Author name cluster

Linlin Wang

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.

11 papers

1 author row

AAAI Conference 2025 Conference Paper

Hierarchical Divide-and-Conquer for Fine-Grained Alignment in LLM-Based Medical Evaluation

Shunfan Zheng
Xiechi Zhang
Gerard de Melo
Xiaoling Wang
Linlin Wang

In the rapidly evolving landscape of large language models (LLMs) for medical applications, ensuring the reliability and accuracy of these models in clinical settings is paramount. Existing benchmarks often focus on fixed-format tasks like multiple-choice QA, which fail to capture the complexity of real-world clinical diagnostics. Moreover, traditional evaluation metrics and LLM-based evaluators struggle with misalignment, often providing oversimplified assessments that do not adequately reflect human judgment. To address these challenges, we introduce HDCEval, a Hierarchical Divide-and-Conquer Evaluation framework tailored for fine-grained alignment in medical evaluation. HDCEval is built on a set of fine-grained medical evaluation guidelines developed in collaboration with professional doctors, encompassing Patient Question Relevance, Medical Knowledge Correctness, and Expression. The framework decomposes complex evaluation tasks into specialized subtasks, each evaluated by expert models trained through Attribute-Driven Token Optimization (ADTO) on a meticulously curated preference dataset. This hierarchical approach ensures that each aspect of the evaluation is handled with expert precision, leading to a significant improvement in alignment with human evaluators.

PDF Details DOI

AAAI Conference 2025 Conference Paper

NLSR: Neuron-Level Safety Realignment of Large Language Models Against Harmful Fine-Tuning

Xin Yi
Shunfan Zheng
Linlin Wang
Gerard de Melo
Xiaoling Wang
Liang He

The emergence of fine-tuning-as-a-service has revealed a new vulnerability in large language models (LLMs). A mere handful of malicious data uploaded by users can subtly manipulate the fine-tuning process, leading to a compromised alignment state. Existing methods to counteract fine-tuning attacks typically require substantial computational resources. Even with parameter-efficient techniques like LoRA, gradient updates remain essential. To address these challenges, we propose Neuron-Level Safety Realignment (NLSR), a training-free framework that restores the safety of LLMs based on the similarity difference of safety-critical neurons before and after fine-tuning. The core of our framework is first to construct a safety reference model from an initially aligned model to amplify safety-related features in neurons. We then utilize this reference model to identify safety-critical neurons, which we prepare as patches. Finally, we selectively restore only those neurons that exhibit significant similarity differences by transplanting these prepared patches, thereby minimally altering the fine-tuned model. Extensive experiments demonstrate significant safety enhancements in fine-tuned models across multiple downstream tasks, while greatly maintaining task-level accuracy. Our findings indicate that safety-critical neurons exhibit significant regional variations after fine-tuning, which can be effectively corrected through neuron transplantation from the reference model without the need for additional training.

PDF Details DOI

AAAI Conference 2024 Conference Paper

MedBench: A Large-Scale Chinese Benchmark for Evaluating Medical Large Language Models

Yan Cai
Linlin Wang
Ye Wang
Gerard de Melo
Ya Zhang
Yanfeng Wang
Liang He

The emergence of various medical large language models (LLMs) in the medical domain has highlighted the need for unified evaluation standards, as manual evaluation of LLMs proves to be time-consuming and labor-intensive. To address this issue, we introduce MedBench, a comprehensive benchmark for the Chinese medical domain, comprising 40,041 questions sourced from authentic examination exercises and medical reports of diverse branches of medicine. In particular, this benchmark is composed of four key components: the Chinese Medical Licensing Examination, the Resident Standardization Training Examination, the Doctor In-Charge Qualification Examination, and real-world clinic cases encompassing examinations, diagnoses, and treatments. MedBench replicates the educational progression and clinical practice experiences of doctors in Mainland China, thereby establish- ing itself as a credible benchmark for assessing the mastery of knowledge and reasoning abilities in medical language learning models. We perform extensive experiments and conduct an in-depth analysis from diverse perspectives, which culminate in the following findings: (1) Chinese medical LLMs underperform on this benchmark, highlighting the need for significant advances in clinical knowledge and diagnostic precision. (2) Several general-domain LLMs surprisingly possess considerable medical knowledge. These findings elucidate both the capabilities and limitations of LLMs within the context of MedBench, with the ultimate goal of aiding the medical research community.

