Shuo Yan

Wentao Ge, Shunian Chen, Guiming Hardy Chen et al.

Multimodal large language models (MLLMs) have broadened the scope of AI applications. Existing automatic evaluation methodologies for MLLMs are mainly limited in evaluating queries without considering user experiences, inadequately addressing the nuances of creative and associative multimodal tasks. However, the open-ended and subjective nature of such tasks poses a significant challenge to the evaluation methodology, where it is difficult to define the ground-truth answers for them. To this end, in our paper, we propose a new evaluation paradigm for MLLMs, which is evaluating MLLMs with per-sample criteria using potent MLLM as the judge. To validate the feasibility and effectiveness of this paradigm, we design a benchmark, dubbed MLLM-Bench, by curating the evaluation samples across six comprehensive cognitive levels. We benchmark 21 popular MLLMs in a pairwise-comparison fashion, showing diverse performance across models. Moreover, the validity of our benchmark manifests itself in reaching 88.02% agreement with human evaluation. We contend that the proposed paradigm explores the potential of MLLMs as effective evaluation tools with the help of per-sample criteria. See online leaderboard at \url{https://mllm-bench.llmzoo.com}.

Hardy Chen, Nancy Lau, Haoqin Tu et al.

Frontier coding agents are increasingly used in workflows where users supervise progress primarily through repeated improvement of a public score, namely the reported score on a public evaluation file with labels in the workspace, rather than through direct inspection of the agent's intermediate outputs. We study whether multi-round user pressure to improve that score induces public score exploitation: behavior that raises the public score through shortcuts without improving hidden private evaluation. We begin with a preliminary single-script tabular classification task, where GPT-5.4 and Claude Opus 4.6 both exploit label information within 10 rounds of user-agent interaction. We then build AgentPressureBench, a 34-task machine-learning repository benchmark spanning three input modalities, and collect 1326 multi-round trajectories from 13 coding agents. On our benchmark, we observe 403 exploitative runs, spanning across all tasks. We also find that stronger models have higher exploitation rates, supported by a significant Spearman rank correlation of 0.77. Our ablation experiments show that higher user pressure leads to earlier exploitation, reducing the average first exploit round by 15.6 rounds (i.e., 19.67 to 4.08). As a mitigation, adding explicit anti-exploit wordings in prompt mostly eliminates exploitation (100% to 8.3%). We hope that our work can bring attention to more careful use of coding agents workflow, and developing more robust coding agents under user pressure. Our project page is at https://ucsc-vlaa.github.io/AgentPressureBench .

Wanying Tan, Shuo Yan, Dazhi Huang et al.

Interpretability in object detection provides crucial confidence support for clinical auxiliary diagnosis. However, in tiny bacteria detection, traditional explanation methods often suffer from blurred foreground boundaries and diffuse feature attribution due to the extreme sparsity of target morphological features and severe interference from complex backgrounds. Such limitations hinder the provision of logically coherent morphological evidence. To bridge this gap, we propose a novel eXplainable AI (XAI) framework, SAM-Sode. The framework innovatively transforms initial feature attribution maps into geometry-aware prompts, leveraging the prior knowledge of the foundation model (SAM3) to achieve spatial refinement and morphological reconstruction of the explanatory mappings. Furthermore, we introduce a dual-constraint mechanism based on physical significance and geometric alignment to perform instance-level denoising, generating coherent explanations that better align with human expert intuition. Experimental results on our self-constructed bacteria dataset with complex circuit backgrounds (containing 2,524 images) and other public datasets demonstrate that the proposed method effectively suppresses background redundancy and significantly enhances the decision-making transparency of tiny object detection.

Kaiyu He, Mian Zhang, Shuo Yan et al.

