arXiv subjects

Yu-Hsiang Wang

Publications and source records attributed to Yu-Hsiang Wang.

12 recordsLinked to original sources

arXiv2026

WaveletDiff: Multilevel Wavelet Diffusion For Time Series Generation

Time series are ubiquitous in many applications that involve forecasting, classification and causal inference tasks, such as healthcare, finance, audio signal processing and climate sciences. Still, large, high-quality time series datasets remain scarce. Synthetic generation can address this limitation; however, current models confined either to the time or frequency domains struggle to reproduce the inherently multi-scaled structure of real-world time series. We introduce WaveletDiff, a new framework that trains diffusion models directly on wavelet coefficients to exploit the inherent multi-resolution structure of time series data. The model combines dedicated transformers for each decomposition level with cross-level attention mechanisms that enable selective information exchange between temporal and frequency scales through adaptive gating. It is also informed by level-specific energy constraints based on Parseval's theorem which preserve time-frequency properties throughout the diffusion process. Comprehensive tests across six real-world datasets from energy, finance, and neuroscience domains demonstrate that WaveletDiff outperforms the diffusion baselines FourierDiffusion, Diffusion-TS, and SigDiffusions on the majority of metrics, with the smallest margin over FourierDiffusion, while still achieving roughly 3x lower discriminative and Context-FID scores. Against the VAE/transformer-based MSDformer, the results are mostly comparable, with WaveletDiff using fewer parameters and less training time on most datasets. The most revealing finding is the significant performance gap on fMRI data (in favor of MSDformer) and EEG (in favor of WaveletDiff). This finding is explained via a careful testing/examination of the properties of wavelet coefficients for generative, as opposed to analyses/decomposition tasks. Our code is available at https://github.com/GarlicWang/WaveletDiff.

cs.LG↗

arXiv2026

Stochastic Control Policies for Robust Molecular Transition Path Sampling

Transition path sampling (TPS) aims to efficiently generate rare molecular transition trajectories between metastable states and is essential for understanding biomolecular mechanisms. Beyond traditional molecular dynamics (MD)-based sampling, machine learning has become central to state-of-the-art TPS. One major class of methods learns control forces during explicit MD rollouts. By preserving the underlying molecular dynamics, these methods tend to produce more physically plausible trajectories than endpoint-conditioned generators that construct paths directly. However, rollout-based control methods have been reported to exhibit unstable and strongly seed-dependent performance. We recast rollout-based control as learning a path-space proposal distribution and investigate stochasticity placement as a design choice for improving exploration and optimization robustness. We develop two stochastic policies: FS-TPS, which directly parameterizes a state-dependent Gaussian distribution over the control policy output, and LaS-TPS, which samples a compact latent control variable and decodes it into structured, cross-atom-correlated force variation. We conduct extensive multi-seed experiments on three biomolecular systems of increasing size: alanine dipeptide, chignolin, and BBL, a fast-folding protein. Stochastic policies consistently improve transition success and path quality over deterministic-policy baselines while substantially reducing sensitivity to random initialization.

cs.LG↗

arXiv2025

THeGAU: Type-Aware Heterogeneous Graph Autoencoder and Augmentation

Heterogeneous Graph Neural Networks (HGNNs) are effective for modeling Heterogeneous Information Networks (HINs), which encode complex multi-typed entities and relations. However, HGNNs often suffer from type information loss and structural noise, limiting their representational fidelity and generalization. We propose THeGAU, a model-agnostic framework that combines a type-aware graph autoencoder with guided graph augmentation to improve node classification. THeGAU reconstructs schema-valid edges as an auxiliary task to preserve node-type semantics and introduces a decoder-driven augmentation mechanism to selectively refine noisy structures. This joint design enhances robustness, accuracy, and efficiency while significantly reducing computational overhead. Extensive experiments on three benchmark HIN datasets (IMDB, ACM, and DBLP) demonstrate that THeGAU consistently outperforms existing HGNN methods, achieving state-of-the-art performance across multiple backbones.

