Author name cluster

Ziyu Shang

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.

14 papers

1 author row

AAAI Conference 2026 Conference Paper

Balanced Knowledge Distillation for Large Language Models with Mix-of-Experts

Jiajun Liu
Yao He
Wenjun Ke
Peng Wang
Ziyu Shang
Guozheng Li
Zijie Xu

Mixture-of-Experts (MoE) architectures have recently become a more prevalent choice for large language models (LLMs) than dense architectures due to their superior performance. However, billions of parameters bring MoE LLMs a huge cost for deployment and inference. To address these issues, knowledge distillation (KD) has become a widely adopted technique to compress LLMs. Existing KD methods for LLMs can be divided into dense-to-dense and moe-to-dense distillation. Dense-to-dense distillation transfers knowledge between single dense LLMs, while moe-to-dense distillation attempts to transfer knowledge between the MoE LLMs and the dense LLMs. However, the architectural mismatch prevents the student from fully absorbing knowledge when distilling MoE LLMs. To address this limitation, we investigate a new distillation setting, moe-to-moe, which aims to fully leverage expert knowledge of teachers and enable the student to absorb it more effectively. Compared to dense-to-dense and moe-to-dense, moe-to-moe suffers from two imbalance issues. First, expert-coverage deficiency reflects an imbalanced knowledge transfer of teacher experts: traditional distillation utilizes only the few experts activated by the teacher router. Second, routing imbalance appears when the student routing distribution drifts from the teacher, which makes it difficult for students to learn how to distribute different experts. To overcome these issues, we propose a novel distillation framework for moe-to-moe, Balanced Distillation (B-Distill), which equally spreads teacher expertise across student experts while regularizing the student router toward teacher-consistent balance. First, to mitigate expert-coverage deficiency, we introduce Monte Carlo exploration, which stochastically perturbs router probabilities so every teacher and student expert is sampled without enlarging the search space. Second, to correct routing imbalance and avert load collapse, we propose an entropy-aware router distillation mechanism that aligns the student router with the teacher while curbing over-concentration. Experiments show that B-Distill outperforms baselines by up to 6.6% in Rouge-L.

PDF Details DOI

AAAI Conference 2026 Conference Paper

Benchmarking and Enhancing Rule Knowledge-Driven Reasoning of Large Language Models

Zijie Xu
Wenjun Ke
Peng Wang
Guozheng Li
Qingjian Ni
Jiajun Liu
Ziyu Shang
Jing Zhou

Large Language Models (LLMs) have demonstrated strong capabilities across diverse tasks under the example-driven learning paradigm. However, in high-stakes domains such as emergency response and industrial safety, historical incidents are scarce, confidential, or both, while concise rule books are abundant. We formalize this underexplored setting as rule knowledge-driven reasoning and ask: Can LLMs reason reliably when rules are plentiful but examples are nearly absent? To study this question, we introduce RULER, an automatic benchmark that generates 32K rigorously verified questions from 1K expert-curated emergency response rules to probe three core abilities: rule memorization, single-rule application, and multi-rule complex reasoning. RULER is further equipped with a hallucination-aware evaluation suite and novel relational metrics. A comprehensive empirical study of five representative LLMs and five enhancement strategies shows that, even when models achieve reliable performance on rule memorization and single-rule application, multi-rule complex reasoning plateaus at 5.4 on a 10-point scale. To address this limitation, we propose RAMPS, a Rule knowledge-Aware Monte Carlo Tree Search Process-reward Supervision framework. RAMPS injects rule knowledge priors into MCTS, distills 12K step-level traces without human annotation, and trains an advantage-based reward model that scores candidate reasoning paths during beam search inference. Experimental results show that RAMPS significantly improves multi-rule complex reasoning performance to 7.7.