PDF Details DOI

YNICL Journal 2024 Journal Article

The brain topological alterations in the structural connectome and correlations with clinical characteristics in type 1 narcolepsy

Huiqin Zhang
Lin Xu
Zhu Ai
Linlin Wang
Lu Wang
Lili Li
Ruilin Zhang
Rong Xue

OBJECTIVE: To explore topological alterations of white matter (WM) structural connectome, and their associations with clinical characteristics in type 1 narcolepsy (NT1). METHODS: 46 NT1 patients and 34 age- and sex-matched healthy controls were recruited for clinical data and diffusion tensor imaging collection. Using graph theory analysis, the topology metrics of structural connectome, rich club organization, and connectivity properties were compared between two groups. Furthermore, partial correlation analysis was performed between the network characteristics of 90 nodes or weakened edges and clinical data using Pearson or Spearman correlation, controlling by age and sex. RESULTS: Between-group comparison reflected that NT1 patients exhibited sleep disorders with comorbidities of impaired cognition and psychological problems. In patients, the global efficiency, local efficiency, and average clustering coefficient were significantly lower, whereas characteristic path length was larger compared to healthy control. Pertinently, nodal path length of left middle frontal gyrus was positively correlated with Pittsburgh Sleep Quality Index scores. The rich club analysis identified six affected nodes: bilateral dorsolateral superior frontal gyrus, bilateral supplementary motor area, left hippocampus, and left pallidum. Furthermore, six significantly weakened structural connections seeding from these rich club nodes have shown significant correlations with clinical index or polysomnography parameters. CONCLUSION: In NT1 patients, WM structural connectome has shown to be disrupted, which were primarily distributed in frontal-parietal cortex, subcortical regions, and particularly cingulate, potentially affecting their clinical manifestations.

Details DOI

AAAI Conference 2023 Conference Paper

A Disentangled-Attention Based Framework with Persona-Aware Prompt Learning for Dialogue Generation

Pingsheng Liu
Zhengjie Huang
Xiechi Zhang
Linlin Wang
Gerard de Melo
Xin Lin
Liang Pang
Liang He

Endowing dialogue agents with personas is the key to delivering more human-like conversations. However, existing persona-grounded dialogue systems still lack informative details of human conversations and tend to reply with inconsistent and generic responses. One of the main underlying causes is that pre-defined persona sentences are generally short and merely superficial descriptions of personal attributes, making appropriate persona selection and understanding non-trivial. Another challenge is that it is crucial to consider the context and the conversation flow to dynamically determine when to invoke different types of persona signals. To address these problems, we propose a disentangled-attention based pre-training architecture, which incorporates persona-aware prompt learning to bridge the connection between the selected persona and response generation. Our model first exploits the conversation flow to select context-relevant personas, and subsequently enriches the superficial persona descriptions with extra personality traits through persona-aware prompting. Finally, the decoder leverages a disentangled-attention mechanism to flexibly control the reliance on personas and dialogue contexts, and incorporates A*-like keyword-based heuristic estimates for controllable generation. Extensive experiments show that our approach can outperform strong baselines and deliver more consistent and engaging responses on the PERSONA-CHAT dataset.

PDF Details DOI

AAAI Conference 2023 Conference Paper

Disentangled CVAEs with Contrastive Learning for Explainable Recommendation

Linlin Wang
Zefeng Cai
Gerard de Melo
Zhu Cao
Liang He

Modern recommender systems are increasingly expected to provide informative explanations that enable users to understand the reason for particular recommendations. However, previous methods struggle to interpret the input IDs of user--item pairs in real-world datasets, failing to extract adequate characteristics for controllable generation. To address this issue, we propose disentangled conditional variational autoencoders (CVAEs) for explainable recommendation, which leverage disentangled latent preference factors and guide the explanation generation with the refined condition of CVAEs via a self-regularization contrastive learning loss. Extensive experiments demonstrate that our method generates high-quality explanations and achieves new state-of-the-art results in diverse domains.

PDF Details DOI

JBHI Journal 2021 Journal Article

Identification of lncRNA Signature Associated With Pan-Cancer Prognosis

Guoqing Bao
Ran Xu
Xiuying Wang
Jianxiong Ji
Linlin Wang
Wenjie Li
Qing Zhang
Bin Huang

Long noncoding RNAs (lncRNAs) have emerged as potential prognostic markers in various human cancers as they participate in many malignant behaviors. However, the value of lncRNAs as prognostic markers among diverse human cancers is still under investigation, and a systematic signature based on these transcripts that related to pan-cancer prognosis has yet to be reported. In this study, we proposed a framework to incorporate statistical power, biological rationale, and machine learning models for pan-cancer prognosis analysis. The framework identified a 5-lncRNA signature ( ENSG00000206567, PCAT29, ENSG00000257989, LOC388282, and LINC00339 ) from TCGA training studies ( n = 1, 878). The identified lncRNAs are significantly associated (all P $\leq$ 1. 48E-11) with overall survival (OS) of the TCGA cohort ( n = 4, 231). The signature stratified the cohort into low- and high-risk groups with significantly distinct survival outcomes (median OS of 9. 84 years versus 4. 37 years, log-rank P = 1. 48E-38) and achieved a time-dependent ROC/AUC of 0. 66 at 5 years. After routine clinical factors involved, the signature demonstrated better performance for long-term prognostic estimation (AUC of 0. 72). Moreover, the signature was further evaluated on two independent external cohorts (TARGET, n = 1, 122; CPTAC, n = 391; National Cancer Institute) which yielded similar prognostic values (AUC of 0. 60 and 0. 75; log-rank P = 8. 6E-09 and P = 2. 7E-06). An indexing system was developed to map the 5-lncRNA signature to prognoses of pan-cancer patients. In silico functional analysis indicated that the lncRNAs are associated with common biological processes driving human cancers. The five lncRNAs, especially ENSG00000206567, ENSG00000257989 and LOC388282 that never reported before, may serve as viable molecular targets common among diverse cancers.

