arXiv subjects

Hung-Hsuan Chen

Publications and source records attributed to Hung-Hsuan Chen.

At least 19 recordsLinked to original sources

arXiv2026

Thinking Deeper, Not Longer: Memory-Efficient Test-Time Reasoning with Depth-Recurrent Transformers for Compositional Generalization

Standard Transformers have a fixed computational depth, limiting their ability to generalize to tasks that require variable-depth reasoning. The usual remedy, Chain-of-Thought (CoT), spends tokens to reason, inflating the key--value cache and making latency grow with the step count, so memory becomes the limiting cost when reasoning is served over large query batches. We study a depth-recurrent Transformer that decouples computational depth from parameter count by iterating a shared-weight block, so that each added reasoning step costs flat memory and linear latency, with no token generation. Three ingredients keep the recurrence stable for 20+ thinking steps: a silent thinking objective that supervises only the final output, LayerScale initialization, and an identity-biased gate that opens a gradient highway across steps. We characterize it on three compositional domains with decreasing structural bias: graph reachability (adjacency masking), nested boolean logic (relative positioning), and unstructured relational text (no positional cue). We find a \emph{computational frontier}: accuracy climbs once the thinking-step count meets the task's complexity, reaching near-perfect performance on the two structured tasks and a lower plateau on unstructured text. How it climbs depends on the structural bias---abruptly from chance on the graph task, gradually on the other two. Depth recurrence extrapolates beyond the training range: it succeeds on the graph task where fixed-depth models barely extrapolate, and on the two sequence tasks comes within two points of fixed-depth Transformers that use $4$--$6.4\times$ more parameters. On the graph task, whose adjacency mask makes propagation depth verifiable, intermediate per-step supervision---a standard recipe for deep iterative models---consistently \emph{harms} this extrapolation. We release the code for reproducibility.

cs.LG↗

arXiv2026

Convergence in Science, Divergence in Religion: Calibrated Framing Differences Across Wikipedia's Language Editions

When Wikipedia's language editions describe the same concept, how differently do they frame it? Prior work measures coverage gaps between editions; we measure framing distance for matched concepts. We analyze 2,799 valid articles from 3,000 possible concept-language observations, spanning 150 Wikidata-anchored concepts, 20 language editions, 4 domains, and a calibration set. Raw embedding distances reflect both content differences and how well the encoder aligns each language pair. Even among calibration concepts with stable cross-cultural denotations (e.g., chemical elements, numbers, colors), the largest language-pair mean distance is 3.6 times the smallest, and distances are typically smaller within language families. We define a baseline-adjusted distance (calibrated distance): the distance between two language versions of a concept minus the mean distance for calibration concepts in the same language pair. This adjustment substantially reduces pair-specific alignment differences and the language-family pattern. Across three multilingual encoders (LaBSE, multilingual MPNet, and CMLM), scientific articles align more closely than calibration articles, and all three rank religion first and science/technology last. Concept-level rankings are highly consistent across encoders (Spearman rho=0.75-0.79 for MPNet and CMLM relative to LaBSE). Religion lies significantly above the calibration baseline under LaBSE. Within politics, divergence concentrates on concepts such as censorship and refugee, while democracy and human rights are among the most aligned. Code, data, and per-language-pair calibration baselines are released.\footnote{https://github.com/hhchen1105/cross-linqual-concept}

cs.CL↗

arXiv2026

CeRA: Breaking the Linear Ceiling of Low-Rank Adaptation with Non-linearity Retained at Inference

Low-Rank Adaptation (LoRA) dominates parameter-efficient fine-tuning (PEFT). However, it faces a ``linear ceiling'': increasing the rank yields diminishing returns in expressive capacity due to linear constraints. We introduce CeRA (Capacity-enhanced Rank Adaptation), a weight-level parallel adapter that injects SiLU gating and dropout to induce non-linearity during inference, thereby placing it in a different function class from adapters whose non-linearity exists during training and collapses to an affine map at inference time. On both the basic arithmetic (GSM8K) and the complex MATH benchmark, CeRA is markedly more parameter-efficient. Across a full rank $\times$ learning rate sweep, CeRA at rank 64 achieves the highest MATH pass@1 of any configuration in the grid (23.6\%), matching or exceeding both a rank-512 LoRA (22.4\%) and DoRA (19.8\%) while using only 1/8 of the parameter budget. With the rank and learning rate fixed, CeRA equals or outperforms LoRA in 10 of 12 matched settings. Spectrally, CeRA's learned updates utilize the singular-value spectrum more broadly than linear adapters, which exhibit rank collapse at high rank, although a scale-matched control shows that this difference stems mostly from output scale and partially from non-linearity. Additionally, dropout appears to contribute to regularization rather than rank expansion. We release the code for reproducibility: https://github.com/hhchen1105/cera.

