expert parameters to approximate ensemble outputs while preserving diversity. Extensive experiments demonstrate that our methods effectively enhance model performance.

The application of AI in psychiatric diagnosis faces significant challenges, including the subjective nature of mental health assessments, symptom overlap across disorders, and privacy constraints limiting data availability. To address these issues, we present MoodAngels, the first specialized multi-agent framework for mood disorder diagnosis. Our approach combines granular-scale analysis of clinical assessments with a structured verification process, enabling more accurate interpretation of complex psychiatric data. Complementing this framework, we introduce MoodSyn, an open-source dataset of 1,173 synthetic psychiatric cases that preserves clinical validity while ensuring patient privacy. Experimental results demonstrate that MoodAngels outperforms conventional methods, with our baseline agent achieving 12.3\% higher accuracy than GPT-4o on real-world cases, and our full multi-agent system delivering further improvements. Together, these contributions provide both an advanced diagnostic tool and a critical research resource for computational psychiatry, bridging important gaps in AI-assisted mental health assessment.

OBJECTIVE: Common Data Elements (CDEs) standardize data collection and sharing across studies, enhancing data interoperability and improving research reproducibility. However, implementing CDEs presents challenges due to the broad range and variety of data elements. This study aims to develop a CDE mapping tool to bridge the gap between local data elements and National Institutes of Health (NIH) CDEs. METHODS: We propose CDEMapper, a large language model (LLM)-powered mapping tool designed to assist in mapping local data elements to NIH CDEs. CDEMapper has 3 core modules: (1) CDE indexing and embeddings. NIH CDEs were indexed and embedded to support semantic search; (2) CDE recommendations. The tool combines Elasticsearch (BM25 methods) with GPT services to recommend candidate CDEs and their permissible values; and (3) Human review. Users review and select the best match for their data elements and value sets. We evaluate the tool's recommendation accuracy and usability against manual annotations and testing. RESULTS: CDEMapper offers a publicly available, LLM-powered, and intuitive user interface that consolidates essential and advanced mapping services into a streamlined pipeline. The evaluation results demonstrated that the augmented BM25 with GPT embeddings and a GPT ranker achieved the overall best performance. The usability test also highlighted the effectiveness and efficiency of our tool. DISCUSSIONS AND CONCLUSIONS: This work opens up the potential of using LLMs to assist with CDE mapping when aligning local data elements with NIH CDEs. Additionally, this effort helps researchers better understand the gaps between their data elements and NIH CDEs while promoting CDE reusability.

Mengxi Xiao, Qianqian Xie, Ziyan Kuang, Zhicheng Liu, Kailai Yang, Min Peng, Weiguang Han, Jimin Huang. Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). 2024.

In real-world applications, users express different behaviors when they interact with different items, including implicit click/like interactions, and explicit comments/reviews interactions. Nevertheless, almost all recommender works are focused on how to describe user preferences by the implicit click/like interactions, to find the synergy of people. For the content-based explicit comments/reviews interactions, some works attempt to utilize them to mine the semantic knowledge to enhance recommender models. However, they still neglect the following two points: (1) The content semantic is a universal world knowledge; how do we extract the multi-aspect semantic information to empower different domains? (2) The user/item ID feature is a fundamental element for recommender models; how do we align the ID and content semantic feature space? In this paper, we propose a 'plugin' semantic knowledge transferring method LoID, which includes two major components: (1) LoRA-based large language model pretraining to extract multi-aspect semantic information; (2) ID-based contrastive objective to align their feature spaces. We conduct extensive experiments with SOTA baselines to demonstrate superiority of our method LoID.

Maintaining factual consistency is a critical issue in abstractive text summarisation, however, it cannot be assessed by traditional automatic metrics used for evaluating text summarisation, such as ROUGE scoring. Recent efforts have been devoted to developing improved metrics for measuring factual consistency using pre-trained language models, but these metrics have restrictive token limits, and are therefore not suitable for evaluating long document text summarisation. Moreover, there is limited research and resources available for evaluating whether existing automatic evaluation metrics are fit for purpose when applied in long document settings. In this work, we evaluate the efficacy of automatic metrics for assessing the factual consistency of long document text summarisation. We create a human-annotated data set for evaluating automatic factuality metrics, LongSciVerify, which contains fine-grained factual consistency annotations for long document summaries from the scientific domain. We also propose a new evaluation framework, LongDocFACTScore, which is suitable for evaluating long document summarisation. This framework allows metrics to be efficiently extended to any length document and outperforms existing state-of-the-art metrics in its ability to correlate with human measures of factuality when used to evaluate long document summarisation data sets. We make our code and LongSciVerify data set publicly available: https://github.com/jbshp/LongDocFACTScore.