PDF Details DOI

AAAI Conference 2026 Conference Paper

Optimizing LoRA Allocation of MoE with the Alignment of Topic Correlation

Hengyuan Xu
Wenjun Ke
Yao He
Jiajun Liu
Dong Nie
Peng Wang
Ziyu Shang
Zijie Xu

Mixture of experts (MoE) dynamically routes inputs to specialized expert networks to scale model capacity with low inference overhead. However, the excessive parameter growth in MoE models poses challenges in low-resource settings. To address these issues, MoE with parameter-efficient fine-tuning (PEFT) methods have emerged as a lightweight adaptation paradigm that distributes knowledge among experts via multiple LoRA blocks. Existing MoE-PEFT methods can be broadly categorized into External and Internal PEFT methods. External PEFT methods incorporate lightweight models into existing MoE architectures without modifying their routing, which limits the model’s parameter efficiency. To overcome these issues, Internal PEFT methods integrate MoE architectures into PEFT, enabling minimal parameter overhead. However, they still face two major challenges: (1) lack of expert functional differentiation, resulting in overlapping specialization across modules, and (2) absence of a structured attribution mechanism to guide expert selection based on semantic relevance. To alleviate these challenges, we propose TopicLoRA, a novel three-stage framework that leverages topic knowledge as semantic anchors to guide expert allocation. Specifically, (1) to address expert redundancy, we construct a topic-level prior graph using Graph Neural Network-enhanced representation learning over Big-Bench categories, enforcing structural separation among expert embeddings, and (2) to introduce semantic attribution, we design a dual-loss training mechanism that softly aligns input-query relevance with topic-guided routing distributions via KL divergence. Extensive experiments on representative datasets (e.g., MMLU, GSM8K, Flanv2) demonstrate that TopicLoRA outperforms state-of-the-art PEFT baselines by 2.40% on average in accuracy. Notably, the maximum improvement is 4.21%. Furthermore, ablation studies demonstrate that our framework's robustness to intricate topics and input sequence variations, which stems from the dual-loss training mechanism.

PDF Details DOI

AAAI Conference 2024 Conference Paper

Cross-Modal and Uni-Modal Soft-Label Alignment for Image-Text Retrieval

Hailang Huang
Zhijie Nie
Ziqiao Wang
Ziyu Shang

Current image-text retrieval methods have demonstrated impressive performance in recent years. However, they still face two problems: the inter-modal matching missing problem and the intra-modal semantic loss problem. These problems can significantly affect the accuracy of image-text retrieval. To address these challenges, we propose a novel method called Cross-modal and Uni-modal Soft-label Alignment (CUSA). Our method leverages the power of uni-modal pre-trained models to provide soft-label supervision signals for the image-text retrieval model. Additionally, we introduce two alignment techniques, Cross-modal Soft-label Alignment (CSA) and Uni-modal Soft-label Alignment (USA), to overcome false negatives and enhance similarity recognition between uni-modal samples. Our method is designed to be plug-and-play, meaning it can be easily applied to existing image-text retrieval models without changing their original architectures. Extensive experiments on various image-text retrieval models and datasets, we demonstrate that our method can consistently improve the performance of image-text retrieval and achieve new state-of-the-art results. Furthermore, our method can also boost the uni-modal retrieval performance of image-text retrieval models, enabling it to achieve universal retrieval. The code and supplementary files can be found at https://github.com/lerogo/aaai24_itr_cusa.

PDF Details DOI

IJCAI Conference 2024 Conference Paper

Domain-Hierarchy Adaptation via Chain of Iterative Reasoning for Few-shot Hierarchical Text Classification