While large language models (LLMs) have been thoroughly evaluated for deductive and inductive reasoning, their proficiency in holistic rule learning in interactive environments remains less explored. We introduce RULEARN, a novel benchmark to assess the rule-learning abilities of LLM agents in interactive settings. In RULEARN, agents strategically interact with simulated environments to gather observations, discern patterns, and solve complex problems. To enhance the rule-learning capabilities for LLM agents, we propose IDEA, a novel reasoning framework that integrates the process of Induction, Deduction, and Abduction. The IDEA agent generates initial hypotheses from limited observations through abduction, devises plans to validate these hypotheses or leverages them to solve problems via deduction, and refines previous hypotheses through induction, dynamically establishing and applying rules that mimic human rule-learning behaviors. Our evaluation of the IDEA framework, which involves five representative LLMs, demonstrates significant improvements over the baseline. Furthermore, our study with human participants reveals notable discrepancies in rule-learning behaviors between humans and LLMs. We believe our benchmark will serve as a valuable and challenging resource, and IDEA will provide crucial insights for the development of LLM agents capable of human-like rule learning in real-world scenarios. Our code and data is publicly available.

Xin Chen, Feng Jiang, Yiqian Zhang et al.

Reasoning-oriented Large Language Models (LLMs) have achieved remarkable progress with Chain-of-Thought (CoT) prompting, yet they remain fundamentally limited by a \emph{blind self-thinking} paradigm: performing extensive internal reasoning even when critical information is missing or ambiguous. We propose Proactive Interactive Reasoning (PIR), a new reasoning paradigm that transforms LLMs from passive solvers into proactive inquirers that interleave reasoning with clarification. Unlike existing search- or tool-based frameworks that primarily address knowledge uncertainty by querying external environments, PIR targets premise- and intent-level uncertainty through direct interaction with the user. PIR is implemented via two core components: (1) an uncertainty-aware supervised fine-tuning procedure that equips models with interactive reasoning capability, and (2) a user-simulator-based policy optimization framework driven by a composite reward that aligns model behavior with user intent. Extensive experiments on mathematical reasoning, code generation, and document editing demonstrate that PIR consistently outperforms strong baselines, achieving up to 32.70\% higher accuracy, 22.90\% higher pass rate, and 41.36 BLEU improvement, while reducing nearly half of the reasoning computation and unnecessary interaction turns. Further reliability evaluations on factual knowledge, question answering, and missing-premise scenarios confirm the strong generalization and robustness of PIR. Model and code are publicly available at: \href{https://github.com/SUAT-AIRI/Proactive-Interactive-R1}

Yuliang Yan, Haochun Tang, Shuo Yan et al.

Large language models (LLMs) are considered valuable Intellectual Properties (IP) for legitimate owners due to the enormous computational cost of training. It is crucial to protect the IP of LLMs from malicious stealing or unauthorized deployment. Despite existing efforts in watermarking and fingerprinting LLMs, these methods either impact the text generation process or are limited in white-box access to the suspect model, making them impractical. Hence, we propose DuFFin, a novel $\textbf{Du}$al-Level $\textbf{Fin}$gerprinting $\textbf{F}$ramework for black-box setting ownership verification. DuFFin extracts the trigger pattern and the knowledge-level fingerprints to identify the source of a suspect model. We conduct experiments on a variety of models collected from the open-source website, including four popular base models as protected LLMs and their fine-tuning, quantization, and safety alignment versions, which are released by large companies, start-ups, and individual users. Results show that our method can accurately verify the copyright of the base protected LLM on their model variants, achieving the IP-ROC metric greater than 0.95. Our code is available at https://github.com/yuliangyan0807/llm-fingerprint.

Chenyu Wang, Shuo Yan, Yixuan Chen et al.

Video generation using diffusion-based models is constrained by high computational costs due to the frame-wise iterative diffusion process. This work presents a Diffusion Reuse MOtion (Dr. Mo) network to accelerate latent video generation. Our key discovery is that coarse-grained noises in earlier denoising steps have demonstrated high motion consistency across consecutive video frames. Following this observation, Dr. Mo propagates those coarse-grained noises onto the next frame by incorporating carefully designed, lightweight inter-frame motions, eliminating massive computational redundancy in frame-wise diffusion models. The more sensitive and fine-grained noises are still acquired via later denoising steps, which can be essential to retain visual qualities. As such, deciding which intermediate steps should switch from motion-based propagations to denoising can be a crucial problem and a key tradeoff between efficiency and quality. Dr. Mo employs a meta-network named Denoising Step Selector (DSS) to dynamically determine desirable intermediate steps across video frames. Extensive evaluations on video generation and editing tasks have shown that Dr. Mo can substantially accelerate diffusion models in video tasks with improved visual qualities.