Sentiment analysis and emotion detection are important research topics in natural language processing (NLP) and benefit many downstream tasks. With the widespread application of large language models (LLMs), researchers have started exploring the application of LLMs based on instruction-tuning in the field of sentiment analysis. However, these models only focus on single aspects of affective classification tasks (e.g. sentimental polarity or categorical emotions), and overlook the regression tasks (e.g. sentiment strength or emotion intensity), which leads to poor performance in downstream tasks. The main reason is the lack of comprehensive affective instruction tuning datasets and evaluation benchmarks, which cover various affective classification and regression tasks. Moreover, although emotional information is useful for downstream tasks, existing downstream datasets lack high-quality and comprehensive affective annotations. In this paper, we propose EmoLLMs, the first series of open-sourced instruction-following LLMs for comprehensive affective analysis based on fine-tuning various LLMs with instruction data, the first multi-task affective analysis instruction dataset (AAID) with 234K data samples based on 3 classification tasks and 2 regression tasks to support LLM instruction tuning, and a comprehensive affective evaluation benchmark (AEB) with 8 regression tasks and 6 classification tasks from various sources and domains to test the generalization ability of LLMs. We propose a series of EmoLLMs by fine-tuning LLMs with AAID to solve various affective instruction tasks. We compare our models with a variety of LLMs and sentiment analysis tools on AEB, where our models outperform all other open-sourced LLMs and sentiment analysis tools, and surpass ChatGPT and GPT-4 in most tasks, which shows that the series of EmoLLMs achieve the ChatGPT-level and GPT-4-level generalization capabilities on affective analysis tasks, and demonstrates our models can be used as affective annotation tools. This project is available at https://github.com/lzw108/EmoLLMs/.

Despite Spanish's pivotal role in the global finance industry, a pronounced gap exists in Spanish financial natural language processing (NLP) and application studies compared to English, especially in the era of large language models (LLMs).To bridge this gap, we unveil Toisón de Oro, the first bilingual framework that establishes instruction datasets, finetuned LLMs, and evaluation benchmark for financial LLMs in Spanish joint with English.We construct a rigorously curated bilingual instruction dataset including over 144K Spanish and English samples from 15 datasets covering 7 tasks.Harnessing this, we introduce FinMA-ES, an LLM designed for bilingual financial applications.We evaluate our model and existing LLMs using FLARE-ES, the first comprehensive bilingual evaluation benchmark with 21 datasets covering 9 tasks.The FLARE-ES benchmark results

Data serves as the fundamental basis for advancing deep learning. The tabular data presented in a structured format is highly valuable for modeling and training.However, even in the era of LLM, obtaining tabular data from sensitive domains remains a challenge due to privacy or copyright concerns. Therefore, exploring the methods for effectively using models like LLMs to generate synthetic tabular data, which is privacy-preserving but similar to original one, is urgent.In this paper, we introduce a new framework HARMONIC for tabular data generation and evaluation by LLMs. In the data generation part of our framework, we employ fine-tuning to generate tabular data and enhance privacy rather than continued pre-training which is often used by previous small-scale LLM-based methods. In particular, we construct an instruction fine-tuning dataset based on the idea of the k-nearest neighbors algorithm to inspire LLMs to discover inter-row relationships. By such fine-tuning, LLMs are trained to remember the format and connections of the data rather than the data itself, which reduces the risk of privacy leakage. The experiments find that our tabular data generation achieves equivalent performance as existing methods but with better privacy by the metric of MLE, DCR, etc.In the evaluation part of our framework, we develop a specific privacy risk metric DLT for LLM synthetic data generation, which quantifies the extent to which the generator itself leaks data. We also developed LLE, a performance evaluation metric for downstream LLM tasks, which is more practical and credible than previous metrics.The experiments show that our data generation method outperform the previous methods in the metrics DLT and LLE.

LLMs have transformed NLP and shown promise in various fields, yet their potential in finance is underexplored due to a lack of comprehensive benchmarks, the rapid development of LLMs, and the complexity of financial tasks. In this paper, we introduce FinBen, the first extensive open-source evaluation benchmark, including 42 datasets spanning 24 financial tasks, covering eight critical aspects: information extraction (IE), textual analysis, question answering (QA), text generation, risk management, forecasting, decision-making, and bilingual (English and Spanish). FinBen offers several key innovations: a broader range of tasks and datasets, the first evaluation of stock trading, novel agent and Retrieval-Augmented Generation (RAG) evaluation, and two novel datasets for regulations and stock trading. Our evaluation of 21 representative LLMs, including GPT-4, ChatGPT, and the latest Gemini, reveals several key findings: While LLMs excel in IE and textual analysis, they struggle with advanced reasoning and complex tasks like text generation and forecasting. GPT-4 excels in IE and stock trading, while Gemini is better at text generation and forecasting. Instruction-tuned LLMs improve textual analysis but offer limited benefits for complex tasks such as QA. FinBen has been used to host the first financial LLMs shared task at the FinNLP-AgentScen workshop during IJCAI-2024, attracting 12 teams. Their novel solutions outperformed GPT-4, showcasing FinBen's potential to drive innovations in financial LLMs. All datasets and code are publicly available for the research community, with results shared and updated regularly on the Open Financial LLM Leaderboard.