Ke Ji
Peng Wang
Wenjun Ke
Guozheng Li
Jiajun Liu
Jingsheng Gao
Ziyu Shang

Recently, various pre-trained language models (PLMs) have been proposed to prove their impressive performances on a wide range of few-shot tasks. However, limited by the unstructured prior knowledge in PLMs, it is difficult to maintain consistent performance on complex hierarchically dependent tasks, especially when the downstream data is extremely scarce. The main challenge is how to transfer the unstructured semantic space in PLMs to the downstream domain hierarchy. Unlike previous work on hierarchical text classification (HTC) which directly performs multi-label classification or uses graph neural network (GNN) to inject label hierarchy, in this work, we study the HTC problem under a few-shot setting to adapt knowledge in PLMs from an unstructured manner to the downstream hierarchy. Technically, we design a simple yet effective method named Hierarchical Iterative Conditional Random Field (HierICRF) to search the most domain-challenging directions and exquisitely crafts domain-hierarchy adaptation as a hierarchical iterative language modeling problem, and then it encourages the model to make hierarchical consistency self-correction during the inference, thereby achieving knowledge transfer with hierarchical consistency preservation. We perform HierICRF on various architectures, and extensive experiments on two popular HTC datasets demonstrate that prompt with HierICRF significantly boosts the few-shot HTC performance with an average Micro-F1 by 28. 80% to 1. 50% and Macro-F1 by 36. 29% to 1. 5% over the previous state-of-the-art (SOTA) baselines under few-shot settings (1->16), while remaining SOTA hierarchical consistency performance.

PDF Details DOI

IJCAI Conference 2024 Conference Paper

Empirical Analysis of Dialogue Relation Extraction with Large Language Models

Guozheng Li
Zijie Xu
Ziyu Shang
Jiajun Liu
Ke Ji
Yikai Guo

Dialogue relation extraction (DRE) aims to extract relations between two arguments within a dialogue, which is more challenging than standard RE due to the higher person pronoun frequency and lower information density in dialogues. However, existing DRE methods still suffer from two serious issues: (1) hard to capture long and sparse multi-turn information, and (2) struggle to extract golden relations based on partial dialogues, which motivates us to discover more effective methods that can alleviate the above issues. We notice that the rise of large language models (LLMs) has sparked considerable interest in evaluating their performance across diverse tasks. To this end, we initially investigate the capabilities of different LLMs in DRE, considering both proprietary models and open-source models. Interestingly, we discover that LLMs significantly alleviate two issues in existing DRE methods. Generally, we have following findings: (1) scaling up model size substantially boosts the overall DRE performance and achieves exceptional results, tackling the difficulty of capturing long and sparse multi-turn information; (2) LLMs encounter with much smaller performance drop from entire dialogue setting to partial dialogue setting compared to existing methods; (3) LLMs deliver competitive or superior performances under both full-shot and few-shot settings compared to current state-of-the-art; (4) LLMs show modest performances on inverse relations but much stronger improvements on general relations, and they can handle dialogues of various lengths especially for longer sequences.

PDF Details DOI

IJCAI Conference 2024 Conference Paper

Fast and Continual Knowledge Graph Embedding via Incremental LoRA

Jiajun Liu
Wenjun Ke
Peng Wang
Jiahao Wang
Jinhua Gao
Ziyu Shang
Guozheng Li
Zijie Xu

Continual Knowledge Graph Embedding (CKGE) aims to efficiently learn new knowledge and simultaneously preserve old knowledge. Dominant approaches primarily focus on alleviating catastrophic forgetting of old knowledge but neglect efficient learning for the emergence of new knowledge. However, in real-world scenarios, knowledge graphs (KGs) are continuously growing, which brings a significant challenge to fine-tuning KGE models efficiently. To address this issue, we propose a fast CKGE framework (FastKGE), incorporating an incremental low-rank adapter (IncLoRA) mechanism to efficiently acquire new knowledge while preserving old knowledge. Specifically, to mitigate catastrophic forgetting, FastKGE isolates and allocates new knowledge to specific layers based on the fine-grained influence between old and new KGs. Subsequently, to accelerate fine-tuning, FastKGE devises an efficient IncLoRA mechanism, which embeds the specific layers into incremental low-rank adapters with fewer training parameters. Moreover, IncLoRA introduces adaptive rank allocation, which makes the LoRA aware of the importance of entities and adjusts its rank scale adaptively. We conduct experiments on four public datasets and two new datasets with a larger initial scale. Experimental results demonstrate that FastKGE can reduce training time by 34%-49% while still achieving competitive link prediction performance against state-of-the-art models on four public datasets (average MRR score of 21. 0% vs. 21. 1%). Meanwhile, on two newly constructed datasets, FastKGE saves 51%-68% training time and improves link prediction performance by 1. 5%.

