arXiv subjects

Cheng-Kuan Lin

Publications and source records attributed to Cheng-Kuan Lin.

9 recordsLinked to original sources

arXiv2026

DeepTable: Structural Attention Biases and Tree Path Encoding for Hierarchical Table Understanding

Large language models (LLMs) have demonstrated strong performance in table understanding. However, they typically process table content and headers as linearized token sequences. This representation weakens the two-dimensional and hierarchical structural relationships encoded by multi-level row and column headers. Existing parameter-efficient fine-tuning methods incorporate basic row and column information but do not explicitly capture the rich structural dependencies induced by hierarchical table headers. We propose DeepTable, a structure-aware approach for table understanding with LLMs. DeepTable comprises two complementary components. Structural Attention Bias (SAB) introduces learnable biases into the attention logits to explicitly represent whether pairs of table tokens share the same row or column. Tree Path Encoding (TPE) represents each table token using the ancestor paths of its row and column headers, preserving its position within the multi-level table structure. We integrate DeepTable with TableLoRA (He et al., 2025) to inject structural information into parameter-efficient adaptation. Across three LLM backbones, DeepTable consistently improves the corresponding TableLoRA baselines on three table question answering benchmarks, achieving average gains of 7.42 points on HiTab, 3.23 points on WikiTQ, and 2.01 BLEU points on FeTaQA. These results demonstrate the effectiveness of the proposed structural biases across different LLM backbones.

cs.CL↗

arXiv2026

NOVA: Normal-Side Modeling for Training-Free Zero-Shot Video Anomaly Detection

Training-free zero-shot video anomaly detection (ZS-VAD) leverages vision-language models (VLMs) to localize anomaly instances from a predefined anomaly vocabulary, without providing any video. Existing CLIP-based methods often emphasize anomaly-side semantics, while the competing normality side remains less carefully formulated. We identify two key limitations in existing solutions: (i) blurred decision boundary: normal prompts may contain ambiguous verbs, such as running, that are semantically close to anomalies, reducing normal and abnormal separation in the VLM embedding space; and (ii) modality gap: poor alignment between features of textual normal anchors and visual frames. We propose NOVA, a training-free ZS-VAD framework that strengthens the normal side at both linguistic and visual levels. NOVA introduces Normality-Aware Prompt Construction (NA), which excludes anomaly-adjacent verbs and biases normal descriptions toward static, low-motion scenes. To overcome the text-vision modality gap, NOVA constructs a Visual Normality Anchor (VNA), which creates a weighted visual normal anchor from the initial frames of each test video, providing a video-specific normal reference without task-specific training or annotations. NOVA achieves 89.86 percent AUC on UCF-Crime and 95.07 percent AUC and 84.82 percent AP on XD-Violence, reaching state-of-the-art performance among comparable training-free zero-shot methods.

cs.CV↗

arXiv2025

Dynamic Participation in Federated Learning: Benchmarks and a Knowledge Pool Plugin

Federated learning (FL) enables clients to collaboratively train a shared model in a distributed manner, setting it apart from traditional deep learning paradigms. However, most existing FL research assumes consistent client participation, overlooking the practical scenario of dynamic participation (DPFL), where clients may intermittently join or leave during training. Moreover, no existing benchmarking framework systematically supports the study of DPFL-specific challenges. In this work, we present the first open-source framework explicitly designed for benchmarking FL models under dynamic client participation. Our framework provides configurable data distributions, participation patterns, and evaluation metrics tailored to DPFL scenarios. Using this platform, we benchmark four major categories of widely adopted FL models and uncover substantial performance degradation under dynamic participation. To address these challenges, we further propose Knowledge-Pool Federated Learning (KPFL), a generic plugin that maintains a shared knowledge pool across both active and idle clients. KPFL leverages dual-age and data-bias weighting, combined with generative knowledge distillation, to mitigate instability and prevent knowledge loss. Extensive experiments demonstrate the significant impact of dynamic participation on FL performance and the effectiveness of KPFL in improving model robustness and generalization.

cs.LG↗

arXiv2025

DDOT: A Derivative-directed Dual-decoder Ordinary Differential Equation Transformer for Dynamic System Modeling

Uncovering the underlying ordinary differential equations (ODEs) that govern dynamic systems is crucial for advancing our understanding of complex phenomena. Traditional symbolic regression methods often struggle to capture the temporal dynamics and intervariable correlations inherent in ODEs. ODEFormer, a state-of-the-art method for inferring multidimensional ODEs from single trajectories, has made notable progress. However, its focus on single-trajectory evaluation is highly sensitive to initial starting points, which may not fully reflect true performance. To address this, we propose the divergence difference metric (DIV-diff), which evaluates divergence over a grid of points within the target region, offering a comprehensive and stable analysis of the variable space. Alongside, we introduce DDOT (Derivative-Directed Dual-Decoder Ordinary Differential Equation Transformer), a transformer-based model designed to reconstruct multidimensional ODEs in symbolic form. By incorporating an auxiliary task predicting the ODE's derivative, DDOT effectively captures both structure and dynamic behavior. Experiments on ODEBench show DDOT outperforms existing symbolic regression methods, achieving an absolute improvement of 4.58% and 1.62% in $P(R^2 > 0.9)$ for reconstruction and generalization tasks, respectively, and an absolute reduction of 3.55% in DIV-diff. Furthermore, DDOT demonstrates real-world applicability on an anesthesia dataset, highlighting its practical impact.

