arXiv subjects

Yuhao Yi

Publications and source records attributed to Yuhao Yi.

At least 19 recordsLinked to original sources

arXiv2026

CorePath: A Breast-Specialized Pathology Foundation Model for Core Needle Biopsy Diagnosis and Risk-Controlled Report Generation

Breast core needle biopsy (CNB) is central to breast cancer diagnosis yet remains challenging because limited tissue sampling, lesion heterogeneity, and subtle morphologic overlap can obscure subtype distinctions. We developed CorePath, a breast-specialized multimodal pathology foundation model fine-tuned from PRISM using 7901 paired CNB whole-slide images and diagnostic reports from two centers. Evaluated across six CNB cohorts and two public breast pathology benchmarks without task-specific retraining, CorePath consistently outperformed PRISM across cancer detection, invasion assessment, and histological subtyping. It achieved weighted area under the receiver operating characteristic curves (AUCs) of 0.9526-0.9735 for five-class CNB histological subtyping across private centers. On public benchmarks, CorePath outperformed leading pathology foundation models, achieving the highest weighted AUCs of 0.7780 for BCNB invasive carcinoma subtyping, 0.8178 for BRACS lesion stratification, and 0.8252 for BRACS fine-grained classification. In report generation, CorePath reduced the overall non-breast hallucinations from 30.1% to 2.8%, demonstrating improved domain fidelity after breast-specific adaptation. CorePath-CRG further combined conformal subtype-confidence gating with Learn-Then-Test risk control to enable selective report release, subtype-level fallback, and deferral. CorePath-CRG achieved zero non-breast hallucinations among released outputs and showed the strongest overall performance in pathologist-validated LLM-based Evaluation Scores and quantitative report-generation metrics across most centers. These results demonstrate that domain-specialized foundation models with statistical risk control offer a promising approach for accurate breast CNB diagnosis and reliable report generation.

cs.CV↗

arXiv2026

Do Pathology Vision-Language Models Truly See Pathology?

Pathology vision-language models (VLMs) have recently progressed rapidly and are commonly evaluated by answer accuracy on pathology VQA benchmarks. However, we dig into current evaluations and identify three overlooked issues: 1) Visual evidence is not always necessary. For instance, Gemini-3-Pro achieves 53.5% average accuracy across 5 VQA benchmarks without any visual input. 2) Domain training can improve accuracy without proportional gains in visual binding. Compared with Qwen2.5-VL-7B, Patho-R1-7B exhibits a 5.8-point lower multimodal gain and a 3.7-point lower attention IoU. 3) Entity-level attention is diffuse and weakly query-specific. On PathVG, attention maps remain highly correlated across different entity queries. These issues can lead to substantial misjudgments of pathology VLMs' actual multimodal capabilities. To this end, we present PathBind, a benchmark comprising 2,600 samples: PathBind-VQA with 1,500 questions across six dimensions, PathBind-PTA with 600 questions from a private pathology teaching atlas, and PathBind-Grounding with 500 expert-curated region-level samples. Each component undergoes task-specific automated filtering and expert review to reduce textual shortcuts and improve entity-region correspondence. We evaluate 18 representative VLMs on VQA samples of PathBind and five existing pathology VQA benchmarks, and further evaluate 10 VLMs on PathBind-Grounding and PathVG. Results show that current pathology VLMs still exhibit a substantial gap between answer-side performance and visual-semantic binding.

cs.CV↗

arXiv2026

PathoSage: Towards Multi-Source Evidence Adjudication in Pathology via Experience-Aware Agentic Workflow

Recent advances in Multimodal Large Language Models (MLLMs) and agent workflows have shown strong promise for computational pathology, yet reliable patch-level reasoning remains challenging. End-to-end pathology MLLMs often hallucinate morphological features, while recent agentic systems usually merge tool outputs and retrieved knowledge into a shared context, making decisions vulnerable to conflicting evidence and context contamination. We propose PathoSage, a three-stage framework that explicitly separates knowledge retrieval, evidence collection, and evidence adjudication for patch-level pathology multimodal reasoning. Its core component, Structured Evidence Deliberation, independently evaluates heterogeneous evidence from tools, performs conflict analysis, and generates the final judgment in a fresh context to reduce anchoring bias. We further introduce a training-free Beta-Bernoulli experience system with continuous credit assignment to model long-term tool reliability and construct similarity-weighted priors for future tool use. Experiments show that PathoSage effectively mitigates VQA hallucinations and classifier disagreement, outperforming strong pathology MLLM and agentic baselines. Our results highlight explicit evidence adjudication and reliability-aware tool modeling as key ingredients for robust pathology agents.

