arXiv subjects

Kuo Zhang

Publications and source records attributed to Kuo Zhang.

8 recordsLinked to original sources

arXiv2026

A radiographic world model for clinical reasoning and evidence generation

Medical imaging artificial intelligence (AI) is commonly developed as separate mappings from radiographs to diagnostic outputs or from clinical descriptions to generated images, although both arise from the same underlying radiographic state. A world-model formulation instead seeks to learn an internal representation of this state that can support both clinical readout and conditional simulation of radiographic observations. Here we introduce MedDream, a radiographic world model that learns a shared continuous latent state from paired chest radiograph-text observations for diagnostic reasoning and report-conditioned evidence generation. MedDream was pretrained on 2.65 million leakage-controlled chest radiograph-text pairs curated from 4.40 million candidates. Across eight clinical datasets and two independent reader cohorts, MedDream outperformed leading diagnostic and generative comparators. For diagnostic reasoning, MedDream showed strong generalization across disease recognition, label-scarce adaptation, severity assessment, and localization, while MedDream-supported review increased mean resident concordance with independent radiologist consensus from 56.3% to 63.0%. For evidence generation, MedDream produced radiographs that preserved clinically relevant pathology and improved downstream performance on held-out real data, with synthetic augmentation increasing external VinDr-CXR macro-AUROC from 76.4% to 81.4%. More importantly, conditioning generation on prespecified subgroup performance gaps enabled targeted evidence construction, increasing weighted F1 by 3.1 percentage points in Asian patients, whereas matched-volume unguided augmentation decreased it by 2.3 points. These findings establish radiographic world models as a path toward medical AI that learns clinically meaningful internal states for interpreting, simulating, and constructing evidence for clinical use.

cs.AI↗

arXiv2026

MOSAIC: Adaptive Inter-layer Composition for Efficient Heterogeneous Vision-Language Models

Vision-Language Models (VLMs) have achieved success using homogeneous Transformers to process multimedia data. Recent studies show that heterogeneous structures interleaving efficient mechanisms, like linear attention, improve both performance and inference latency over homogeneous designs. However, these efforts rely on handcrafted static mixing patterns, which are sub-optimal and difficult to adapt to specific hardware. To bridge this gap, we propose Multi-Objective Search for Adaptive Inter-layer Composition (MOSAIC), a hardware-aware search method that automatically transforms homogeneous models into optimized heterogeneous architectures. MOSAIC integrates diverse efficiency mechanisms--including linear, sparse, and low-rank operators--into a unified search space. By formulating the selection as a multi-objective Mixed Integer Programming (MIP) problem, our method identifies optimal configurations that maximize downstream performance under strict hardware latency constraints. To mitigate performance degradation from structural transitions, we introduce a two-stage parameter recovery process: global off-policy distillation to stabilize internal representations, followed by a dual-teacher on-policy distillation leveraging a 235B oracle for knowledge expansion and the original 4B teacher for distributional stability. We validate MOSAIC through MOSAIC-4B, derived from Qwen3-VL-4B-Instruct. Results demonstrate that MOSAIC-4B matches the baseline's performance across multiple benchmarks while requiring less than 2% of the original training cost. Furthermore, it substantially improves inference efficiency, achieving 1.76x prefilling and 2.54x decoding speedups.

cs.CV↗

arXiv2026

Orthogonal degree-similarity of edge-deleted strongly regular graphs

Godsil and Sun asked whether, for a strongly regular graph $X$ and any two different edges $e$ and $f$, the edge-deleted graphs $X\setminus e$ and $X\setminus f$ are degree-similar. We give an affirmative answer to the problem of Godsil and Sun. In fact, we prove the stronger statement that if $X$ is a $1$-walk-regular graph, then for any two edges $e$ and $f$ of $X$, the graphs $X\setminus e$ and $X\setminus f$ are orthogonally degree-similar. The proof is based on an edge version of the orthogonal-intertwiner method: the equality of the Gram matrices of the projected endpoint vectors in every eigenspace yields an orthogonal matrix commuting with the adjacency matrix and sending one pair of ordered endpoint vectors to the other.

math.CO↗

arXiv2026

Smith normal forms for coalescences at cospectral vertices

Let $L_\mu(G)=A(G)-\mu D(G)$ be the generalized $\mu$-adjacency matrix of a finite graph $G$. Fan, Xing, Zhang, and Wang constructed pairs of non-degree-similar trees for which the Smith normal forms of the matrices $tI-L_\mu(G)$ over $\mathbb{Q}(\mu)[t]$ coincide, and conjectured that their construction remains valid when the attached rooted path is replaced by an arbitrary rooted tree. We prove this conjecture as a consequence of a more general coalescence theorem: if a finite graph $H$ has two vertices $u$ and $v$ that are cospectral for $L_\mu(H)$, then, for every finite rooted graph $R$ with root $r$, the matrices \[ tI-L_\mu(R(r)\odot H(u)) \quad\text{and}\quad tI-L_\mu(R(r)\odot H(v)) \] have the same Smith normal form over $\mathbb{Q}(\mu)[t]$, where $\odot$ denotes coalescence of rooted graphs. The proof uses an orthogonal intertwiner over a real closed extension field.