cs.LG↗

arXiv2026

More Women, Same Stereotypes: Unpacking the Gender Bias Paradox in Large Language Models

Large Language Models (LLMs) have revolutionized natural language processing, yet concerns persist regarding their tendency to reflect or amplify social biases. This study introduces a novel evaluation framework to uncover gender biases in LLMs: using free-form storytelling to surface biases embedded within the models. A systematic analysis of ten prominent LLMs shows a consistent pattern of overrepresenting female characters across occupations, likely due to supervised fine-tuning (SFT) and reinforcement learning from human feedback (RLHF). Paradoxically, despite this overrepresentation, the occupational gender distributions produced by these LLMs align more closely with human stereotypes than with real-world labor data. This highlights the challenge and importance of implementing balanced mitigation measures to promote fairness and prevent the establishment of potentially new biases. We release the prompts and LLM-generated stories at GitHub.

cs.CL↗

arXiv2026

Structure-Preserving Graph Contrastive Learning for Mathematical Information Retrieval

This paper introduces Variable Substitution as a domain-specific graph augmentation technique for graph contrastive learning (GCL) in the context of searching for mathematical formulas. Standard GCL augmentation techniques often distort the semantic meaning of mathematical formulas, particularly for small and highly structured graphs. Variable Substitution, on the other hand, preserves the core algebraic relationships and formula structure. To demonstrate the effectiveness of our technique, we apply it to a classic GCL-based retrieval model. Experiments show that this straightforward approach significantly improves retrieval performance compared to generic augmentation strategies. We release the code on GitHub.\footnote{https://github.com/lazywulf/formula_ret_aug}.

cs.IR↗

arXiv2026

SCPL: Enhancing Neural Network Training Throughput with Decoupled Local Losses and Model Parallelism

Adopting large-scale AI models in enterprise information systems is often hindered by high training costs and long development cycles, posing a significant managerial challenge. The standard end-to-end backpropagation (BP) algorithm is a primary driver of modern AI, but it is also the source of inefficiency in training deep networks. This paper introduces a new training methodology, Supervised Contrastive Parallel Learning (SCPL), that addresses this issue by decoupling BP and transforming a long gradient flow into multiple short ones. This design enables the simultaneous computation of parameter gradients in different layers, achieving superior model parallelism and enhancing training throughput. Detailed experiments are presented to demonstrate the efficiency and effectiveness of our model compared to BP, Early Exit, GPipe, and Associated Learning (AL), a state-of-the-art method for decoupling backpropagation. By mitigating a fundamental performance bottleneck, SCPL provides a practical pathway for organizations to develop and deploy advanced information systems more cost-effectively and with greater agility. The experimental code is released for reproducibility. https://github.com/minyaho/scpl/

cs.LG↗

arXiv2026

From Volumes to Slices: Computationally Efficient Contrastive Learning for Sequential Abdominal CT Analysis

The requirement for expert annotations limits the effectiveness of deep learning for medical image analysis. Although 3D self-supervised methods like volume contrast learning (VoCo) are powerful and partially address the labeling scarcity issue, their high computational cost and memory consumption are barriers. We propose 2D-VoCo, an efficient adaptation of the VoCo framework for slice-level self-supervised pre-training that learns spatial-semantic features from unlabeled 2D CT slices via contrastive learning. The pre-trained CNN backbone is then integrated into a CNN-LSTM architecture to classify multi-organ injuries. In the RSNA 2023 Abdominal Trauma dataset, 2D-VoCo pre-training significantly improves mAP, precision, recall, and RSNA score over training from scratch. Our framework provides a practical method to reduce the dependency on labeled data and enhance model performance in clinical CT analysis. We release the code for reproducibility. https://github.com/tkz05/2D-VoCo-CT-Classifier