cs.AI↗

arXiv2026

Accelerating Byzantine-Robust Distributed Learning with Compressed Communication via Double Momentum and Variance Reduction

In collaborative and distributed learning, Byzantine robustness reflects a major facet of optimization algorithms. Such distributed algorithms are often accompanied by transmitting a large number of parameters, so communication compression is essential for an effective solution. In this paper, we propose Byz-DM21, a novel Byzantine-robust and communication-efficient stochastic distributed learning algorithm. Our key innovation is a novel gradient estimator based on a double-momentum mechanism, integrating recent advancements in error feedback techniques. Using this estimator, we design both standard and accelerated algorithms that eliminate the need for large batch sizes while maintaining robustness against Byzantine workers. We prove that the Byz-DM21 algorithm has a smaller neighborhood size and converges to $\varepsilon$-stationary points in $\mathcal{O}(\varepsilon^{-4})$ iterations. To further enhance efficiency, we introduce a distributed variant called Byz-VR-DM21, which incorporates local variance reduction at each node to progressively eliminate variance from random approximations. We show that Byz-VR-DM21 provably converges to $\varepsilon$-stationary points in $\mathcal{O}(\varepsilon^{-3 })$ iterations. Additionally, we extend our results to the case where the functions satisfy the Polyak-{\L}ojasiewicz condition. Finally, numerical experiments demonstrate the effectiveness of the proposed method.

cs.LG↗

arXiv2025

Patho-AgenticRAG: Towards Multimodal Agentic Retrieval-Augmented Generation for Pathology VLMs via Reinforcement Learning

Although Vision Language Models (VLMs) have shown strong generalization in medical imaging, pathology presents unique challenges due to ultra-high resolution, complex tissue structures, and nuanced clinical semantics. These factors make pathology VLMs prone to hallucinations, i.e., generating outputs inconsistent with visual evidence, which undermines clinical trust. Existing RAG approaches in this domain largely depend on text-based knowledge bases, limiting their ability to leverage diagnostic visual cues. To address this, we propose Patho-AgenticRAG, a multimodal RAG framework with a database built on page-level embeddings from authoritative pathology textbooks. Unlike traditional text-only retrieval systems, it supports joint text-image search, enabling direct retrieval of textbook pages that contain both the queried text and relevant visual cues, thus avoiding the loss of critical image-based information. Patho-AgenticRAG also supports reasoning, task decomposition, and multi-turn search interactions, improving accuracy in complex diagnostic scenarios. Experiments show that Patho-AgenticRAG significantly outperforms existing multimodal models in complex pathology tasks like multiple-choice diagnosis and visual question answering. Our project is available at the Patho-AgenticRAG repository: https://github.com/Wenchuan-Zhang/Patho-AgenticRAG.

cs.CV↗

arXiv2025

Adding links wisely: how an influencer seeks for leadership in opinion dynamics?

This paper investigates the problem of leadership development for an external influencer using the Friedkin-Johnsen (FJ) opinion dynamics model, where the influencer is modeled as a fully stubborn agent and leadership is quantified by social power. The influencer seeks to maximize her social power by strategically adding a limited number of links to regular agents. This optimization problem is shown to be equivalent to maximizing the absorbing probability to the influencer in an augmented Markov chain. The resulting objective function is both monotone and submodular, enabling the use of a greedy algorithm to compute an approximate solution. To handle large-scale networks efficiently, a random walk sampling over the Markov chain is employed to reduce computational complexity. Analytical characterizations of the solution are provided for both low and high stubbornness of regular agents. Specific network topologies are also examined: for complete graphs with rank-one weight matrices, the problem reduces to a hyperbolic 0-1 programmming problem, which is solvable in polynomial time; for symmetric ring graphs with circulant weight matrices and uniform agent stubbornness, the optimal strategy involves selecting agents that are sufficiently dispersed across the network. Numerical simulations are presented for illustration.

