Chengwei Pan

Assuming the coefficient matrix is a nonzero singular matrix, we demonstrate that the invertible solutions of the Yang–Baxter-like matrix equation must possess at least two elementary divisors. This establishes a necessary condition for an invertible matrix to satisfy the Yang–Baxter-like matrix equation. Building on this finding, we derive several meaningful corollaries. Additionally, we provide some examples to illustrate our results.

Large Language Models (LLMs) are increasingly deployed in medicine. However, their utility for non-generative clinical prediction is under-evaluated, and they are often assumed to be inferior to specialized models, creating potential for misuse and misunderstanding. To address this, our ClinicRealm benchmark systematically evaluates 15 GPT-style LLMs, 5 BERT-style models, and 11 traditional methods on unstructured clinical notes and structured Electronic Health Records (EHR) across predictive performance, reasoning, fairness, etc. Our findings reveal a significant shift: on clinical notes, leading zero-shot LLMs (e.g., DeepSeek-V3.1-Think, GPT-5) now decisively outperform finetuned BERT models. On structured EHRs, while specialized models excel with ample data, advanced LLMs demonstrate potent zero-shot capabilities, often surpassing conventional models in data-scarce settings. Notably, leading open-source LLMs match or exceed their proprietary counterparts. This provides compelling evidence that modern LLMs are competitive tools for clinical prediction, necessitating a re-evaluation of model selection strategies by health data scientists and developers.

Assuming the coefficient matrix is a nonzero singular matrix, we demonstrate that the invertible solutions of the Yang-Baxter-like matrix equation must possess at least two elementary divisors. This establishes a necessary condition for an invertible matrix to satisfy the Yang-Baxter-like matrix equation. Building on this finding, we derive several meaningful corollaries. Additionally, we provide some examples to illustrate our results.

Multi-Object Tracking (MOT) aims to detect and uniquely identify objects within a video sequence. Traditional methods often focus on processing data between consecutive frames or primarily rely on motion cues to represent object states. However, these approaches often overlook the influence of historical data on the association process. To address this gap, this paper introduces a Topological Chronology Graph Model (TCGM) designed to bridge the historical and current states of objects. The TCGM consists of two main components: the Topological Chronology Graph Builder (TCGB) and the Spatio-Temporal Feature Integrator (STFI). TCGB encodes historical trajectory data and establishes associative links with newly detected objects to model spatio-temporal topological relationships. Meanwhile, the STFI integrates current and historical data through a spatio-temporal transformer to comprehensively represent the state of each object. Our algorithm has been rigorously evaluated on the MOT17, MOT20, DanceTrack and KITTI benchmarks, achieving competitive results.

Multi-object tracking (MOT) plays a critical role in applications such as autonomous driving and surveillance. Camera-based approaches offer rich texture features for object association, while LiDAR-based methods provide accurate geometric information for spatial reasoning. Although each modality addresses different challenges, their intrinsic discrepancies hinder effective cross-modal fusion and unified representation learning. To overcome these limitations, we propose IMH-MOT, an interactive multi-hierarchical MOT framework comprising three key modules. The Multi-modality Alignment Module (MMAM) enhances spatial representations by sampling and clustering instance-level point clouds. From different modalities are motion cues integrated by the Multi-modality Motion Estimation Module (MMEM) to build a unified motion model. To mitigate the impact of occlusion on single-frame appearance features, the Long-term Appearance Module (LAM) captures temporal appearance consistency by constructing a long-term appearance embedding. Guided by modality-aware cues from MMAM, MMEM generates reliable spatial representations, while LAM encodes robust long-term appearance features. These components are jointly integrated through a Multi-hierarchical Data Association (MHDA) strategy, enabling stable and accurate tracking. Extensive experiments on the KITTI MOT benchmark demonstrate the effectiveness of our framework, achieving 80.90% HOTA, 89.73% MOTA, and 470 IDSW, outperforming state-of-the-art methods in both standard and challenging scenarios.