cs.CV↗

arXiv2026

GraphFusionSBR: Denoising Multi-Channel Graphs for Session-Based Recommendation

Session-based recommendation systems must capture implicit user intents from sessions. However, existing models suffer from issues such as item interaction dominance and noisy sessions. We propose a multi-channel recommendation model, including a knowledge graph channel, a session hypergraph channel, and a session line graph channel, to capture information from multiple sources. Our model adaptively removes redundant edges in the knowledge graph channel to reduce noise. Knowledge graph representations cooperate with hypergraph representations for prediction to alleviate item dominance. We also generate in-session attention for denoising. Finally, we maximize mutual information between the hypergraph and line graph channels as an auxiliary task. Experiments demonstrate that our method enhances the accuracy of various recommendations, including e-commerce and multimedia recommendations. We release the code on GitHub for reproducibility.\footnote{https://github.com/hohehohe0509/DSR-HK}

cs.IR↗

arXiv2025

Contrastive ECOC: Learning Output Codes for Adversarial Defense

Although one-hot encoding is commonly used for multiclass classification, it is not always the most effective encoding mechanism. Error Correcting Output Codes (ECOC) address multiclass classification by mapping each class to a unique codeword used as a label. Traditional ECOC methods rely on manually designed or randomly generated codebooks, which are labor-intensive and may yield suboptimal, dataset-agnostic results. This paper introduces three models for automated codebook learning based on contrastive learning, allowing codebooks to be learned directly and adaptively from data. Across four datasets, our proposed models demonstrate superior robustness to adversarial attacks compared to two baselines. The source is available at https://github.com/YuChou20/Automated-Codebook-Learning-with-Error-Correcting-Output-Code-Technique.

cs.LG↗

arXiv2025

DeInfoReg: A Decoupled Learning Framework for Better Training Throughput

This paper introduces Decoupled Supervised Learning with Information Regularization (DeInfoReg), a novel approach that transforms a long gradient flow into multiple shorter ones, thereby mitigating the vanishing gradient problem. Integrating a pipeline strategy, DeInfoReg enables model parallelization across multiple GPUs, significantly improving training throughput. We compare our proposed method with standard backpropagation and other gradient flow decomposition techniques. Extensive experiments on diverse tasks and datasets demonstrate that DeInfoReg achieves superior performance and better noise resistance than traditional BP models and efficiently utilizes parallel computing resources. The code for reproducibility is available at: https://github.com/ianzih/Decoupled-Supervised-Learning-for-Information-Regularization/.

cs.LG↗

arXiv2025

Flexible Bivariate Beta Mixture Model: A Probabilistic Approach for Clustering Complex Data Structures

Clustering is essential in data analysis and machine learning, but traditional algorithms like $k$-means and Gaussian Mixture Models (GMM) often fail with nonconvex clusters. To address the challenge, we introduce the Flexible Bivariate Beta Mixture Model (FBBMM), which utilizes the flexibility of the bivariate beta distribution to handle diverse and irregular cluster shapes. Using the Expectation Maximization (EM) algorithm and Sequential Least Squares Programming (SLSQP) optimizer for parameter estimation, we validate FBBMM on synthetic and real-world datasets, demonstrating its superior performance in clustering complex data structures, offering a robust solution for big data analytics across various domains. We release the experimental code at https://github.com/yung-peng/MBMM-and-FBBMM.

cs.LG↗

arXiv2025

Dynamic DropConnect: Enhancing Neural Network Robustness through Adaptive Edge Dropping Strategies

Dropout and DropConnect are well-known techniques that apply a consistent drop rate to randomly deactivate neurons or edges in a neural network layer during training. This paper introduces a novel methodology that assigns dynamic drop rates to each edge within a layer, uniquely tailoring the dropping process without incorporating additional learning parameters. We perform experiments on synthetic and openly available datasets to validate the effectiveness of our approach. The results demonstrate that our method outperforms Dropout, DropConnect, and Standout, a classic mechanism known for its adaptive dropout capabilities. Furthermore, our approach improves the robustness and generalization of neural network training without increasing computational complexity. The complete implementation of our methodology is publicly accessible for research and replication purposes at https://github.com/ericabd888/Adjusting-the-drop-probability-in-DropConnect-based-on-the-magnitude-of-the-gradient/.

cs.LG↗

arXiv2024

Understanding Gradient Boosting Classifier: Training, Prediction, and the Role of $γ_j$