eess.SY↗

arXiv2025

Denoising Mutual Knowledge Distillation in Bi-Directional Multiple Instance Learning

Multiple Instance Learning is the predominant method for Whole Slide Image classification in digital pathology, enabling the use of slide-level labels to supervise model training. Although MIL eliminates the tedious fine-grained annotation process for supervised learning, whether it can learn accurate bag- and instance-level classifiers remains a question. To address the issue, instance-level classifiers and instance masks were incorporated to ground the prediction on supporting patches. These methods, while practically improving the performance of MIL methods, may potentially introduce noisy labels. We propose to bridge the gap between commonly used MIL and fully supervised learning by augmenting both the bag- and instance-level learning processes with pseudo-label correction capabilities elicited from weak to strong generalization techniques. The proposed algorithm improves the performance of dual-level MIL algorithms on both bag- and instance-level predictions. Experiments on public pathology datasets showcase the advantage of the proposed methods.

cs.CV↗

arXiv2025

Patho-R1: A Multimodal Reinforcement Learning-Based Pathology Expert Reasoner

Recent advances in vision language models (VLMs) have enabled broad progress in the general medical field. However, pathology still remains a more challenging subdomain, with current pathology specific VLMs exhibiting limitations in both diagnostic accuracy and reasoning plausibility. Such shortcomings are largely attributable to the nature of current pathology datasets, which are primarily composed of image description pairs that lack the depth and structured diagnostic paradigms employed by real world pathologists. In this study, we leverage pathology textbooks and real world pathology experts to construct high-quality, reasoning-oriented datasets. Building on this, we introduce Patho-R1, a multimodal RL-based pathology Reasoner, trained through a three-stage pipeline: (1) continued pretraining on 3.5 million image-text pairs for knowledge infusion; (2) supervised fine-tuning on 500k high-quality Chain-of-Thought samples for reasoning incentivizing; (3) reinforcement learning using Group Relative Policy Optimization and Decoupled Clip and Dynamic sAmpling Policy Optimization strategies for multimodal reasoning quality refinement. To further assess the alignment quality of our dataset, we propose Patho-CLIP, trained on the same figure-caption corpus used for continued pretraining. Comprehensive experimental results demonstrate that both Patho-CLIP and Patho-R1 achieve robust performance across a wide range of pathology-related tasks, including zero-shot classification, cross-modal retrieval, Visual Question Answering, and Multiple Choice Question. Our project is available at the Patho-R1 repository: https://github.com/Wenchuan-Zhang/Patho-R1.

cs.CV↗

arXiv2024

Byzantine-Robust and Communication-Efficient Distributed Learning via Compressed Momentum Filtering

Distributed learning has become the standard approach for training large-scale machine learning models across private data silos. While distributed learning enhances privacy preservation and training efficiency, it faces critical challenges related to Byzantine robustness and communication reduction. Existing Byzantine-robust and communication-efficient methods rely on full gradient information either at every iteration or at certain iterations with a probability, and they only converge to an unnecessarily large neighborhood around the solution. Motivated by these issues, we propose a novel Byzantine-robust and communication-efficient stochastic distributed learning method that imposes no requirements on batch size and converges to a smaller neighborhood around the optimal solution than all existing methods, aligning with the theoretical lower bound. Our key innovation is leveraging Polyak Momentum to mitigate the noise caused by both biased compressors and stochastic gradients, thus defending against Byzantine workers under information compression. We provide proof of tight complexity bounds for our algorithm in the context of non-convex smooth loss functions, demonstrating that these bounds match the lower bounds in Byzantine-free scenarios. Finally, we validate the practical significance of our algorithm through an extensive series of experiments, benchmarking its performance on both binary classification and image classification tasks.

