Author name cluster

Zhixu Li

Possible papers associated with this exact author name in Arrow. This page groups case-insensitive exact name matches and is not a full identity disambiguation profile.

12 papers

2 author rows

AAAI Conference 2026 Conference Paper

Adaptive Hallucination Alleviation in Multimodal Large Language Models: From Strategic Data Selection to Severity-Guided Training

Yuanyi Xu
Xiangru Zhu
Sihang Jiang
Zhixu Li
Bei Yang
Xiaoxiao Xu
Yanghua Xiao
Wei Wang

Multimodal Large Language Models (MLLMs) have recently achieved strong performance across a variety of multimodal tasks. However, they still suffer from various forms of hallucination, which hinder their practical deployment. Prior approaches often struggle to efficiently construct high-quality hallucination-related samples and to process them in a fine-grained manner, resulting in limited effectiveness in hallucination alleviation. To address this issue, we propose a data sampling strategy that selects samples better suited for hallucination-oriented training, thereby enhancing training effectiveness. In addition, we introduce a quantitative method for measuring hallucination severity and assign individualized weights to training samples accordingly. Building on this, we present Hallucination-Differentiated Direct Preference Optimization (HD-DPO), a novel preference optimization framework. During fine-tuning, HD-DPO incorporates these weights into both the formulation of customized loss functions and the modulation of localized visual attention, enabling fine-grained optimization. Experimental results demonstrate that our method outperforms existing fine-tuning strategies across multiple benchmarks and generalizes well to diverse MLLM architectures, effectively reducing hallucination rates and enhancing overall model performance.

PDF Details DOI

AAAI Conference 2025 Conference Paper

Enhancing Long-and Short-Term Representations for Next POI Recommendations via Frequency and Hierarchical Contrastive Learning

Jiajie Chen
Yu Sang
Peng-Fei Zhang
Jiaan Wang
Jianfeng Qu
Zhixu Li

Next POI recommendation aids users in predicting their destinations of interest and plays an increasingly vital role in location-based social services. Recent works focus on analyzing both long-term and short-term interests in POI recommendation to gain a deeper understanding of user profiles. However, these methods for modeling long-term user’s sequences primarily rely on the Transformer model, which functions as a low-pass filter, often leading to the loss of high-frequency information. Additionally, long-term and short-term sequences are typically modeled independently, with short-term sequences often defined solely by the most recent check-ins, overlooking their interactions and dependencies. Therefore, we propose Enhancing Long-and Short-Term Representations for Next POI Recommendations via Frequency and Hierarchical Contrastive Learning (FHCRec). FHCRec captures both high-frequency and low-frequency information in long-term sequences to model richer long-term user’s preference representations. Moreover, it harnesses the characteristics of the short-term subsequences embedded within long-term sequences to enhance short-term preference characterization via local and global hierarchical contrastive learning, resulting in more personalized short-term preferences. The enhanced long-term and short-term preferences are integrated to improve model recommendation performance. Extensive experiments on three real-world datasets demonstrate the effectiveness of our method.

PDF Details DOI

ICLR Conference 2025 Conference Paper

Evaluating Semantic Variation in Text-to-Image Synthesis: A Causal Perspective

Xiangru Zhu
Penglei Sun
Yaoxian Song
Yanghua Xiao
Zhixu Li
Chengyu Wang 0001
Jun Huang 0007
Bei Yang

Accurate interpretation and visualization of human instructions are crucial for text-to-image (T2I) synthesis. However, current models struggle to capture semantic variations from word order changes, and existing evaluations, relying on indirect metrics like text-image similarity, fail to reliably assess these challenges. This often obscures poor performance on complex or uncommon linguistic patterns by the focus on frequent word combinations. To address these deficiencies, we propose a novel metric called SemVarEffect and a benchmark named SemVarBench, designed to evaluate the causality between semantic variations in inputs and outputs in T2I synthesis. Semantic variations are achieved through two types of linguistic permutations, while avoiding easily predictable literal variations. Experiments reveal that the CogView-3-Plus and Ideogram 2 performed the best, achieving a score of 0.2/1. Semantic variations in object relations are less understood than attributes, scoring 0.07/1 compared to 0.17-0.19/1. We found that cross-modal alignment in UNet or Transformers plays a crucial role in handling semantic variations, a factor previously overlooked by a focus on textual encoders. Our work establishes an effective evaluation framework that advances the T2I synthesis community's exploration of human instruction understanding. Our benchmark and code are available at https://github.com/zhuxiangru/SemVarBench.