Recent advancements in large language models (LLMs) focus on aligning to heterogeneous human expectations and values via multi-objective preference alignment. However, existing methods are dependent on the policy model parameters, which require high-cost repetition of their alignment algorithms for each new policy model, and they cannot expand to unseen objectives due to their static alignment objectives. In this work, we propose Meta-Objective Aligner (MetaAligner), the first policy-agnostic and generalizable method for multi-objective preference alignment. MetaAligner models multi-objective alignment into three stages: (1) dynamic objectives reformulation algorithm reorganizes traditional alignment datasets to supervise the model on performing flexible alignment across different objectives; (2) conditional weak-to-strong correction paradigm aligns the weak outputs of fixed policy models to approach strong outputs with higher preferences in the corresponding alignment objectives, enabling plug-and-play inferences on any policy models, which significantly reduces training costs and facilitates alignment on close-source policy models; (3) generalizable inference method flexibly adjusts target objectives by updating their text descriptions in the prompts, facilitating generalizable alignment to unseen objectives. Experimental results show that MetaAligner achieves significant and balanced improvements in multi-objective alignments on 10 state-of-the-art policy models, and saves up to 93.63% of GPU training hours compared to previous alignment methods. The model also effectively aligns unseen objectives, marking the first step towards generalizable multi-objective preference alignment.

Large language models (LLMs) have demonstrated notable potential in conducting complex tasks and are increasingly utilized in various financial applications. However, high-quality sequential financial investment decision-making remains challenging. These tasks require multiple interactions with a volatile environment for every decision, demanding sufficient intelligence to maximize returns and manage risks. Although LLMs have been used to develop agent systems that surpass human teams and yield impressive investment returns, opportunities to enhance multi-source information synthesis and optimize decision-making outcomes through timely experience refinement remain unexplored. Here, we introduce FinCon, an LLM-based multi-agent framework tailored for diverse financial tasks. Inspired by effective real-world investment firm organizational structures, FinCon utilizes a manager-analyst communication hierarchy. This structure allows for synchronized cross-functional agent collaboration towards unified goals through natural language interactions and equips each agent with greater memory capacity than humans. Additionally, a risk-control component in FinCon enhances decision quality by episodically initiating a self-critiquing mechanism to update systematic investment beliefs. The conceptualized beliefs serve as verbal reinforcement for the future agent’s behavior and can be selectively propagated to the appropriate node that requires knowledge updates. This feature significantly improves performance while reducing unnecessary peer-to-peer communication costs. Moreover, FinCon demonstrates strong generalization capabilities in various financial tasks, including stock trading and portfolio management.

Early suicidal ideation detection using social media is crucial for mental health surveillance. Simultaneously, emojis from the posts can help us better understand users' emotions and predict mental health conditions. However, research in emoji-based suicide analysis remains underexplored, with few resources available, which can restrict the development of studying emoji usage patterns among users with suicidal ideation. In this work, we build a derived suicide-related emoji dataset named SuicidEmoji, which contains 25k emoji posts (2,329 suicide-related posts and 22,722 posts for the control group users) filtered from about 1.3 million crawled Reddit data. To the best of our knowledge, SuicidEmoji is the first suicide-related emoji dataset. Based on SuicidEmoji, we propose two novel tasks: emoji-aware suicidal ideation detection and emoji prediction, for which we build two benchmark subdatasets from SuicidEmoji to evaluate the performance of advanced methods including pre-trained language models (PLMs) and large language models (LLMs). We analyze the experimental results of two PLMs and the highly capable LLMs, which reveal the significance and challenges of emoji-based suicide-related NLP tasks. The dataset is avaliable at https://github.com/TianlinZhang668/SuicidEmoji.

