Author name cluster

Shasha 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.

11 papers

1 author row

AAAI Conference 2026 Conference Paper

Beyond Plain Demos: A Demo-Centric Anchoring Paradigm for In-Context Learning in Alzheimer’s Disease Detection

Puzhen Su
haoran yin
Miao Yongzhu
Jintao Tang
Shasha Li
Ting Wang

Detecting Alzheimer’s disease (AD) from narrative transcripts challenges large language models (LLMs): pre-training rarely covers this out-of-distribution task, and all transcript demos describe the same scene, producing highly homogeneous contexts. These factors cripple both the model’s built-in task knowledge (task cognition) and its ability to surface subtle, class-discriminative cues (contextual perception). Because cognition is fixed after pre-training, improving in-context learning (ICL) for AD detection hinges on enriching perception through better demonstration (demo) sets. We demonstrate that standard ICL quickly saturates, its demos lack diversity (context width) and fail to convey fine-grained signals (context depth), and that recent task vector (TV) approaches improve broad task adaptation by injecting TV into the LLMs' hidden states (HSs), they are ill-suited for AD detection due to the mismatch of injection granularity, strength and position. To address these bottlenecks, we introduce DA4ICL, a demo-centric anchoring framework that jointly expands context width via Diverse and Contrastive Retrieval (DCR) and deepens each demo's signal via Projected Vector Anchoring (PVA) at every Transformer layer. Across three AD benchmarks, DA4ICL achieves large, stable gains over both ICL and TV baselines, charting a new paradigm for fine-grained, OOD and low-resource LLM adaptation.

PDF Details DOI

AAAI Conference 2026 Conference Paper

Leveraging Image as Compressed Visual Prompt and Hierarchical Visual Knowledge for Effective Image Utilization in MLLMs

Shezheng Song
Kangcheng Ding
Shan Zhao
Shasha Li
Xiaopeng Li
Chengyu Wang
Qian Wan
Bin Ji

Multimodal Large Language Models (MLLMs) integrate text and images for complex reasoning tasks, but efficiently utilizing image remains a challenge due to redundancy and noise. Traditional methods take the entire image features as visual prompt into the MLLMs, leading to excessive visual tokens that disrupt textual information expression. Thus, recent studies treat image features as visual knowledge, storing them in the feed-forward network for retrieval when needed. These methods, completely removing images from the input, may hinder the activation of image-related knowledge. Besides, current visual knowledge focuses on fine-grained details but overlooks the hierarchical process of visual perception. As described in feature integration theory, global structure is first processed before details are integrated. Ignoring this process may lead to a fragmented visual understanding, making it difficult to capture high-level semantic relationships. To overcome these issues, we propose a novel image utilization mechanism in MLLMs. We leverage a compression-based attention mechanism to generate the compressed visual prompt, which not only mitigates the interference of excessively long visual prompts but also preserves crucial visual information necessary for activating knowledge in the MLLM. Furthermore, we extract hierarchical visual features as visual knowledge using wavelet transforms, allowing the model to capture both global structures and fine-grained details. Experiments show that our method achieves state-of-the-art performance.

PDF Details DOI

NeurIPS Conference 2025 Conference Paper

Rethinking Residual Distribution in Locate-then-Edit Model Editing

Xiaopeng Li
Shangwen Wang
Shasha Li
Shezheng Song
Bin Ji
Ma Jun
Jie Yu

Model editing enables targeted updates to the knowledge of large language models (LLMs) with minimal retraining. Among existing approaches, locate-then-edit methods constitute a prominent paradigm: they first identify critical layers, then compute residuals at the final critical layer based on the target edit, and finally apply least-squares-based multi-layer updates via $\textbf{residual distribution}$. While empirically effective, we identify a counterintuitive failure mode: residual distribution, a core mechanism in these methods, introduces weight shift errors that undermine editing precision. Through theoretical and empirical analysis, we show that such errors increase with the distribution distance, batch size, and edit sequence length, ultimately leading to inaccurate or suboptimal edits. To address this, we propose the $\textbf{B}$oundary $\textbf{L}$ayer $\textbf{U}$pdat$\textbf{E (BLUE)}$ strategy to enhance locate-then-edit methods. Sequential batch editing experiments on three LLMs and two datasets demonstrate that BLUE not only delivers an average performance improvement of 35. 59\%, significantly advancing the state of the art in model editing, but also enhances the preservation of LLMs' general capabilities. Our code is available at https: //github. com/xpq-tech/BLUE.