Shuo Yan, Yuliang Yan, Bin Ma et al.

Recently, extensive deep learning architectures and pretraining strategies have been explored to support downstream protein applications. Additionally, domain-specific models incorporating biological knowledge have been developed to enhance performance in specialized tasks. In this work, we introduce $\textbf{Protap}$, a comprehensive benchmark that systematically compares backbone architectures, pretraining strategies, and domain-specific models across diverse and realistic downstream protein applications. Specifically, Protap covers five applications: three general tasks and two novel specialized tasks, i.e., enzyme-catalyzed protein cleavage site prediction and targeted protein degradation, which are industrially relevant yet missing from existing benchmarks. For each application, Protap compares various domain-specific models and general architectures under multiple pretraining settings. Our empirical studies imply that: (i) Though large-scale pretraining encoders achieve great results, they often underperform supervised encoders trained on small downstream training sets. (ii) Incorporating structural information during downstream fine-tuning can match or even outperform protein language models pretrained on large-scale sequence corpora. (iii) Domain-specific biological priors can enhance performance on specialized downstream tasks. Code and datasets are publicly available at https://github.com/Trust-App-AI-Lab/protap.

Chenao Li, Shuo Yan, Enyan Dai

Enzyme-catalyzed protein cleavage is essential for many biological functions. Accurate prediction of cleavage sites can facilitate various applications such as drug development, enzyme design, and a deeper understanding of biological mechanisms. However, most existing models are restricted to an individual enzyme, which neglects shared knowledge of enzymes and fails generalize to novel enzymes. Thus, we introduce a unified protein cleavage site predictor named UniZyme, which can generalize across diverse enzymes. To enhance the enzyme encoding for the protein cleavage site prediction, UniZyme employs a novel biochemically-informed model architecture along with active-site knowledge of proteolytic enzymes. Extensive experiments demonstrate that UniZyme achieves high accuracy in predicting cleavage sites across a range of proteolytic enzymes, including unseen enzymes. The code is available in https://anonymous.4open.science/r/UniZyme-4A67.

Shuo Yan, Xiaolin Wen, Shaolun Ruan et al.

Large Language Model (LLM)-based agentic systems have shown growing promise in tackling complex, multi-step tasks through autonomous planning, reasoning, and interaction with external environments. However, the stochastic nature of LLM generation introduces intrinsic behavioral inconsistency: the same agent may succeed in one execution but fail in another under identical inputs. Diagnosing such inconsistencies remains a major challenge for developers, as agent execution logs are often lengthy, unstructured, and difficult to compare across runs. Existing debugging and evaluation tools primarily focus on inspecting single executions, offering limited support for understanding how and why agent behaviors diverge across repeated runs. To address this challenge, we introduce InconLens, a visual analytics system designed to support interactive diagnosis of LLM-based agentic systems with a particular focus on cross-run behavioral analysis. InconLens introduces information nodes as an intermediate abstraction that captures canonical informational milestones shared across executions, enabling semantic alignment and inspection of agent reasoning trajectories across multiple runs. We demonstrate the effectiveness of InconLens through a detailed case study and further validate its usability and analytical value via expert interviews. Our results show that InconLens enables developers to more efficiently identify divergence points, uncover latent failure modes, and gain actionable insights into improving the reliability and stability of agentic systems.