As an integral part of people's daily lives, social media is becoming a rich source for automatic mental health analysis. As traditional discriminative methods bear poor generalization ability and low interpretability, the recent large language models (LLMs) have been explored for interpretable mental health analysis on social media, which aims to provide detailed explanations along with predictions in zero-shot or few-shot settings. The results show that LLMs still achieve unsatisfactory classification performance in a zero-shot/few-shot manner, which further significantly affects the quality of the generated explanations. Domain-specific finetuning is an effective solution, but faces two critical challenges: 1) lack of high-quality training data. 2) no open-source foundation LLMs. To alleviate these problems, we formally model interpretable mental health analysis as a text generation task, and build the first multi-task and multi-source interpretable mental health instruction (IMHI) dataset with 105K data samples to support LLM instruction tuning and evaluation. The raw social media data are collected from 10 existing sources covering 8 mental health analysis tasks. We prompt ChatGPT with expert-designed few-shot prompts to obtain explanations. To ensure the reliability of the explanations, we perform strict automatic and human evaluations on the correctness, consistency, and quality of generated data. Based on the IMHI dataset and LLaMA2 foundation models, we train MentaLLaMA, the first open-source instruction-following LLM series for interpretable mental health analysis on social media. We evaluate MentaLLaMA and other advanced methods on the IMHI benchmark, the first holistic evaluation benchmark for interpretable mental health analysis. The results show that MentaLLaMA approaches state-of-the-art discriminative methods in correctness and generates human-level explanations. MentaLLaMA models also show strong generalizability to unseen tasks. The project is available at https://github.com/SteveKGYang/MentaLLaMA.

Temporal reasoning is a crucial natural language processing (NLP) task, providing a nuanced understanding of time-sensitive contexts within textual data. Although recent advancements in Large Language Models (LLMs) have demonstrated their potential in temporal reasoning, the predominant focus has been on tasks such as temporal expression detection, normalization, and temporal relation extraction. These tasks are primarily designed for the extraction of direct and past temporal cues from given contexts and to engage in simple reasoning processes. A significant gap remains when considering complex reasoning tasks such as event forecasting, which requires multi-step temporal reasoning on events and prediction on the future timestamp. Another notable limitation of existing methods is their incapability to illustrate their reasoning process for explaining their prediction, hindering explainability. In this paper, we introduce the first task of explainable temporal reasoning, to predict an event's occurrence at a future timestamp based on context which requires multiple reasoning over multiple events, and subsequently provide a clear explanation for their prediction. Our task offers a comprehensive evaluation of both the LLMs' complex temporal reasoning ability, the future event prediction ability, and explainability-a critical attribute for AI applications. To support this task, we present the first instruction-tuning dataset of explainable temporal reasoning (ExpTime) with 26k derived from the temporal knowledge graph datasets, using a novel knowledge-graph-instructed-generation strategy. Based on the dataset, we propose the first open-source LLM series TimeLlaMA based on the foundation LLM LlaMA2, with the ability of instruction following for explainable temporal reasoning. We compare the performance of our method and a variety of LLMs, where our method achieves the state-of-the-art performance of temporal prediction and explanation generation. We also explore the impact of instruction tuning and different training sizes of instruction-tuning data, highlighting LLM's capabilities and limitations in complex temporal prediction and explanation generation.

MOTIVATION: Large Language Models (LLMs) have the potential to revolutionize the field of Natural Language Processing, excelling not only in text generation and reasoning tasks but also in their ability for zero/few-shot learning, swiftly adapting to new tasks with minimal fine-tuning. LLMs have also demonstrated great promise in biomedical and healthcare applications. However, when it comes to Named Entity Recognition (NER), particularly within the biomedical domain, LLMs fall short of the effectiveness exhibited by fine-tuned domain-specific models. One key reason is that NER is typically conceptualized as a sequence labeling task, whereas LLMs are optimized for text generation and reasoning tasks. RESULTS: We developed an instruction-based learning paradigm that transforms biomedical NER from a sequence labeling task into a generation task. This paradigm is end-to-end and streamlines the training and evaluation process by automatically repurposing pre-existing biomedical NER datasets. We further developed BioNER-LLaMA using the proposed paradigm with LLaMA-7B as the foundational LLM. We conducted extensive testing on BioNER-LLaMA across three widely recognized biomedical NER datasets, consisting of entities related to diseases, chemicals, and genes. The results revealed that BioNER-LLaMA consistently achieved higher F1-scores ranging from 5% to 30% compared to the few-shot learning capabilities of GPT-4 on datasets with different biomedical entities. We show that a general-domain LLM can match the performance of rigorously fine-tuned PubMedBERT models and PMC-LLaMA, biomedical-specific language model. Our findings underscore the potential of our proposed paradigm in developing general-domain LLMs that can rival SOTA performances in multi-task, multi-domain scenarios in biomedical and health applications. AVAILABILITY AND IMPLEMENTATION: Datasets and other resources are available at https://github.com/BIDS-Xu-Lab/BioNER-LLaMA.