PDF Details

AAAI Conference 2025 Conference Paper

SWEA: Updating Factual Knowledge in Large Language Models via Subject Word Embedding Altering

Xiaopeng Li
Shasha Li
Shezheng Song
Huijun Liu
Bin Ji
Xi Wang
Jun Ma
Jie Yu

The general capabilities of large language models (LLMs) make them the infrastructure for various AI applications, but updating their inner knowledge requires significant resources. Recent model editing is a promising technique for efficiently updating a small amount of knowledge of LLMs and has attracted much attention. In particular, local editing methods, which directly update model parameters, are proven suitable for updating small amounts of knowledge. Local editing methods update weights by computing least squares closed-form solutions and identify edited knowledge by vector-level matching in inference, which achieve promising results. However, these methods still require a lot of time and resources to complete the computation. Moreover, vector-level matching lacks reliability, and such updates disrupt the original organization of the model's parameters. To address these issues, we propose a detachable and expandable Subject Word Embedding Altering (SWEA) framework, which finds the editing embeddings through token-level matching and adds them to the subject word embeddings in Transformer input. To get these editing embeddings, we propose optimizing then suppressing fusion method, which first optimizes learnable embedding vectors for the editing target and then suppresses the Knowledge Embedding Dimensions (KEDs) to obtain final editing embeddings. We thus propose SWEAOS method for editing factual knowledge in LLMs. We demonstrate the overall state-of-the-art (SOTA) performance of SWEAOS on the CounterFact and zsRE datasets. To further validate the reasoning ability of SWEAOS in editing knowledge, we evaluate it on the more complex RippleEdits benchmark. The results demonstrate that SWEAOS possesses SOTA reasoning ability.

PDF Details DOI

AAAI Conference 2025 Conference Paper

Towards Verifiable Text Generation with Generative Agent

Bin Ji
Huijun Liu
Mingzhe Du
Shasha Li
Xiaodong Liu
Jun Ma
Jie Yu
See-Kiong Ng

Text generation with citations makes it easy to verify the factuality of Large Language Models’ (LLMs) generations. Existing one-step generation studies expose distinct shortages in answer refinement and in-context demonstration matching. In light of these challenges, we propose R2-MGA, a Retrieval and Reflection Memory-augmented Generative Agent. Specifically, it first retrieves the memory bank to obtain the best-matched memory snippet, then reflects the retrieved snippet as a reasoning rationale, next combines the snippet and the rationale as the best-matched in-context demonstration. Additionally, it is capable of in-depth answer refinement with two specifically designed modules. We evaluate R2-MGA across five LLMs on the ALCE benchmark. The results reveal R2-MGA’ exceptional capabilities in text generation with citations. In particular, compared to the selected baselines, it delivers up to +58.8% and +154.7% relative performance gains on answer correctness and citation quality, respectively. Extensive analyses strongly support the motivations of R2-MGA.

PDF Details DOI

AAAI Conference 2024 Conference Paper

A Dual-Way Enhanced Framework from Text Matching Point of View for Multimodal Entity Linking

Shezheng Song
Shan Zhao
Chengyu Wang
Tianwei Yan
Shasha Li
Xiaoguang Mao
Meng Wang