Details DOI

AAAI Conference 2018 Conference Paper

Link Prediction via Subgraph Embedding-Based Convex Matrix Completion

Zhu Cao
Linlin Wang
Gerard de Melo

Link prediction is of fundamental importance in network science and machine learning. Early methods consider only simple topological features, while subsequent supervised approaches typically rely on human-labeled data and feature engineering. In this work, we present a new representation learning-based approach called SEMAC that jointly exploits ﬁne-grained node features as well as the overall graph topology. In contrast to the SGNS or SVD methods espoused in previous representation-based studies, our model represents nodes in terms of subgraph embeddings acquired via a form of convex matrix completion to iteratively reduce the rank, and thereby, more effectively eliminate noise in the representation. Thus, subgraph embeddings and convex matrix completion are elegantly integrated into a novel link prediction framework. Experimental results on several datasets show the effectiveness of our method compared to previous work.

PDF Details

IJCAI Conference 2017 Conference Paper

Multiple-Weight Recurrent Neural Networks

Zhu Cao
Linlin Wang
Gerard de Melo

Recurrent neural networks (RNNs) have enjoyed great success in speech recognition, natural language processing, etc. Many variants of RNNs have been proposed, including vanilla RNNs, LSTMs, and GRUs. However, current architectures are not particularly adept at dealing with tasks involving multi-faceted contents. In this work, we solve this problem by proposing Multiple-Weight RNNs and LSTMs, which rely on multiple weight matrices in an attempt to mimic the human ability of switching between contexts. We present a framework for adapting RNN-based models and analyze the properties of this approach. Our detailed experimental results show that our model outperforms previous work across a range of different tasks and datasets.

PDF Details

YNIMG Journal 2016 Journal Article

Alterations of functional connectivities from early to middle adulthood: Clues from multivariate pattern analysis of resting-state fMRI data

Lixia Tian
Lin Ma
Linlin Wang

In contrast to extended research interests in the maturation and aging of human brain, alterations of brain structure and function from early to middle adulthood have been much less studied. The aim of the present study was to investigate the extent and pattern of the alterations of functional interactions between brain regions from early to middle adulthood. We carried out the study by multivariate pattern analysis of resting-state fMRI (RS-fMRI) data of 63 adults aged 18 to 45years. Specifically, using elastic net, we performed brain age estimation and age-group classification (young adults aged 18–28years vs. middle-aged adults aged 35–45years) based on the resting-state functional connectivities (RSFCs) between 160 regions of interest (ROIs) evaluated on the RS-fMRI data of each subject. The results indicate that the estimated brain ages were significantly correlated with the chronological age (R =0. 78, MAE=4. 81), and a classification rate of 94. 44% and area under the receiver operating characteristic curve (AUC) of 0. 99 were obtained when classifying the young and middle-aged adults. These results provide strong evidence that functional interactions between brain regions undergo notable alterations from early to middle adulthood. By analyzing the RSFCs that contribute to brain age estimation/age-group classification, we found that a majority of the RSFCs were inter-network, and we speculate that inter-network RSFCs might mature late but age early as compared to intra-network ones. In addition, the strengthening/weakening of the RSFCs associated with the left/right hemispheric ROIs, the weakening of cortico-cerebellar RSFCs and the strengthening of the RSFCs between the default mode network and other networks contributed much to both brain age estimation and age-group classification. All these alterations might reflect that aging of brain function is already in progress in middle adulthood. Overall, the present study indicated that the RSFCs undergo notable alterations from early to middle adulthood and highlighted the necessity of careful considerations of possible influences of these alterations in related studies.

Details DOI

AAAI Conference 2016 Conference Paper

Morphological Segmentation with Window LSTM Neural Networks

Linlin Wang
Zhu Cao
Yu Xia
Gerard de Melo

Morphological segmentation, which aims to break words into meaning-bearing morphemes, is an important task in natural language processing. Most previous work relies heavily on linguistic preprocessing. In this paper, we instead propose novel neural network architectures that learn the structure of input sequences directly from raw input words and are subsequently able to predict morphological boundaries. Our architectures rely on Long Short Term Memory (LSTM) units to accomplish this, but exploit windows of characters to capture more contextual information. Experiments on multiple languages conﬁrm the effectiveness of our models on this task.

PDF Details