cs.LG↗

arXiv2024

Structure fault diameter of hypercubes

Structure connectivity and substructure connectivity are innovative indicators for assessing network reliability and fault tolerance. Similarly, fault diameter evaluates fault tolerance and transmission delays in networks. This paper extends the concept of fault diameter by introducing two new variants: structure fault diameter and substructure fault diameter, derived from structure connectivity and substructure connectivity respectively. For a connected graph $G$ with $W$-structure connectivity $\kappa(G;W)$ or $W$-substructure connectivity $\kappa^s(G;W)$, the $W$-structure fault diameter $D_f(G;W)$ and $W$-substructure fault diameter $D_f^s(G;W)$ are defined as the maximum diameter of any subgraph of $G$ resulting from removing up to $\kappa(G;W)-1$ $W$-structures or $\kappa^s(G;W)-1$ $W$-substructures. For the $n$-dimensional hypercube $Q_n$ with $n \geq 3$ and $1 \leq m \leq n - 2$, we determine both $D_f(Q_n;Q_m)$ and $D_f^s(Q_n;Q_1)$. These findings generalize existing results for the diameter and fault diameter of $Q_n$, providing a broader understanding of the hypercube's structural properties under fault conditions.

math.CO↗

arXiv2024

MP-PolarMask: A Faster and Finer Instance Segmentation for Concave Images

While there are a lot of models for instance segmentation, PolarMask stands out as a unique one that represents an object by a Polar coordinate system. With an anchor-box-free design and a single-stage framework that conducts detection and segmentation at one time, PolarMask is proved to be able to balance efficiency and accuracy. Hence, it can be easily connected with other downstream real-time applications. In this work, we observe that there are two deficiencies associated with PolarMask: (i) inability of representing concave objects and (ii) inefficiency in using ray regression. We propose MP-PolarMask (Multi-Point PolarMask) by taking advantage of multiple Polar systems. The main idea is to extend from one main Polar system to four auxiliary Polar systems, thus capable of representing more complicated convex-and-concave-mixed shapes. We validate MP-PolarMask on both general objects and food objects of the COCO dataset, and the results demonstrate significant improvement of 13.69% in AP_L and 7.23% in AP over PolarMask with 36 rays.

cs.CV↗

arXiv2023

Tracking Players in a Badminton Court by Two Cameras

This study proposes a simple method for multi-object tracking (MOT) of players in a badminton court. We leverage two off-the-shelf cameras, one on the top of the court and the other on the side of the court. The one on the top is to track players' trajectories, while the one on the side is to analyze the pixel features of players. By computing the correlations between adjacent frames and engaging the information of the two cameras, MOT of badminton players is obtained. This two-camera approach addresses the challenge of player occlusion and overlapping in a badminton court, providing player trajectory tracking and multi-angle analysis. The presented system offers insights into the positions and movements of badminton players, thus serving as a coaching or self-training tool for badminton players to improve their gaming strategies.

cs.CV↗

arXiv2021

A Local Diagnosis Algorithm for Hypercube-like Networks under the BGM Diagnosis Model

System diagnosis is process of identifying faulty nodes in a system. An efficient diagnosis is crucial for a multiprocessor system. The BGM diagnosis model is a modification of the PMC diagnosis model, which is a test-based diagnosis. In this paper, we present a specific structure and propose an algorithm for diagnosing a node in a system under the BGM model. We also give a polynomial-time algorithm that a node in a hypercube-like network can be diagnosed correctly in three test rounds under the BGM diagnosis model.

cs.DM↗

arXiv2021

The Component Diagnosability of General Networks

The processor failures in a multiprocessor system have a negative impact on its distributed computing efficiency. Because of the rapid expansion of multiprocessor systems, the importance of fault diagnosis is becoming increasingly prominent. The $h$-component diagnosability of $G$, denoted by $ct_{h}(G)$, is the maximum number of nodes of the faulty set $F$ that is correctly identified in a system, and the number of components in $G-F$ is at least $h$. In this paper, we determine the $(h+1)$-component diagnosability of general networks under the PMC model and MM$^{*}$ model. As applications, the component diagnosability is explored for some well-known networks, including complete cubic networks, hierarchical cubic networks, generalized exchanged hypercubes, dual-cube-like networks, hierarchical hypercubes, Cayley graphs generated by transposition trees (except star graphs), and DQcube as well. Furthermore, we provide some comparison results between the component diagnosability and other fault diagnosabilities.

cs.DC↗