Contrastive learning has recently been introduced into neural topic models (NTMs) to improve latent semantic discovery, but existing methods suffer from the sample bias problem owing to word frequency-based sampling strategy, which may result in false negative samples with similar semantics to the prototypes. We propose the novel graph contrastive neural topic model (GCTM), based on the graph-based sampling strategy, guided by the in-depth correlation and irrelevance information among documents and words. We model the input document as the document word bipartite graph (DWBG) and construct positive and negative word co-occurrence graphs (WCGs), to capture in-depth semantic correlation and irrelevance among words. Based on the DWBG and WCGs, we design the document-word information propagation (DWIP) process to perform the edge perturbation of DWBG, based on multi-hop correlations/irrelevance among documents and words. This yields the desired negative and positive samples, which are utilized for GCL together with the prototypes to improve learning document topic representations and latent topics. Experiments on several benchmark datasets demonstrate the effectiveness of our method for topic coherence and document representation learning compared with existing state-of-the-art methods.

Factual inconsistency with source documents in automatically generated summaries can lead to misinformation or pose risks. Existing factual consistency (FC) metrics are constrained by their performance, efficiency, and explainability. Recent advances in Large language models (LLMs) have demonstrated remarkable potential in text evaluation but their effectiveness in assessing FC in summarization remains underexplored. Prior research has mostly focused on proprietary LLMs, leaving essential factors that affect their assessment capabilities unexplored. Additionally, current FC evaluation benchmarks are restricted to news articles, casting doubt on the generality of the FC methods tested on them. In this paper, we first address the gap by introducing TreatFact—a dataset of LLM-generated summaries of clinical texts, annotated for FC by domain experts. Moreover, we benchmark 11 LLMs for FC evaluation across news and clinical domains and analyse the impact of model size, prompts, pre-training and fine-tuning data. Our findings reveal that despite proprietary models prevailing on the task, open-source LLMs lag behind. Nevertheless, there is potential for enhancing the performance of open-source LLMs through increasing model size, expanding pre-training data, and developing well-curated fine-tuning data. Experiments on TreatFact suggest that both previous methods and LLM-based evaluators are unable to capture factual inconsistencies in clinical summaries, posing a new challenge for FC evaluation.

Temporal relation extraction aims to infer the temporal order of either two events in the document. Because of the nature of events in real life, severe imbalanced temporal relation classes exist in the temporal relation extraction task. Even though various methods have been proposed to improve the overall performance, the accuracy of these methods on minority temporal classes is limited. In this work, we present a contrastive prototypical learning architecture to address this problem, which explicitly models the spatial similarity between instances in the embedding space so that instances from minority classes can be distinguished from the large classes. To make it compatible with current temporal relation extraction settings, we propose a novel sampling memory queue-based method so that the architecture can be applied to a limited batch size scenario. We further design a context encoding layer that incorporates both contextualized information and linguistic features such as tense information and dependency. Our extensive experiments on TimeBank-Dense, TDDiscourse, and MATRES datasets demonstrate that our model can significantly improve the performance of minority relation classes and, therefore increase the overall learning ability.

Given the overwhelming and rapidly increasing volumes of the published biomedical literature, automatic biomedical text summarization has long been a highly important task. Recently, great advances in the performance of biomedical text summarization have been facilitated by pre-trained language models (PLMs) based on fine-tuning. However, existing summarization methods based on PLMs do not capture domain-specific knowledge. This can result in generated summaries with low coherence, including redundant sentences, or excluding important domain knowledge conveyed in the full-text document. Furthermore, the black-box nature of the transformers means that they lack explainability, i.e. it is not clear to users how and why the summary was generated. The domain-specific knowledge and explainability are crucial for the accuracy and transparency of biomedical text summarization methods. In this article, we aim to address these issues by proposing a novel domain knowledge-enhanced graph topic transformer (DORIS) for explainable biomedical text summarization. The model integrates the graph neural topic model and the domain-specific knowledge from the Unified Medical Language System (UMLS) into the transformer-based PLM, to improve the explainability and accuracy. Experimental results on four biomedical literature datasets show that our model outperforms existing state-of-the-art (SOTA) PLM-based summarization methods on biomedical extractive summarization. Furthermore, our use of graph neural topic modeling means that our model possesses the desirable property of being explainable, i.e. it is straightforward for users to understand how and why the model selects particular sentences for inclusion in the summary. The domain-specific knowledge helps our model to learn more coherent topics, to better explain the performance.

Language is the principal tool for human communication, in which humor is one of the most attractive parts.Producing natural language like humans using computers, a.k.a,Natural Language Generation (NLG), has been widely used for dialogue systems, chatbots, text summarization, as well as AI-Generated Content (AIGC), e.g., idea generation, and scriptwriting.However, the humor aspect of natural language is relatively under-investigated, especially in the age of pre-trained language models.In this work, we aim to preliminarily test whether NLG can generate humor as humans do.We build the largest dataset consisting of numerous Chinese Comical Crosstalk scripts (called C 3 in short), which is for a popular Chinese performing art called 'Xiangsheng' or '相 声' since 1800s 1 .We benchmark various generation approaches including training-from-scratch Seq2seq, fine-tuned middle-scale PLMs, and large-scale PLMs with and without fine-tuning.Moreover, we also conduct a human assessment, showing that 1) large-scale pretraining largely improves crosstalk generation quality; and 2) even the scripts generated from the best PLM is far from what we expect.We conclude humor generation could be largely improved using large-scale PLMs, but it is still in its infancy.The data and benchmarking code are publicly available in https://github.com/ anonNo2/crosstalk-generation.* Benyou is the corresponding author. 1 For convenience for non-Chinese speakers, we called 'crosstalk' for 'Xiangsheng' in this paper. Roles Script (in Chinese)