Details

AAAI Conference 2025 Conference Paper

SQLFixAgent: Towards Semantic-Accurate Text-to-SQL Parsing via Consistency-Enhanced Multi-Agent Collaboration

Jipeng Cen
Jiaxin Liu
Zhixu Li
Jingjing Wang

While fine-tuned large language models (LLMs) excel in generating grammatically valid SQL in Text-to-SQL parsing, they often struggle to ensure semantic accuracy in queries, leading to user confusion and diminished system usability. To tackle this challenge, we introduce SQLFixAgent, a new consistency-enhanced multi-agent collaborative framework designed for detecting and repairing erroneous SQL. Our framework comprises a core agent, SQLRefiner, alongside two auxiliary agents: SQLReviewer and QueryCrafter. The SQLReviewer agent employs the rubber duck debugging method to identify potential semantic mismatches between SQL and user query. If the error is detected, the QueryCrafter agent generates multiple SQL as candidate repairs using a fine-tuned SQLTool. Subsequently, leveraging similar repair retrieval and failure memory reflection, the SQLRefiner agent selects the most fitting SQL statement from the candidates as the final repair. We evaluated our proposed framework on five Text-to-SQL benchmarks. The experimental results show that our method consistently enhances the performance of the baseline model, specifically achieving an execution accuracy improvement of over 3% on the Bird benchmark. Our framework also has a higher token efficiency compared to other advanced methods, making it more competitive.

PDF Details DOI

ECAI Conference 2024 Conference Paper

Empowering Biomedical Named Entity Recognition through Multi-Tagger Collaboration

Jin Zhao 0004
Jian Xie
Tinghui Zhu
Qian Guo
Zhixu Li
Yanghua Xiao

Biomedical Named Entity Recognition (BioNER) plays a crucial role in automatically identifying specific categories of entities from biomedical texts. Currently, region-based methods have shown promising performance in BioNER. However, existing paradigms in the region-based methods suffer from inherent limitations, including the generation of negative samples, and the ignorance of token dependencies. To overcome these limitations, we propose a new paradigm, implemented as Token Cascade Tagger (TCT), which combines span identification and category classification. The TCT utilizes category information to enhance the correlation between the heads and tails of entities, effectively reducing the generation of negative samples. Additionally, we introduce a Token Dependency Tagger (TDT) that captures token dependencies within entity spans by identifying the longest span in a sentence. The TDT filters out incorrect spans and further improves the accuracy of span detection obtained from the TCT. Furthermore, we employ a multi-task learning framework to optimize both the TCT and TDT, leading to superior performance in BioNER. Extensive experiments on publicly available biomedical datasets demonstrate our method outperforms the previous state-of-the-art methods, achieving 92. 44%, 92. 54%, and 81. 26% on NCBI-Disease, BC5CDR, and GENIA, respectively, in terms of F1 score.