Multimodal Entity Linking (MEL) aims at linking ambiguous mentions with multimodal information to entity in Knowledge Graph (KG) such as Wikipedia, which plays a key role in many applications. However, existing methods suffer from shortcomings, including modality impurity such as noise in raw image and ambiguous textual entity representation, which puts obstacles to MEL. We formulate multimodal entity linking as a neural text matching problem where each multimodal information (text and image) is treated as a query, and the model learns the mapping from each query to the relevant entity from candidate entities. This paper introduces a dual-way enhanced (DWE) framework for MEL: (1) our model refines queries with multimodal data and addresses semantic gaps using cross-modal enhancers between text and image information. Besides, DWE innovatively leverages fine-grained image attributes, including facial characteristic and scene feature, to enhance and refine visual features. (2)By using Wikipedia descriptions, DWE enriches entity semantics and obtains more comprehensive textual representation, which reduces between textual representation and the entities in KG. Extensive experiments on three public benchmarks demonstrate that our method achieves state-of-the-art (SOTA) performance, indicating the superiority of our model. The code is released on https://github.com/season1blue/DWE.

PDF Details DOI

AAAI Conference 2024 Conference Paper

PMET: Precise Model Editing in a Transformer

Xiaopeng Li
Shasha Li
Shezheng Song
Jing Yang
Jun Ma
Jie Yu

Model editing techniques modify a minor proportion of knowledge in Large Language Models (LLMs) at a relatively low cost, which have demonstrated notable success. Existing methods assume Transformer Layer (TL) hidden states are values of key-value memories of the Feed-Forward Network (FFN). They usually optimize the TL hidden states to memorize target knowledge and use it to update the weights of the FFN in LLMs. However, the information flow of TL hidden states comes from three parts: Multi-Head Self-Attention (MHSA), FFN, and residual connections. Existing methods neglect the fact that the TL hidden states contains information not specifically required for FFN. Consequently, the performance of model editing decreases. To achieve more precise model editing, we analyze hidden states of MHSA and FFN, finding that MHSA encodes certain general knowledge extraction patterns. This implies that MHSA weights do not require updating when new knowledge is introduced. Based on above findings, we introduce PMET, which simultaneously optimizes Transformer Component (TC, namely MHSA and FFN) hidden states, while only using the optimized TC hidden states of FFN to precisely update FFN weights. Our experiments demonstrate that PMET exhibits state-of-the-art performance on both the \textsc{counterfact} and zsRE datasets. Our ablation experiments substantiate the effectiveness of our enhancements, further reinforcing the finding that the MHSA encodes certain general knowledge extraction patterns and indicating its storage of a small amount of factual knowledge. Our code is available at \url{https://github.com/xpq-tech/PMET}.

PDF Details DOI

AAAI Conference 2022 Conference Paper

Context-Aware Transfer Attacks for Object Detection

Zikui Cai
Xinxin Xie
Shasha Li
Mingjun Yin
Chengyu Song
Srikanth V. Krishnamurthy
Amit K. Roy-Chowdhury
M. Salman Asif

Blackbox transfer attacks for image classifiers have been extensively studied in recent years. In contrast, little progress has been made on transfer attacks for object detectors. Object detectors take a holistic view of the image and the detection of one object (or lack thereof) often depends on other objects in the scene. This makes such detectors inherently context-aware and adversarial attacks in this space are more challenging than those targeting image classifiers. In this paper, we present a new approach to generate contextaware attacks for object detectors. We show that by using cooccurrence of objects and their relative locations and sizes as context information, we can successfully generate targeted mis-categorization attacks that achieve higher transfer success rates on blackbox object detectors than the state-of-theart. We test our approach on a variety of object detectors with images from PASCAL VOC and MS COCO datasets and demonstrate up to 20 percentage points improvement in performance compared to the other state-of-the-art methods.