Shuo Yan, Ruochen Li, Ziming Luo et al. · microsoft-research

Large language model (LLM) agents have demonstrated remarkable potential in advancing scientific discovery. However, their capability in the fundamental yet crucial task of reproducing code from research papers, especially in the NLP domain, remains underexplored. This task includes unique complex reasoning challenges in the intellectual synthesis of abstract concepts and the comprehension of code repositories with interdependent files. Motivated by this gap, we present LMR-BENCH, a benchmark designed to systematically evaluate the capability of LLM agents on code reproduction from Language Modeling Research. It consists of 28 code reproduction tasks derived from 23 research papers published in top-tier NLP venues over the past five years, spanning nine fundamental categories. Models are provided with a research paper, a code repository containing one or more masked functions, and instructions for implementing these functions. We conduct extensive experiments in standard prompting and LLM agent settings with state-of-the-art LLMs, evaluating the accuracy of unit tests and performing LLM-based evaluation of code correctness. Experimental results reveal that even the most advanced models still exhibit persistent limitations in scientific reasoning and code synthesis, highlighting critical gaps in LLM agents' ability to autonomously reproduce scientific research

Yu Bai, Zitong Yu, Haowen Tian et al.

We propose Spatial-Aware Correlated Multiple Instance Learning (SAC-MIL) for performing WSI classification. SAC-MIL consists of a positional encoding module to encode position information and a SAC block to perform full instance correlations. The positional encoding module utilizes the instance coordinates within the slide to encode the spatial relationships instead of the instance index in the input WSI sequence. The positional encoding module can also handle the length extrapolation issue where the training and testing sequences have different lengths. The SAC block is an MLP-based method that performs full instance correlation in linear time complexity with respect to the sequence length. Due to the simple structure of MLP, it is easy to deploy since it does not require custom CUDA kernels, compared to Transformer-based methods for WSI classification. SAC-MIL has achieved state-of-the-art performance on the CAMELYON-16, TCGA-LUNG, and TCGA-BRAC datasets. The code will be released upon acceptance.

Bo Zhang, Xu Xinan, Shuo Yan et al.

Recent pseudo-bag augmentation methods for Multiple Instance Learning (MIL)-based Whole Slide Image (WSI) classification sample instances from a limited number of bags, resulting in constrained diversity. To address this issue, we propose Contrastive Cross-Bag Augmentation ($C^2Aug$) to sample instances from all bags with the same class to increase the diversity of pseudo-bags. However, introducing new instances into the pseudo-bag increases the number of critical instances (e.g., tumor instances). This increase results in a reduced occurrence of pseudo-bags containing few critical instances, thereby limiting model performance, particularly on test slides with small tumor areas. To address this, we introduce a bag-level and group-level contrastive learning framework to enhance the discrimination of features with distinct semantic meanings, thereby improving model performance. Experimental results demonstrate that $C^2Aug$ consistently outperforms state-of-the-art approaches across multiple evaluation metrics.

Bo Zhang, Yifan Zhang, Shuo Yan et al.

In light of the spatial domain's limited capacity for modeling global context in 3D medical image segmentation, emerging approaches have begun to incorporate frequency domain representations. However, straightforward feature extraction strategies often overlook the unique properties of frequency domain information, such as conjugate symmetry. They also fail to account for the fundamental differences in data distribution between the spatial and frequency domains, which can ultimately dilute or obscure the complementary strengths that frequency-based representations offer. In this paper, we propose SSFMamba, a Mamba based Symmetry-driven Spatial-Frequency feature fusion network for 3D medical image segmentation. SSFMamba employs a complementary dual-branch architecture that extracts features from both the spatial and frequency domains, and leverages a Mamba block to fuse these heterogeneous features to preserve global context while reinforcing local details. In the frequency domain branch, we harness Mamba's exceptional capability to extract global contextual information in conjunction with the synergistic effect of frequency domain features to further enhance global modeling. Moreover, we design a 3D multi-directional scanning mechanism to strengthen the fusion of local and global cues. Extensive experiments on the BraTS2020 and BraTS2023 datasets demonstrate that our approach consistently outperforms state-of-the-art methods across various evaluation metrics.