Details

AAAI Conference 2024 Conference Paper

Improving the Robustness of Knowledge-Grounded Dialogue via Contrastive Learning

Jiaan Wang
Jianfeng Qu
Kexin Wang
Zhixu Li
Wen Hua
Ximing Li
An Liu

Knowledge-grounded dialogue (KGD) learns to generate an informative response based on a given dialogue context and external knowledge (e.g., knowledge graphs; KGs). Recently, the emergence of large language models (LLMs) and pre-training techniques has brought great success to knowledge-grounded dialogue. However, when building KGD systems in real applications, there are various real-world noises that are inevitable to face. For example, the dialogue context might involve perturbations such as misspellings and abbreviations. In addition, KGs typically suffer from incompletion and also might contain erroneous and outdated facts. Such real-world noises pose a challenge to the robustness of KGD systems and hinder their applications in the real world. In this paper, we propose an entity-based contrastive learning framework for improving the robustness of KGD. Specifically, we make use of the entity information in a KGD sample to create both its positive and negative samples which involve semantic-irrelevant and semantic-relevant perturbations, respectively. The contrastive learning framework ensures the KGD model is aware of these two types of perturbations, thus could generate informative responses with the potentially noisy inputs in real applications. Experimental results on three widely-used benchmark datasets show that our method achieves new state-of-the-art performance in terms of automatic evaluation scores, verifying its effectiveness and potentiality. Furthermore, we show that our method is able to generate better responses than comparison models in both the noisy and the few-shot settings.

PDF Details DOI

AAAI Conference 2024 Conference Paper

Talk Funny! A Large-Scale Humor Response Dataset with Chain-of-Humor Interpretation

Yuyan Chen
Yichen Yuan
Panjun Liu
Dayiheng Liu
Qinghao Guan
Mengfei Guo
Haiming Peng
Bang Liu

Humor is a crucial part of human communication. Understanding humor and generating humorous responses in dialogue can provide natural and empathic human-computer interactions. However, most existing pre-trained language models (PLMs) perform unsatisfactorily in humor generation. On one hand, the serious shortage of humor corpus and datasets pose challenges for constructing models that can understand and generate humorous expressions. On the other hand, humor generation relies on rich knowledge and commonsense, which is often tacit and unspoken. In this paper, we construct the largest Chinese Explainable Humor Response Dataset to date with chain-of-humor and humor mind map annotations, which can be used to comprehensively evaluate as well as improve the humorous response ability of PLMs. We further design humor-related auxiliary tasks to further enhance PLMs' humorous response performance. Extensive evaluations demonstrate that our proposed dataset and auxiliary tasks effectively help PLMs to generate humorous responses, laying the groundwork for future humor research.

PDF Details DOI

AAAI Conference 2023 Conference Paper

Competition or Cooperation? Exploring Unlabeled Data via Challenging Minimax Game for Semi-supervised Relation Extraction

Yu Hong
Jiahang Li
Jianchuan Feng
Chenghua Huang
Zhixu Li
Jianfeng Qu
Yanghua Xiao
Wei Wang

Semi-Supervised Relation Extraction aims at learning well-performed RE models with limited labeled and large-scale unlabeled data. Existing methods mainly suffer from semantic drift and insufficient supervision, which severely limit the performance. To address these problems, recent work tends to design dual modules to work cooperatively for mutual enhancement. However, the consensus of two modules greatly restricts the model from exploring diverse relation expressions in unlabeled set, which hinders the performance as well as model generalization. To tackle this problem, in this paper, we propose a novel competition-based method AdvSRE. We set up a challenging minimax game on unlabeled data between two modules, Generator and Discriminator, and assign them with conflicting objectives. During the competition game, one module may find any possible chance to beat the other, which develops two modules' abilities until relation expressions cannot be further explored. To exploit label information, Discriminator is further asked to predict specific relation for each sentence. Experiment results on two benchmarks show new state-of-the-art performance over baselines, demonstrating the effectiveness of proposed AdvSRE.

PDF Details DOI

AAAI Conference 2023 Conference Paper

GANTEE: Generative Adversarial Network for Taxonomy Enterance Evaluation

Zhouhong Gu
Sihang Jiang
Jingping Liu
Yanghua Xiao
Hongwei Feng
Zhixu Li
Jiaqing Liang
Zhong Jian