cs.LG↗

arXiv2025

Financial Risk Relation Identification through Dual-view Adaptation

A multitude of interconnected risk events -- ranging from regulatory changes to geopolitical tensions -- can trigger ripple effects across firms. Identifying inter-firm risk relations is thus crucial for applications like portfolio management and investment strategy. Traditionally, such assessments rely on expert judgment and manual analysis, which are, however, subjective, labor-intensive, and difficult to scale. To address this, we propose a systematic method for extracting inter-firm risk relations using Form 10-K filings -- authoritative, standardized financial documents -- as our data source. Leveraging recent advances in natural language processing, our approach captures implicit and abstract risk connections through unsupervised fine-tuning based on chronological and lexical patterns in the filings. This enables the development of a domain-specific financial encoder with a deeper contextual understanding and introduces a quantitative risk relation score for transparency, interpretable analysis. Extensive experiments demonstrate that our method outperforms strong baselines across multiple evaluation settings. Our codes are available at https://github.com/cnclabs/codes.fin.relation.

cs.CL↗

arXiv2025

DMol: A Highly Efficient and Chemical Motif-Preserving Molecule Generation Platform

We introduce a new graph diffusion model for small molecule generation, DMol, which outperforms the state-of-the-art DiGress model in terms of validity by roughly 1.5% across all benchmarking datasets while reducing the number of diffusion steps by at least 10-fold, and the running time to roughly one half. The performance improvements are a result of a careful change in the objective function and a graph noise scheduling approach which, at each diffusion step, allows one to only change a subset of nodes of varying size in the molecule graph. Another relevant property of the method is that it can be easily combined with junction-tree-like graph representations that arise by compressing a collection of relevant ring structures into supernodes. Unlike classical junction-tree techniques that involve VAEs and require complicated reconstruction steps, compressed DMol directly performs graph diffusion on a graph that compresses only a carefully selected set of frequent carbon rings into supernodes, which results in straightforward sample generation. This compressed DMol method offers additional validity improvements over generic DMol of roughly 2%, increases the novelty of the method, and further improves the running time due to reductions in the graph size.

cs.LG↗

arXiv2025

Strategy Method Effects in Centipede Games: An Optimal Design Approach

We explore the twin questions of when and why the strategy method creates behavioral distortions in the elicitation of choices in laboratory studies of sequential games. While such distortions have been widely documented, the theoretical forces driving these distortions remain poorly understood. In this paper, we compare behavior in six optimally designed centipede games, implemented under three different choice elicitation methods: the direct response method, the reduced strategy method and the full strategy method. These methods elicit behavioral strategies, reduced strategies, and complete strategies, respectively. We find significant behavioral differences across these elicitation methods -- differences that cannot be explained by standard game theory, but are consistent with the predictions of the Dynamic Cognitive Hierarchy solution (Lin and Palfrey, 2024), combined with quantal responses.

econ.GN↗

arXiv2025

Let's Fuse Step by Step: A Generative Fusion Decoding Algorithm with LLMs for Robust and Instruction-Aware ASR and OCR

We propose "Generative Fusion Decoding" (GFD), a novel shallow fusion framework designed to integrate large language models (LLMs) into cross-modal text recognition systems for automatic speech recognition (ASR) and optical character recognition (OCR). We derive the necessary formulations to enable GFD to operate across mismatched token spaces of different models by calculating likelihood at the byte level, thereby enabling seamless fusion and synchronous progression during the decoding process. GFD is plug-and-play by design, making it readily compatible with various auto-regressive models without the need for any re-training. GFD proves effective for general ASR and OCR tasks through intermediate and frequent interactions with LLMs, surpassing cascaded methods in English and Mandarin benchmarks. In addition, GFD transfers in-context learning abilities of LLMs and allows for adaptive ASR in instruction-aware and long-context settings, yielding significant WER reductions of up to 17.7\%.

cs.CL↗

arXiv2024

CLUE: Concept-Level Uncertainty Estimation for Large Language Models

Large Language Models (LLMs) have demonstrated remarkable proficiency in various natural language generation (NLG) tasks. Previous studies suggest that LLMs' generation process involves uncertainty. However, existing approaches to uncertainty estimation mainly focus on sequence-level uncertainty, overlooking individual pieces of information within sequences. These methods fall short in separately assessing the uncertainty of each component in a sequence. In response, we propose a novel framework for Concept-Level Uncertainty Estimation (CLUE) for LLMs. We leverage LLMs to convert output sequences into concept-level representations, breaking down sequences into individual concepts and measuring the uncertainty of each concept separately. We conduct experiments to demonstrate that CLUE can provide more interpretable uncertainty estimation results compared with sentence-level uncertainty, and could be a useful tool for various tasks such as hallucination detection and story generation.