cs.LG↗

arXiv2024

LiD-FL: Towards List-Decodable Federated Learning

Federated learning is often used in environments with many unverified participants. Therefore, federated learning under adversarial attacks receives significant attention. This paper proposes an algorithmic framework for list-decodable federated learning, where a central server maintains a list of models, with at least one guaranteed to perform well. The framework has no strict restriction on the fraction of honest workers, extending the applicability of Byzantine federated learning to the scenario with more than half adversaries. Under proper assumptions on the loss function, we prove a convergence theorem for our method. Experimental results, including image classification tasks with both convex and non-convex losses, demonstrate that the proposed algorithm can withstand the malicious majority under various attacks.

cs.LG↗

arXiv2023

Near-Optimal Resilient Aggregation Rules for Distributed Learning Using 1-Center and 1-Mean Clustering with Outliers

Byzantine machine learning has garnered considerable attention in light of the unpredictable faults that can occur in large-scale distributed learning systems. The key to secure resilience against Byzantine machines in distributed learning is resilient aggregation mechanisms. Although abundant resilient aggregation rules have been proposed, they are designed in ad-hoc manners, imposing extra barriers on comparing, analyzing, and improving the rules across performance criteria. This paper studies near-optimal aggregation rules using clustering in the presence of outliers. Our outlier-robust clustering approach utilizes geometric properties of the update vectors provided by workers. Our analysis show that constant approximations to the 1-center and 1-mean clustering problems with outliers provide near-optimal resilient aggregators for metric-based criteria, which have been proven to be crucial in the homogeneous and heterogeneous cases respectively. In addition, we discuss two contradicting types of attacks under which no single aggregation rule is guaranteed to improve upon the naive average. Based on the discussion, we propose a two-phase resilient aggregation framework. We run experiments for image classification using a non-convex loss function. The proposed algorithms outperform previously known aggregation rules by a large margin with both homogeneous and heterogeneous data distributions among non-faulty workers. Code and appendix are available at https://github.com/jerry907/AAAI24-RASHB.

cs.LG↗

arXiv2022

Dynamic Curing and Network Design in SIS Epidemic Processes

This paper studies efficient algorithms for dynamic curing policies and the corresponding network design problems to guarantee the fast extinction of epidemic spread in a susceptible-infected-susceptible (SIS) model. We consider a Markov process-based SIS epidemic model. We provide a computationally efficient curing algorithm based on the curing policy proposed by Drakopoulos, Ozdaglar, and Tsitsiklis (2014). Since the corresponding optimization problem is NP-hard, finding optimal policies is intractable for large graphs. We provide approximation guarantees on the curing budget of the proposed dynamic curing algorithm. We also present a curing algorithm fair to demographic groups. When the total infection rate is high, the original curing policy includes a waiting period in which no measure is taken to mitigate the spread until the rate slows down. To avoid the waiting period, we study network design problems to reduce the total infection rate by deleting edges or reducing the weight of edges. Then the curing processes become continuous since the total infection rate is restricted by network design. We provide algorithms with provable guarantees for the considered network design problems. In summary, the proposed curing and network design algorithms together provide an effective and computationally efficient approach that mitigates SIS epidemic spread in networks.

cs.DS↗

arXiv2022

A Sample-Based Algorithm for Approximately Testing $r$-Robustness of a Digraph

One of the intensely studied concepts of network robustness is $r$-robustness, which is a network topology property quantified by an integer $r$. It is required by mean subsequence reduced (MSR) algorithms and their variants to achieve resilient consensus. However, determining $r$-robustness is intractable for large networks. In this paper, we propose a sample-based algorithm to approximately test $r$-robustness of a digraph with $n$ vertices and $m$ edges. For a digraph with a moderate assumption on the minimum in-degree, and an error parameter $0<\epsilon\leq 1$, the proposed algorithm distinguishes $(r+\epsilon n)$-robust graphs from graphs which are not $r$-robust with probability $(1-\delta)$. Our algorithm runs in $\exp(O((\ln{\frac{1}{\epsilon\delta}})/\epsilon^2))\cdot m$ time. The running time is linear in the number of edges if $\epsilon$ is a constant.