PDF Details DOI

IJCAI Conference 2024 Conference Paper

Learning Multi-Granularity and Adaptive Representation for Knowledge Graph Reasoning

Ziyu Shang
Peng Wang
Wenjun Ke
Jiajun Liu
Hailang Huang
Guozheng Li
Chenxiao Wu
Jianghan Liu

Knowledge graph reasoning (KGR) aims to infer new factual triples from existing knowledge graphs (KGs). Recently, a new category of methods, possessing both transductive and inductive reasoning capabilities, has been proposed to tackle this task via learning entity-independent representations from local neighboring structures. However, these methods are plagued by inefficiency issues and they exclusively capture evidence from well-designed local structures, ignoring the correlation between the query and different structures within KGs. In this work, we first propose a novel multi-granularity and adaptive representation framework, MulGA, exploiting the connectivity subgraph to uniformly and hierarchically model query-related triples, relation paths, and subgraphs without explicitly extracting any graph structure, hence mitigating inefficiency issues. Second, we introduce a message-passing mechanism across connectivity subgraphs, facilitating all entities to attain query-related structural representations of diverse granularity levels, i. e. , triple and relation paths of different lengths. Third, we design a self-attention-based merging mechanism that allocates weights to different granularities and then consolidates them into subgraph granularity representations for reasoning. The systematic experiments have been conducted on 15 benchmarks and MulGA achieves a significant improvement in MRR by an average of 1. 5% on transductive and 2. 7% on inductive tasks than existing state-of-the-art methods. Moreover, MulGA boasts faster convergence speed, competitive inference time, and alleviates the over-smoothing prevalent in graph neural networks.

PDF Details DOI

IJCAI Conference 2024 Conference Paper

Meta In-Context Learning Makes Large Language Models Better Zero and Few-Shot Relation Extractors

Guozheng Li
Peng Wang
Jiajun Liu
Yikai Guo
Ke Ji
Ziyu Shang
Zijie Xu

Relation extraction (RE) is an important task that aims to identify the relationships between entities in texts. While large language models (LLMs) have revealed remarkable in-context learning (ICL) capability for general zero and few-shot learning, recent studies indicate that current LLMs still struggle with zero and few-shot RE. Previous studies are mainly dedicated to design prompt formats and select good examples for improving ICL-based RE. Although both factors are vital for ICL, if one can fundamentally boost the ICL capability of LLMs in RE, the zero and few-shot RE performance via ICL would be significantly improved. To this end, we introduce Micre (Meta In-Context learning of LLMs for Relation Extraction), a new meta-training framework for zero and few-shot RE where an LLM is tuned to do ICL on a diverse collection of RE datasets (i. e. , learning to learn in context for RE). Through meta-training, the model becomes more effectively to learn a new RE task in context by conditioning on a few training examples with no parameter updates or task-specific templates at inference time, enabling better zero and few-shot task generalization. We experiment Micre on various LLMs with different model scales and 12 public RE datasets, and then evaluate it on unseen RE benchmarks under zero and few-shot settings. Micre delivers comparable or superior performance compared to a range of baselines including supervised fine-tuning and typical in-context learning methods. We find that the gains are particular significant for larger model scales, and using a diverse set of the meta-training RE datasets is key to improvements. Empirically, we show that Micre can transfer the relation semantic knowledge via relation label name during inference on target RE datasets.

PDF Details DOI

AAAI Conference 2024 Conference Paper

OntoFact: Unveiling Fantastic Fact-Skeleton of LLMs via Ontology-Driven Reinforcement Learning