The latest large language models (LLMs) such as ChatGPT, exhibit strong capabilities in automated mental health analysis.However, existing relevant studies bear several limitations, including inadequate evaluations, lack of prompting strategies, and ignorance of exploring LLMs for explainability.To bridge these gaps, we comprehensively evaluate the mental health analysis and emotional reasoning ability of LLMs on 11 datasets across 5 tasks.We explore the effects of different prompting strategies with unsupervised and distantly supervised emotional information.Based on these prompts, we explore LLMs for interpretable mental health analysis by instructing them to generate explanations for each of their decisions.We convey strict human evaluations to assess the quality of the generated explanations, leading to a novel dataset with 163 humanassessed explanations 1 .We benchmark existing automatic evaluation metrics on this dataset to guide future related works.According to the results, ChatGPT shows strong in-context learning ability but still has a significant gap with advanced task-specific methods.Careful prompt engineering with emotional cues and expertwritten few-shot examples can also effectively improve performance on mental health analysis.In addition, ChatGPT generates explanations that approach human performance, showing its great potential in explainable mental health analysis.

Although large language models (LLMs) have shown great performance in natural language processing (NLP) in the financial domain, there are no publicly available financially tailored LLMs, instruction tuning datasets, and evaluation benchmarks, which is critical for continually pushing forward the open-source development of financial artificial intelligence (AI). This paper introduces PIXIU, a comprehensive framework including the first financial LLM based on fine-tuning LLaMA with instruction data, the first instruction data with 128K data samples to support the fine-tuning, and an evaluation benchmark with 8 tasks and 15 datasets. We first construct the large-scale multi-task instruction data considering a variety of financial tasks, financial document types, and financial data modalities. We then propose a financial LLM called FinMA by fine-tuning LLaMA with the constructed dataset to be able to follow instructions for various financial tasks. To support the evaluation of financial LLMs, we propose a standardized benchmark that covers a set of critical financial tasks, including six financial NLP tasks and two financial prediction tasks. With this benchmark, we conduct a detailed analysis of FinMA and several existing LLMs, uncovering their strengths and weaknesses in handling critical financial tasks. The model, datasets, benchmark, and experimental results are open-sourced to facilitate future research in financial AI.

Citation graphs can be helpful in generating high-quality summaries of scientific papers, where references of a scientific paper and their correlations can provide additional knowledge for contextualising its background and main contributions. Despite the promising contributions of citation graphs, it is still challenging to incorporate them into summarization tasks. This is due to the difficulty of accurately identifying and leveraging relevant content in references for a source paper, as well as capturing their correlations of different intensities. Existing methods either ignore references or utilize only abstracts indiscriminately from them, failing to tackle the challenge mentioned above. To fill that gap, we propose a novel citation-aware scientific paper summarization framework based on the citation graph, able to accurately locate and incorporate the salient contents from references, as well as capture varying relevance between source papers and their references. Specifically, we first build a domain-specific dataset PubMedCite with about 192K biomedical scientific papers and a large citation graph preserving 917K citation relationships between them. It is characterized by preserving the salient contents extracted from full texts of references, and the weighted correlation between the salient contents of references and the source paper. Based on it, we design a self-supervised citation-aware summarization framework (CitationSum) with graph contrastive learning, which boosts the summarization generation by efficiently fusing the salient information in references with source paper contents under the guidance of their correlations. Experimental results show that our model outperforms the state-of-the-art methods, due to efficiently leveraging the information of references and citation correlations.

Recently, neural topic models (NTMs) have been incorporated into pre-trained language models (PLMs), to capture the global semantic information for text summarization. However, in these methods, there remain limitations in the way they capture and integrate the global semantic information. In this paper, we propose a novel model, the graph contrastive topic enhanced language model (GRETEL), that incorporates the graph contrastive topic model with the pre-trained language model, to fully leverage both the global and local contextual semantics for long document extractive summarization. To better capture and incorporate the global semantic information into PLMs, the graph contrastive topic model integrates the hierarchical transformer encoder and the graph contrastive learning to fuse the semantic information from the global document context and the gold summary. To this end, GRETEL encourages the model to efficiently extract salient sentences that are topically related to the gold summary, rather than redundant sentences that cover sub-optimal topics. Experimental results on both general domain and biomedical datasets demonstrate that our proposed method outperforms SOTA methods.