The Gradient Boosting Classifier (GBC) is a widely used machine learning algorithm for binary classification, which builds decision trees iteratively to minimize prediction errors. This document explains the GBC's training and prediction processes, focusing on the computation of terminal node values $γ_j$, which are crucial to optimizing the logistic loss function. We derive $γ_j$ through a Taylor series approximation and provide a step-by-step pseudocode for the algorithm's implementation. The guide explains the theory of GBC and its practical application, demonstrating its effectiveness in binary classification tasks. We provide a step-by-step example in the appendix to help readers understand.

cs.LG↗

arXiv2024

Multivariate Beta Mixture Model: Probabilistic Clustering With Flexible Cluster Shapes

This paper introduces the multivariate beta mixture model (MBMM), a new probabilistic model for soft clustering. MBMM adapts to diverse cluster shapes because of the flexible probability density function of the multivariate beta distribution. We introduce the properties of MBMM, describe the parameter learning procedure, and present the experimental results, showing that MBMM fits diverse cluster shapes on synthetic and real datasets. The code is released anonymously at https://github.com/hhchen1105/mbmm/.

cs.LG↗

arXiv2024

The Effectiveness of Graph Contrastive Learning on Mathematical Information Retrieval

This paper details an empirical investigation into using Graph Contrastive Learning (GCL) to generate mathematical equation representations, a critical aspect of Mathematical Information Retrieval (MIR). Our findings reveal that this simple approach consistently exceeds the performance of the current leading formula retrieval model, TangentCFT. To support ongoing research and development in this field, we have made our source code accessible to the public at https://github.com/WangPeiSyuan/GCL-Formula-Retrieval/.

cs.IR↗

arXiv2023

Toward Efficient and Incremental Spectral Clustering via Parametric Spectral Clustering

Spectral clustering is a popular method for effectively clustering nonlinearly separable data. However, computational limitations, memory requirements, and the inability to perform incremental learning challenge its widespread application. To overcome these limitations, this paper introduces a novel approach called parametric spectral clustering (PSC). By extending the capabilities of spectral clustering, PSC addresses the challenges associated with big data and real-time scenarios and enables efficient incremental clustering with new data points. Experimental evaluations conducted on various open datasets demonstrate the superiority of PSC in terms of computational efficiency while achieving clustering quality mostly comparable to standard spectral clustering. The proposed approach has significant potential for incremental and real-time data analysis applications, facilitating timely and accurate clustering in dynamic and evolving datasets. The findings of this research contribute to the advancement of clustering techniques and open new avenues for efficient and effective data analysis. We publish the experimental code at https://github.com/109502518/PSC_BigData.

cs.LG↗

arXiv2023

Detecting Inactive Cyberwarriors from Online Forums

The proliferation of misinformation has emerged as a new form of warfare in the information age. This type of warfare involves cyberwarriors, who deliberately propagate messages aimed at defaming opponents or fostering unity among allies. In this study, we investigate the level of activity exhibited by cyberwarriors within a large online forum, and remarkably, we discover that only a minute fraction of cyberwarriors are active users. Surprisingly, despite their expected role of actively disseminating misinformation, cyberwarriors remain predominantly silent during peacetime and only spring into action when necessary. Moreover, we analyze the challenges associated with identifying cyberwarriors and provide evidence that detecting inactive cyberwarriors is considerably more challenging than identifying their active counterparts. Finally, we discuss potential methodologies to more effectively identify cyberwarriors during their inactive phases, offering insights into better capturing their presence and actions. The experimental code is released for reproducibility: \url{https://github.com/Ryaninthegame/Detect-Inactive-Spammers-on-PTT}.

cs.SI↗

arXiv2023

TTSWING: a Dataset for Table Tennis Swing Analysis

We introduce TTSWING, a novel dataset designed for table tennis swing analysis. This dataset comprises comprehensive swing information obtained through 9-axis sensors integrated into custom-made racket grips, accompanied by anonymized demographic data of the players. We detail the data collection and annotation procedures. Furthermore, we conduct pilot studies utilizing diverse machine learning models for swing analysis. TTSWING holds tremendous potential to facilitate innovative research in table tennis analysis and is a valuable resource for the scientific community. We release the dataset and experimental codes at https://github.com/DEPhantom/TTSWING.

cs.LG↗