Ziyu Shang
Wenjun Ke
Nana Xiu
Peng Wang
Jiajun Liu
Yanhui Li
Zhizhao Luo
Ke Ji

Large language models (LLMs) have demonstrated impressive proficiency in information retrieval, while they are prone to generating incorrect responses that conflict with reality, a phenomenon known as intrinsic hallucination. The critical challenge lies in the unclear and unreliable fact distribution within LLMs trained on vast amounts of data. The prevalent approach frames the factual detection task as a question-answering paradigm, where the LLMs are asked about factual knowledge and examined for correctness. However, existing studies primarily focused on deriving test cases only from several specific domains, such as movies and sports, limiting the comprehensive observation of missing knowledge and the analysis of unexpected hallucinations. To address this issue, we propose OntoFact, an adaptive framework for detecting unknown facts of LLMs, devoted to mining the ontology-level skeleton of the missing knowledge. Specifically, we argue that LLMs could expose the ontology-based similarity among missing facts and introduce five representative knowledge graphs (KGs) as benchmarks. We further devise a sophisticated ontology-driven reinforcement learning (ORL) mechanism to produce error-prone test cases with specific entities and relations automatically. The ORL mechanism rewards the KGs for navigating toward a feasible direction for unveiling factual errors. Moreover, empirical efforts demonstrate that dominant LLMs are biased towards answering Yes rather than No, regardless of whether this knowledge is included. To mitigate the overconfidence of LLMs, we leverage a hallucination-free detection (HFD) strategy to tackle unfair comparisons between baselines, thereby boosting the result robustness. Experimental results on 5 datasets, using 32 representative LLMs, reveal a general lack of fact in current LLMs. Notably, ChatGPT exhibits fact error rates of 51.6% on DBpedia and 64.7% on YAGO, respectively. Additionally, the ORL mechanism demonstrates promising error prediction scores, with F1 scores ranging from 70% to 90% across most LLMs. Compared to the exhaustive testing, ORL achieves an average recall of 80% while reducing evaluation time by 35.29% to 63.12%.

PDF Details DOI

IJCAI Conference 2024 Conference Paper

Recall, Retrieve and Reason: Towards Better In-Context Relation Extraction

Guozheng Li
Peng Wang
Wenjun Ke
Yikai Guo
Ke Ji
Ziyu Shang
Jiajun Liu
Zijie Xu

Relation extraction (RE) aims to identify relations between entities mentioned in texts. Although large language models (LLMs) have demonstrated impressive in-context learning (ICL) abilities in various tasks, they still suffer from poor performances compared to most supervised fine-tuned RE methods. Utilizing ICL for RE with LLMs encounters two challenges: (1) retrieving good demonstrations from training examples, and (2) enabling LLMs exhibit strong ICL abilities in RE. On the one hand, retrieving good demonstrations is a non-trivial process in RE, which easily results in low relevance regarding entities and relations. On the other hand, ICL with an LLM achieves poor performance in RE while RE is different from language modeling in nature or the LLM is not large enough. In this work, we propose a novel recall-retrieve-reason RE framework that synergizes LLMs with retrieval corpora (training examples) to enable relevant retrieving and reliable in-context reasoning. Specifically, we distill the consistently ontological knowledge from training datasets to let LLMs generate relevant entity pairs grounded by retrieval corpora as valid queries. These entity pairs are then used to retrieve relevant training examples from the retrieval corpora as demonstrations for LLMs to conduct better ICL via instruction tuning. Extensive experiments on different LLMs and RE datasets demonstrate that our method generates relevant and valid entity pairs and boosts ICL abilities of LLMs, achieving competitive or new state-of-the-art performance on sentence-level RE compared to previous supervised fine-tuning methods and ICL-based methods.