To tackle the understanding of complex questions, Question Decomposition Meaning Representation (QDMR) decomposes a complex question into a sequence of atomic simple questions. However, state-of-the-art QDMR parsers neglect the type information of simple questions and the dependency information between simple questions, leading to limited overall performance. In this paper, we propose a Decomposition Graph Reconstruction (DGR) model to induce the information by introducing two additional tasks: simple question type classification and dependency relationship detection. DGR can generate sub-questions with the expected type and organize sub-questions with correct logical connections. Experimental results on the BREAK QDMR dataset demonstrate the effectiveness of our method compared with previous methods.

Biomedical text summarization is a critical task for comprehension of an ever-growing amount of biomedical literature. Pre-trained language models (PLMs) with transformer-based architectures have been shown to greatly improve performance in biomedical text mining tasks. However, existing methods for text summarization generally fine-tune PLMs on the target corpora directly and do not consider how fine-grained domain knowledge, such as PICO elements used in evidence-based medicine, can help to identify the context needed for generating coherent summaries. To fill the gap, we propose KeBioSum, a novel knowledge infusion training framework, and experiment using a number of PLMs as bases, for the task of extractive summarization on biomedical literature. We investigate generative and discriminative training techniques to fuse domain knowledge (i.e., PICO elements) into knowledge adapters and apply adapter fusion to efficiently inject the knowledge adapters into the basic PLMs for fine-tuning the extractive summarization task. Experimental results from the extractive summarization task on three biomedical literature datasets show that existing PLMs (BERT, RoBERTa, BioBERT, and PubMedBERT) are improved by incorporating the KeBioSum knowledge adapters, and our model outperforms the strong baselines.

Due to the overload of published scientific articles, citation recommendation has long been a critical research problem for automatically recommending the most relevant citations of given articles. Relational topic models (RTMs) have shown promise on citation prediction via joint modeling of document contents and citations. However, existing RTMs can only capture pairwise or direct (first-order) citation relationships among documents. The indirect (high-order) citation links have been explored in graph neural network–based methods, but these methods suffer from the well-known explainability problem. In this article, we propose a model called Graph Neural Collaborative Topic Model that takes advantage of both relational topic models and graph neural networks to capture high-order citation relationships and to have higher explainability due to the latent topic semantic structure. Experiments on three real-world citation datasets show that our model outperforms several competitive baseline methods on citation recommendation. In addition, we show that our approach can learn better topics than the existing approaches. The recommendation results can be well explained by the underlying topics.

Qianqian Xie, Jimin Huang, Pan Du, Min Peng, Jian-Yun Nie. Proceedings of the 2021 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies. 2021.

Graph Neural Networks (GNNs) such as GCN can effectively learn document representations via the semantic relation graph among documents and words. However, despite a few exceptions, most of the previous work in this line of research does not consider the underlying topical semantics inherited in document contents and the relation graph, making the representations less effective and hard to interpret. In a few recent studies trying to incorporate latent topics into GNNs, the topics have been learned independently from the relation graph modeling. Intuitively, topic extraction can benefit much from the information propagation of the relation graph structure - directly and indirectly connected documents and words have similar topics. In this paper, we propose a novel Graph Topic Neural Network (GTNN) model to mine latent topic semantics for interpretable document representation learning, taking into account the document-document, document-word, and word-word relationships in the graph. We also show that our model can be viewed as semi-amortized inference for relational topic model based on Poisson distribution, with high order correlations. We test our model in several settings: unsupervised, semi-supervised, and supervised representation learning, for both connected and unconnected documents. In all the cases, our model outperforms the state-of-the-art models for these tasks.

Summary Software repositories are crucial resources for many software tasks, including code retrieval and annotation. Programming forums provide questions and answers (Q&A) from software developers, containing abundant code‐description posts for exchanging knowledge about programming issues. However, most posts provide personal opinions of users that are often not adequately confirmed or outdated. Mining software repositories in such open and unrestricted forums is challenging. Since the posts can be arbitrary and noisy, it is difficult to get unified labels for supervised noise elimination. Different from existing mining approaches, this paper proposes Code‐Description Mining Framework (CodeMF), an unsupervised framework to eliminate noisy posts and extract high quality software repositories from programming forums. CodeMF treats all social features of the posts as discrete‐time signals for kernel principal component analysis and further performs wavelet transform feature fusion to find the delicate changes (noises in temporal signals). We conduct comprehensive experiments on StackOverflow. Experimental results demonstrate that CodeMF can effectively reduce running time and improve precision via mining high‐quality software repositories for various programming languages, especially for the large‐scale codebases. To further illustrate the effect of CodeMF applied in software tasks, we introduce it to improve the performance of query‐expansion code search. Meanwhile, for SQL and C# programs, compared to the state‐of‐the‐art query‐expansion method QECK, the improvement of QECK CodeMF is 2% and 6% on Recall@10, and 4% and 14% on mean reciprocal rank, respectively.