cs.CL↗

arXiv2024

SMILEtrack: SiMIlarity LEarning for Occlusion-Aware Multiple Object Tracking

Despite recent progress in Multiple Object Tracking (MOT), several obstacles such as occlusions, similar objects, and complex scenes remain an open challenge. Meanwhile, a systematic study of the cost-performance tradeoff for the popular tracking-by-detection paradigm is still lacking. This paper introduces SMILEtrack, an innovative object tracker that effectively addresses these challenges by integrating an efficient object detector with a Siamese network-based Similarity Learning Module (SLM). The technical contributions of SMILETrack are twofold. First, we propose an SLM that calculates the appearance similarity between two objects, overcoming the limitations of feature descriptors in Separate Detection and Embedding (SDE) models. The SLM incorporates a Patch Self-Attention (PSA) block inspired by the vision Transformer, which generates reliable features for accurate similarity matching. Second, we develop a Similarity Matching Cascade (SMC) module with a novel GATE function for robust object matching across consecutive video frames, further enhancing MOT performance. Together, these innovations help SMILETrack achieve an improved trade-off between the cost ({\em e.g.}, running speed) and performance (e.g., tracking accuracy) over several existing state-of-the-art benchmarks, including the popular BYTETrack method. SMILETrack outperforms BYTETrack by 0.4-0.8 MOTA and 2.1-2.2 HOTA points on MOT17 and MOT20 datasets. Code is available at https://github.com/pingyang1117/SMILEtrack_Official

cs.CV↗

arXiv2023

MiniSUPERB: Lightweight Benchmark for Self-supervised Speech Models

SUPERB was proposed to evaluate the generalizability of self-supervised learning (SSL) speech models across various tasks. However, it incurs high computational costs due to the large datasets and diverse tasks. In this paper, we introduce MiniSUPERB, a lightweight benchmark that efficiently evaluates SSL speech models with comparable results to SUPERB but lower computational costs significantly. We carefully select representative tasks, sample datasets, and extract model representations offline. Our approach achieves a Spearman's rank correlation of 0.954 and 0.982 with SUPERB Paper and SUPERB Challenge, respectively. Additionally, we reduce the computational cost by 97% in terms of Multiply-ACcumulate operations (MACs). Furthermore, we evaluate SSL speech models in few-shot scenarios and observe significant variations in their performance. To our knowledge, this is the first study to examine both the computational cost of the model itself and the cost of evaluating it on a benchmark.

eess.AS↗

arXiv2015

A Fast Hyperplane-Based Minimum-Volume Enclosing Simplex Algorithm for Blind Hyperspectral Unmixing

Hyperspectral unmixing (HU) is a crucial signal processing procedure to identify the underlying materials (or endmembers) and their corresponding proportions (or abundances) from an observed hyperspectral scene. A well-known blind HU criterion, advocated by Craig in early 1990's, considers the vertices of the minimum-volume enclosing simplex of the data cloud as good endmember estimates, and it has been empirically and theoretically found effective even in the scenario of no pure pixels. However, such kind of algorithms may suffer from heavy simplex volume computations in numerical optimization, etc. In this work, without involving any simplex volume computations, by exploiting a convex geometry fact that a simplest simplex of N vertices can be defined by N associated hyperplanes, we propose a fast blind HU algorithm, for which each of the N hyperplanes associated with the Craig's simplex of N vertices is constructed from N-1 affinely independent data pixels, together with an endmember identifiability analysis for its performance support. Without resorting to numerical optimization, the devised algorithm searches for the N(N-1) active data pixels via simple linear algebraic computations, accounting for its computational efficiency. Monte Carlo simulations and real data experiments are provided to demonstrate its superior efficacy over some benchmark Craig-criterion-based algorithms in both computational efficiency and estimation accuracy.

cs.CG↗

arXiv2008

Enhanced Sensing Characteristics in MEMS-based Formaldehyde Gas Sensor

This study has successfully demonstrated a novel self-heating formaldehyde gas sensor based on a thin film of NiO sensing layer. A new fabrication process has been developed in which the Pt micro heater and electrodes are deposited directly on the substrate and the NiO thin film is deposited above on the micro heater to serve as sensing layer. Pt electrodes are formed below the sensing layer to measure the electrical conductivity changes caused by formaldehyde oxidation at the oxide surface. Furthermore, the upper sensing layer and NiO/Al2O3 co-sputtering significantly increases the sensitivity of the gas sensor, improves its detection limit capability. The microfabricated formaldehyde gas sensor presented in this study is suitable not only for industrial process monitoring, but also for the detection of formaldehyde concentrations in buildings in order to safeguard human health.

cs.OH↗