PDF Details DOI

AAAI Conference 2024 Conference Paper

Towards Continual Knowledge Graph Embedding via Incremental Distillation

Jiajun Liu
Wenjun Ke
Peng Wang
Ziyu Shang
Jinhua Gao
Guozheng Li
Ke Ji
Yanhe Liu

Traditional knowledge graph embedding (KGE) methods typically require preserving the entire knowledge graph (KG) with significant training costs when new knowledge emerges. To address this issue, the continual knowledge graph embedding (CKGE) task has been proposed to train the KGE model by learning emerging knowledge efficiently while simultaneously preserving decent old knowledge. However, the explicit graph structure in KGs, which is critical for the above goal, has been heavily ignored by existing CKGE methods. On the one hand, existing methods usually learn new triples in a random order, destroying the inner structure of new KGs. On the other hand, old triples are preserved with equal priority, failing to alleviate catastrophic forgetting effectively. In this paper, we propose a competitive method for CKGE based on incremental distillation (IncDE), which considers the full use of the explicit graph structure in KGs. First, to optimize the learning order, we introduce a hierarchical strategy, ranking new triples for layer-by-layer learning. By employing the inter- and intra-hierarchical orders together, new triples are grouped into layers based on the graph structure features. Secondly, to preserve the old knowledge effectively, we devise a novel incremental distillation mechanism, which facilitates the seamless transfer of entity representations from the previous layer to the next one, promoting old knowledge preservation. Finally, we adopt a two-stage training paradigm to avoid the over-corruption of old knowledge influenced by under-trained new knowledge. Experimental results demonstrate the superiority of IncDE over state-of-the-art baselines. Notably, the incremental distillation mechanism contributes to improvements of 0.2%-6.5% in the mean reciprocal rank (MRR) score. More exploratory experiments validate the effectiveness of IncDE in proficiently learning new knowledge while preserving old knowledge across all time steps.

PDF Details DOI

AAAI Conference 2024 Conference Paper

Unify Named Entity Recognition Scenarios via Contrastive Real-Time Updating Prototype

Yanhe Liu
Peng Wang
Wenjun Ke
Guozheng Li
Xiye Chen
Jiteng Zhao
Ziyu Shang

Supervised named entity recognition (NER) aims to classify entity mentions into a fixed number of pre-defined types. However, in real-world scenarios, unknown entity types are continually involved. Naive fine-tuning will result in catastrophic forgetting on old entity types. Existing continual methods usually depend on knowledge distillation to alleviate forgetting, which are less effective on long task sequences. Moreover, most of them are specific to the class-incremental scenario and cannot adapt to the online scenario, which is more common in practice. In this paper, we propose a unified framework called Contrastive Real-time Updating Prototype (CRUP) that can handle different scenarios for NER. Specifically, we train a Gaussian projection model by a regularized contrastive objective. After training on each batch, we store the mean vectors of representations belong to new entity types as their prototypes. Meanwhile, we update existing prototypes belong to old types only based on representations of the current batch. The final prototypes will be used for the nearest class mean classification. In this way, CRUP can handle different scenarios through its batch-wise learning. Moreover, CRUP can alleviate forgetting in continual scenarios only with current data instead of old data. To comprehensively evaluate CRUP, we construct extensive benchmarks based on various datasets. Experimental results show that CRUP significantly outperforms baselines in continual scenarios and is also competitive in the supervised scenario.

PDF Details DOI

AAAI Conference 2023 Conference Paper

IterDE: An Iterative Knowledge Distillation Framework for Knowledge Graph Embeddings

Jiajun Liu
Peng Wang
Ziyu Shang
Chenxiao Wu

Knowledge distillation for knowledge graph embedding (KGE) aims to reduce the KGE model size to address the challenges of storage limitations and knowledge reasoning efficiency. However, current work still suffers from the performance drops when compressing a high-dimensional original KGE model to a low-dimensional distillation KGE model. Moreover, most work focuses on the reduction of inference time but ignores the time-consuming training process of distilling KGE models. In this paper, we propose IterDE, a novel knowledge distillation framework for KGEs. First, IterDE introduces an iterative distillation way and enables a KGE model to alternately be a student model and a teacher model during the iterative distillation process. Consequently, knowledge can be transferred in a smooth manner between high-dimensional teacher models and low-dimensional student models, while preserving good KGE performances. Furthermore, in order to optimize the training process, we consider that different optimization objects between hard label loss and soft label loss can affect the efficiency of training, and then we propose a soft-label weighting dynamic adjustment mechanism that can balance the inconsistency of optimization direction between hard and soft label loss by gradually increasing the weighting of soft label loss. Our experimental results demonstrate that IterDE achieves a new state-of-the-art distillation performance for KGEs compared to strong baselines on the link prediction task. Significantly, IterDE can reduce the training time by 50% on average. Finally, more exploratory experiments show that the soft-label weighting dynamic adjustment mechanism and more fine-grained iterations can improve distillation performance.

PDF Details DOI