PDF Details

TCS Journal 2022 Journal Article

Smallest number of vertices in a 2-arc-strong digraph without good pairs

Ran Gu
Gregory Gutin
Shasha Li
Yongtang Shi
Zhenyu Taoqiu

Branchings play an important role in digraph theory and algorithms. In particular, a chapter in the monograph of Bang-Jensen and Gutin, Digraphs: Theory, Algorithms and Application, Ed. 2, 2009 is wholly devoted to branchings. The well-known Edmonds Branching Theorem provides a characterization for the existence of k arc-disjoint out-branchings rooted at the same vertex. A short proof of the theorem by Lovász (1976) leads to a polynomial-time algorithm for finding such out-branchings. A natural related problem is to characterize digraphs having an out-branching and an in-branching which are arc-disjoint. Such a pair of branchings is called a good pair. Bang-Jensen, Bessy, Havet and Yeo (2022) pointed out that it is NP-complete to decide if a given digraph has a good pair. They also showed that every digraph of independence number at most 2 and arc-connectivity at least 2 has a good pair, which settled a conjecture of Thomassen for digraphs of independence number 2. Then they asked for the smallest number n n g p of vertices in a 2-arc-strong digraph which has no good pair. They proved that 7 ≤ n n g p ≤ 10. In this paper, we prove that n n g p = 10, which solves the open problem.

Details DOI

NeurIPS Conference 2021 Conference Paper

Adversarial Attacks on Black Box Video Classifiers: Leveraging the Power of Geometric Transformations

Shasha Li
Abhishek Aich
Shitong Zhu
Salman Asif
Chengyu Song
Amit Roy-Chowdhury
Srikanth Krishnamurthy

When compared to the image classification models, black-box adversarial attacks against video classification models have been largely understudied. This could be possible because, with video, the temporal dimension poses significant additional challenges in gradient estimation. Query-efficient black-box attacks rely on effectively estimated gradients towards maximizing the probability of misclassifying the target video. In this work, we demonstrate that such effective gradients can be searched for by parameterizing the temporal structure of the search space with geometric transformations. Specifically, we design a novel iterative algorithm GEOmetric TRAnsformed Perturbations (GEO-TRAP), for attacking video classification models. GEO-TRAP employs standard geometric transformation operations to reduce the search space for effective gradients into searching for a small group of parameters that define these operations. This group of parameters describes the geometric progression of gradients, resulting in a reduced and structured search space. Our algorithm inherently leads to successful perturbations with surprisingly few queries. For example, adversarial examples generated from GEO-TRAP have better attack success rates with ~73. 55% fewer queries compared to the state-of-the-art method for video adversarial attacks on the widely used Jester dataset. Overall, our algorithm exposes vulnerabilities of diverse video classification models and achieves new state-of-the-art results under black-box settings on two large datasets.

PDF Details

YNICL Journal 2019 Journal Article

Visual network alterations in brain functional connectivity in chronic low back pain: A resting state functional connectivity and machine learning study

Wei Shen
Yiheng Tu
Randy L. Gollub
Ana Ortiz
Vitaly Napadow
Siyi Yu
Georgia Wilson
Joel Park

Chronic low back pain (cLBP) is associated with widespread functional and structural changes in the brain. This study aims to investigate the resting state functional connectivity (rsFC) changes of visual networks in cLBP patients and the feasibility of distinguishing cLBP patients from healthy controls using machine learning methods. cLBP (n = 90) and control individuals (n = 74) were enrolled and underwent resting-state BOLD fMRI scans. Primary, dorsal, and ventral visual networks derived from independent component analysis were used as regions of interest to compare resting state functional connectivity changes between the cLBP patients and healthy controls. We then applied a support vector machine classifier to distinguish the cLBP patients and control individuals. These results were further verified in a new cohort of subjects. We found that the functional connectivity between the primary visual network and the somatosensory/motor areas were significantly enhanced in cLBP patients. The rsFC between the primary visual network and S1 was negatively associated with duration of cLBP. In addition, we found that the rsFC of the visual network could achieve a classification accuracy of 79.3% in distinguishing cLBP patients from HCs, and these results were further validated in an independent cohort of subjects (accuracy = 66.7%). Our results demonstrate significant changes in the rsFC of the visual networks in cLBP patients. We speculate these alterations may represent an adaptation/self-adjustment mechanism and cross-model interaction between the visual, somatosensory, motor, attention, and salient networks in response to cLBP. Elucidating the role of the visual networks in cLBP may shed light on the pathophysiology and development of the disorder.

Details DOI