Deep generative models (DGMs) have shown strong performance in semi-supervised learning (SSL), which incorporate discrete class information into the learning process. Yet existing methods generally overfit to the given labeled data, for only considering the conditional probability of labels. In this paper, we propose a novel discriminative regularized deep generative method for SSL, which fully exploits the discriminative and geometric information of data to address the aforementioned issue. Our method introduces the cluster and manifold assumption that maximizes the classification margin between clusters and simultaneously smooths the predictions of the data which is close in the sub-manifold of each cluster, to regularize the learning of the classifier in DGMs. To derive the regularization based on introduced assumptions, we adopt the generated data of DGMs along with labelled and unlabelled data, to model the data manifold and yield clusters based on the Gumbel-softmax distribution. Experimental results on both text and image datasets demonstrate the effectiveness and flexibility of our method, and prove that two introduced assumptions are complementary in guiding the classification boundary, thus improving the discriminative ability of the classifier.

Existing generative adversarial networks (GANs) with manifold regularization for semi-supervised learning (SSL) have shown promising performance in image generation and semi-supervised learning (SSL), which penalize the smoothness of classifier over data manifold based on the smoothness assumption. However, the smoothness assumption is valid for data points in high density region while not hold for data points in low density region, thus they tend to misclassify boundary instances in low density region. In this paper, we propose a novel discriminative regularization method for semi-supervised learning with conditional generative adversarial nets (CGANs). In our method, the discriminative information from class conditional data distribution captured by CGANs is utilized to improve the discrimination of classifier. Different from regular manifold regularization, the discriminative regularization encourages the classifier invariance to local perturbations on the sub-manifold of each cluster, and distinct classification outputs for data points in different clusters. Moreover, our method can be easily implemented via the stochastic approximation without constructing the Laplacian graph or computing the Jacobian of classifier explicitly. Experimental results on benchmark datasets show that our method can achieve competitive performance against previous advanced methods.

Sparse topic models (STMs) are widely used for learning a semantically rich latent sparse representation of short texts in large scale, mainly by imposing sparse priors or appropriate regularizers on topic models. However, it is difficult for these STMs to model the sparse structure and pattern of the corpora accurately, since their sparse priors always fail to achieve real sparseness, and their regularizers bypass the prior information of the relevance between sparse coefficients. In this paper, we propose a novel Bayesian hierarchical topic models called Bayesian Sparse Topical Coding with Poisson Distribution (BSTC-P) on the basis of Sparse Topical Coding with Sparse Groups (STCSG). Different from traditional STMs, it focuses on imposing hierarchical sparse prior to leverage the prior information of relevance between sparse coefficients. Furthermore, we propose a sparsity-enhanced BSTC, Bayesian Sparse Topical Coding with Normal Distribution (BSTC-N), via mathematic approximation. We adopt superior hierarchical sparse inducing prior, with the purpose of achieving the sparsest optimal solution. Experimental results on datasets of Newsgroups and Twitter show that both BSTC-P and BSTC-N have better performance on finding clear latent semantic representations. Therefore, they yield better performance than existing works on document classification tasks.

Min Peng, Qianqian Xie, Yanchun Zhang, Hua Wang, Xiuzhen Zhang, Jimin Huang, Gang Tian. Proceedings of the 56th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). 2018.

Learning low dimensional representations from a large number of short corpora has a profound practical significance but with vital challenge in content analysis and data mining applications. In this paper, we propose a novel topic model called Block Bayesian Sparse Topic Coding (Block-BSTC), which is capable of discovering the latent semantic representation of short texts. The Block-BSTC relaxes the normalization constraint of the inferred representations with word embeddings and block sparse Bayesian learning, which is convenient to directly control the sparsity of word codes with exploiting the intra-block correlations. Furthermore, the experimental results show that Block-BSTC achieves great performance on the sparsity ratio of word codes. Meanwhile, it can improve the accuracy of document classification.

Compressing textstreams generated by social networks can both reduce storage consumption and improve efficiency such as fast searching. However, the compression process is a challenge due to the large scale of textstreams. In this article, we propose a textstream compression framework based on compressed sensing theory and design a series of matching parallel procedures. The new approach uses a linear projection technique in the textstream compression process, achieving fast compression speed and low compression ratio. Two processes are executed by designing elaborated parallel procedures for efficient compressing and decompressing of large-scale textstreams. The decompression process is implemented for approximate solutions of underdetermined linear systems. Experimental results show that the new method can efficiently achieve the compression and decompression tasks on a large amount of text generated by social networks.