math.CO↗

arXiv2025

An Automated Classifier of Harmful Brain Activities for Clinical Usage Based on a Vision-Inspired Pre-trained Framework

Timely identification of harmful brain activities via electroencephalography (EEG) is critical for brain disease diagnosis and treatment, which remains limited application due to inter-rater variability, resource constraints, and poor generalizability of existing artificial intelligence (AI) models. In this study, a convolutional neural network model, VIPEEGNet, was developed and validated using EEGs recorded from Massachusetts General Hospital/Harvard Medical School. The VIPEEGNet was developed and validated using two independent datasets, collected between 2006 and 2020. The development cohort included EEG recordings from 1950 patients, with 106,800 EEG segments annotated by at least one experts (ranging from 1 to 28). The online testing cohort consisted of EEG segments from a subset of an additional 1,532 patients, each annotated by at least 10 experts. For the development cohort (n=1950), the VIPEEGNet achieved high accuracy, with an AUROC for binary classification of seizure, LPD, GPD, LRDA, GRDA, and "other" categories at 0.972 (95% CI, 0.957-0.988), 0.962 (95% CI, 0.954-0.970), 0.972 (95% CI, 0.960-0.984), 0.938 (95% CI, 0.917-0.959), 0.949 (95% CI, 0.941-0.957), and 0.930 (95% CI, 0.926-0.935). For multi classification, the sensitivity of VIPEEGNET for the six categories ranges from 36.8% to 88.2% and the precision ranges from 55.6% to 80.4%, and performance similar to human experts. Notably, the external validation showed Kullback-Leibler Divergence (KLD)of 0.223 and 0.273, ranking top 2 among the existing 2,767 competing algorithms, while we only used 2.8% of the parameters of the first-ranked algorithm.

cs.LG↗

arXiv2023

Flexible Coherent Optical Access: Architectures, Algorithms, and Demonstrations

To cope with the explosive bandwidth demand, significant progress has been made in the ITU-T standardization sector to define a higher-speed passive optical network (PON) with a 50Gb/s line rate. Recently, 50G PON becomes mature gradually, which means it is time to discuss beyond 50G PON. For ensuring an acceptable optical power budget, beyond 50G PON will potentially use coherent technologies, which can simultaneously promote the applications of flexible multiple access such as time/frequency-domain multiple access (TFDMA). In this paper, we will introduce the architectures, algorithms, and demonstrations for TFDMA-based coherent PON. The system architectures based on an ultra-simple coherent transceiver and specific signal spectra are designed to greatly reduce the cost of ONUs. Meanwhile, fast and low-complexity digital signal processing (DSP) algorithms are proposed for dealing with upstream and downstream signals. Based on the architectures and algorithms, we experimentally demonstrate the first real-time TFDMA-based coherent PON, which can support at most 256 end users, and peak line rates of 100Gb/s and 200Gb/s in the upstream and downstream scenarios, respectively. In conclusion, the proposed technologies for the coherent PON make it more possible to be applied in the future beyond 50G PON.

cs.NI↗

arXiv2022

Safe Reinforcement Learning of Dynamic High-Dimensional Robotic Tasks: Navigation, Manipulation, Interaction

Safety is a crucial property of every robotic platform: any control policy should always comply with actuator limits and avoid collisions with the environment and humans. In reinforcement learning, safety is even more fundamental for exploring an environment without causing any damage. While there are many proposed solutions to the safe exploration problem, only a few of them can deal with the complexity of the real world. This paper introduces a new formulation of safe exploration for reinforcement learning of various robotic tasks. Our approach applies to a wide class of robotic platforms and enforces safety even under complex collision constraints learned from data by exploring the tangent space of the constraint manifold. Our proposed approach achieves state-of-the-art performance in simulated high-dimensional and dynamic tasks while avoiding collisions with the environment. We show safe real-world deployment of our learned controller on a TIAGo++ robot, achieving remarkable performance in manipulation and human-robot interaction tasks.

cs.RO↗

arXiv2022

Regularized Deep Signed Distance Fields for Reactive Motion Generation

Autonomous robots should operate in real-world dynamic environments and collaborate with humans in tight spaces. A key component for allowing robots to leave structured lab and manufacturing settings is their ability to evaluate online and real-time collisions with the world around them. Distance-based constraints are fundamental for enabling robots to plan their actions and act safely, protecting both humans and their hardware. However, different applications require different distance resolutions, leading to various heuristic approaches for measuring distance fields w.r.t. obstacles, which are computationally expensive and hinder their application in dynamic obstacle avoidance use-cases. We propose Regularized Deep Signed Distance Fields (ReDSDF), a single neural implicit function that can compute smooth distance fields at any scale, with fine-grained resolution over high-dimensional manifolds and articulated bodies like humans, thanks to our effective data generation and a simple inductive bias during training. We demonstrate the effectiveness of our approach in representative simulated tasks for whole-body control (WBC) and safe Human-Robot Interaction (HRI) in shared workspaces. Finally, we provide proof of concept of a real-world application in a HRI handover task with a mobile manipulator robot.

cs.RO↗