eess.SY↗

arXiv2020

Edge Deletion Algorithms for Minimizing Spread in SIR Epidemic Models

This paper studies algorithmic strategies to effectively reduce the number of infections in susceptible-infected-recovered (SIR) epidemic models. We consider a Markov chain SIR model and its two instantiations in the deterministic SIR (D-SIR) model and the independent cascade SIR (IC-SIR) model. We investigate the problem of minimizing the number of infections by restricting contacts under realistic constraints. Under moderate assumptions on the reproduction number, we prove that the infection numbers are bounded by supermodular functions in the D-SIR model and the IC-SIR model for large classes of random networks. We propose efficient algorithms with approximation guarantees to minimize infections. The theoretical results are illustrated by numerical simulations.

math.OC↗

arXiv2020

Shifting Opinions in a Social Network Through Leader Selection

We study the French-DeGroot opinion dynamics in a social network with two polarizing parties. We consider a network in which the leaders of one party are given, and we pose the problem of selecting the leader set of the opposing party so as to shift the average opinion to a desired value. When each party has only one leader, we express the average opinion in terms of the transition matrix and the stationary distribution of random walks in the network. The analysis shows balance of influence between the two leader nodes. We show that the problem of selecting at most $k$ absolute leaders to shift the average opinion is $\mathbf{NP}$-hard. Then, we reduce the problem to a problem of submodular maximization with a submodular knapsack constraint and an additional cardinality constraint and propose a greedy algorithm with upper bound search to approximate the optimum solution. We also conduct experiments in random networks and real-world networks to show the effectiveness of the algorithm.

cs.SI↗

arXiv2020

Disagreement and Polarization in Two-Party Social Networks

We investigate disagreement and polarization in a social network with two polarizing sources of information. First, we define disagreement and polarization indices in two-party leader-follower models of opinion dynamics. We then give expressions for the indices in terms of a graph Laplacian. The expressions show a relationship between these quantities and the concepts of resistance distance and biharmonic distance. We next study the problem of designing the network so as to minimize disagreement and polarization. We give conditions for optimal disagreement and polarization, and further, we show that a linear combination of disagreement and polarization of the follower nodes is a convex function of the edge weights between followers. We propose algorithms to address some related continuous and discrete optimization problems and also present analytic results for some interesting examples.

cs.SI↗

arXiv2020

Edge corona product as an approach to modeling complex simplical networks

Many graph products have been applied to generate complex networks with striking properties observed in real-world systems. In this paper, we propose a simple generative model for simplicial networks by iteratively using edge corona product. We present a comprehensive analysis of the structural properties of the network model, including degree distribution, diameter, clustering coefficient, as well as distribution of clique sizes, obtaining explicit expressions for these relevant quantities, which agree with the behaviors found in diverse real networks. Moreover, we obtain exact expressions for all the eigenvalues and their associated multiplicities of the normalized Laplacian matrix, based on which we derive explicit formulas for mixing time, mean hitting time and the number of spanning trees. Thus, as previous models generated by other graph products, our model is also an exactly solvable one, whose structural properties can be analytically treated. More interestingly, the expressions for the spectra of our model are also exactly determined, which is sharp contrast to previous models whose spectra can only be given recursively at most. This advantage makes our model a good test-bed and an ideal substrate network for studying dynamical processes, especially those closely related to the spectra of normalized Laplacian matrix, in order to uncover the influences of simplicial structure on these processes.

cs.DM↗

arXiv2020

Diffusion and Consensus in a Weakly Coupled Network of Networks

We study diffusion and consensus dynamics in a Network of Networks model. In this model, there is a collection of sub-networks, connected to one another using a small number of links. We consider a setting where the links between networks have small weights, or are used less frequently than links within each sub-network. Using spectral perturbation theory, we analyze the diffusion rate and convergence rate of the investigated systems. Our analysis shows that the first order approximation of the diffusion and convergence rates is independent of the topologies of the individual graphs; the rates depend only on the number of nodes in each graph and the topology of the connecting edges. The second order analysis shows a relationship between the diffusion and convergence rates and the information centrality of the connecting nodes within each sub-network. We further highlight these theoretical results through numerical examples.

cs.SI↗