Taxonomy is formulated as directed acyclic graphs or trees of concepts that support many downstream tasks. Many new coming concepts need to be added to an existing taxonomy. The traditional taxonomy expansion task aims only at finding the best position for new coming concepts in the existing taxonomy. However, they have two drawbacks when being applied to the real-scenarios. The previous methods suffer from low-efficiency since they waste much time when most of the new coming concepts are indeed noisy concepts. They also suffer from low-effectiveness since they collect training samples only from the existing taxonomy, which limits the ability of the model to mine more hypernym-hyponym relationships among real concepts. This paper proposes a pluggable framework called Generative Adversarial Network for Taxonomy Entering Evaluation (GANTEE) to alleviate these drawbacks. A generative adversarial network is designed in this framework by discriminative models to alleviate the first drawback and the generative model to alleviate the second drawback. Two discriminators are used in GANTEE to provide long-term and short-term rewards, respectively. Moreover, to further improve the efficiency, pre-trained language models are used to retrieve the representation of the concepts quickly. The experiments on three real-world large-scale datasets with two different languages show that GANTEE improves the performance of the existing taxonomy expansion methods in both effectiveness and efficiency.

PDF Details DOI

IS Journal 2020 Journal Article

Personalized Geographical Influence Modeling for POI Recommendation

Yanan Zhang
Guanfeng Liu
An Liu
Yifan Zhang
Zhixu Li
Xiangliang Zhang
Qing Li

Point-of-interest (POI) recommendation has great significance in helping users find favorite places from a large number of candidate venues. One challenging in POI recommendation is to effectively exploit geographical information since users usually care about the physical distance to the recommended POIs. Though spatial relevance has been widely considered in recent recommendation methods, it is modeled only from the POI perspective, failing to capture user personalized preference to spatial distance. Moreover, these methods suffer from a diversity-deficiency problem since they are often based on collaborative filtering which always favors popular POIs. To overcome these problems, we propose in this article a personalized geographical influence modeling method called PGIM, which jointly learns users’ geographical preference and diversity preference for POI recommendation. Specifically, we model geographical preference from three aspects: user global tolerance, user local tolerance, and spatial distance. We also extract user diversity preference from interactions among users for diversity-promoting recommendation. Experimental results on three real-world datasets demonstrate the superiority of PGIM.

Details DOI

IJCAI Conference 2019 Conference Paper

DMRAN: A Hierarchical Fine-Grained Attention-Based Network for Recommendation

Huizhao Wang
Guanfeng Liu
An Liu
Zhixu Li
Kai Zheng

The conventional methods for the next-item recommendation are generally based on RNN or one- dimensional attention with time encoding. They are either hard to preserve the long-term dependencies between different interactions, or hard to capture fine-grained user preferences. In this paper, we propose a Double Most Relevant Attention Network (DMRAN) that contains two layers, i. e. , Item level Attention and Feature Level Self- attention, which are to pick out the most relevant items from the sequence of user’s historical behaviors, and extract the most relevant aspects of relevant items, respectively. Then, we can capture the fine-grained user preferences to better support the next-item recommendation. Extensive experiments on two real-world datasets illustrate that DMRAN can improve the efficiency and effectiveness of the recommendation compared with the state-of-the-art methods.

PDF Details

AAAI Conference 2019 Conference Paper

Where to Go Next: A Spatio-Temporal Gated Network for Next POI Recommendation

Pengpeng Zhao
Haifeng Zhu
Yanchi Liu
Jiajie Xu
Zhixu Li
Fuzhen Zhuang
Victor S. Sheng
Xiaofang Zhou

Next Point-of-Interest (POI) recommendation is of great value for both location-based service providers and users. However, the state-of-the-art Recurrent Neural Networks (RNNs) rarely consider the spatio-temporal intervals between neighbor check-ins, which are essential for modeling user check-in behaviors in next POI recommendation. To this end, in this paper, we propose a new Spatio-Temporal Gated Network (STGN) by enhancing long-short term memory network, where spatio-temporal gates are introduced to capture the spatio-temporal relationships between successive checkins. Specifically, two pairs of time gate and distance gate are designed to control the short-term interest and the longterm interest updates, respectively. Moreover, we introduce coupled input and forget gates to reduce the number of parameters and further improve efficiency. Finally, we evaluate the proposed model using four real-world datasets from various location-based social networks. The experimental results show that our model significantly outperforms the state-ofthe-art approaches for next POI recommendation.

PDF Details