arXiv subjects

Yu-Wei Chen

Publications and source records attributed to Yu-Wei Chen.

7 recordsLinked to original sources

arXiv2026

Near-Optimal Private Tests for Simple and MLR Hypotheses

We develop a near-optimal testing procedure under the framework of Gaussian differential privacy for simple as well as one- and two-sided tests under monotone likelihood ratio conditions. Our mechanism is based on a private mean estimator with data-driven clamping bounds, whose population risk matches the private minimax rate up to logarithmic factors. Using this estimator, we construct private test statistics that achieve the same asymptotic relative efficiency as the non-private, most powerful tests while maintaining conservative type I error control. In addition to our theoretical results, our numerical experiments show that our private tests outperform competing DP methods and offer comparable power to the non-private most powerful tests, even at moderately small sample sizes and privacy loss budgets.

stat.ML↗

arXiv2026

Judge Model for Large-scale Multimodality Benchmarks

We propose a dedicated multimodal Judge Model designed to provide reliable, explainable evaluation across a diverse suite of tasks. Our benchmark spans text, audio, image, and video modalities, drawing from carefully sampled public datasets with fixed seeds to ensure reproducibility and minimize train test leakage. Instead of simple scoring, our framework aggregates multimodal judgments, analyzes the quality and reasoning consistency of model outputs, and generates diagnostic feedback. We evaluate several MLLMs, including Gemini 2.5, Phi 4, and Qwen 2.5, across 280 multimodal samples and compare judge model assessments with human annotators. Results show strong alignment between the Judge Model and human scores, demonstrating its potential as a scalable, interpretable evaluation pipeline for future multimodal AI research.

cs.LG↗

arXiv2025

DynFaceRestore: Balancing Fidelity and Quality in Diffusion-Guided Blind Face Restoration with Dynamic Blur-Level Mapping and Guidance

Blind Face Restoration aims to recover high-fidelity, detail-rich facial images from unknown degraded inputs, presenting significant challenges in preserving both identity and detail. Pre-trained diffusion models have been increasingly used as image priors to generate fine details. Still, existing methods often use fixed diffusion sampling timesteps and a global guidance scale, assuming uniform degradation. This limitation and potentially imperfect degradation kernel estimation frequently lead to under- or over-diffusion, resulting in an imbalance between fidelity and quality. We propose DynFaceRestore, a novel blind face restoration approach that learns to map any blindly degraded input to Gaussian blurry images. By leveraging these blurry images and their respective Gaussian kernels, we dynamically select the starting timesteps for each blurry image and apply closed-form guidance during the diffusion sampling process to maintain fidelity. Additionally, we introduce a dynamic guidance scaling adjuster that modulates the guidance strength across local regions, enhancing detail generation in complex areas while preserving structural fidelity in contours. This strategy effectively balances the trade-off between fidelity and quality. DynFaceRestore achieves state-of-the-art performance in both quantitative and qualitative evaluations, demonstrating robustness and effectiveness in blind face restoration. Project page at https://nycu-acm.github.io/DynFaceRestore/

cs.CV↗

arXiv2025

Particle Filter for Bayesian Inference on Privatized Data

Differential Privacy (DP) is a probabilistic framework that protects privacy while preserving data utility. To protect the privacy of the individuals in the dataset, DP requires adding a precise amount of noise to a statistic of interest; however, this noise addition alters the resulting sampling distribution, making statistical inference challenging. One of the main DP goals in Bayesian analysis is to make statistical inference based on the private posterior distribution. While existing methods have strengths in specific conditions, they can be limited by poor mixing, strict assumptions, or low acceptance rates. We propose a novel particle filtering algorithm, which features (i) consistent estimates, (ii) Monte Carlo error estimates and asymptotic confidence intervals, (iii) computational efficiency, and (iv) accommodation to a wide variety of priors, models, and privacy mechanisms with minimal assumptions. We empirically evaluate our algorithm through a variety of simulation settings as well as an application to a 2021 Canadian census dataset, demonstrating the efficacy and adaptability of the proposed sampler.

stat.CO↗

arXiv2025

Optimal Survey Design for Private Mean Estimation

This work identifies the first privacy-aware stratified sampling scheme that minimizes the variance for general private mean estimation under the Laplace, Discrete Laplace (DLap) and Truncated-Uniform-Laplace (TuLap) mechanisms within the framework of differential privacy (DP). We view stratified sampling as a subsampling operation, which amplifies the privacy guarantee; however, to have the same final privacy guarantee for each group, different nominal privacy budgets need to be used depending on the subsampling rate. Ignoring the effect of DP, traditional stratified sampling strategies risk significant variance inflation. We phrase our optimal survey design as an optimization problem, where we determine the optimal subsampling sizes for each group with the goal of minimizing the variance of the resulting estimator. We establish strong convexity of the variance objective, propose an efficient algorithm to identify the integer-optimal design, and offer insights on the structure of the optimal design.

stat.ML↗

arXiv2024

Always Clear Days: Degradation Type and Severity Aware All-In-One Adverse Weather Removal

All-in-one adverse weather removal is an emerging topic on image restoration, which aims to restore multiple weather degradations in an unified model, and the challenge are twofold. First, discover and handle the property of multi-domain in target distribution formed by multiple weather conditions. Second, design efficient and effective operations for different degradations. To resolve this problem, most prior works focus on the multi-domain caused by different weather types. Inspired by inter\&intra-domain adaptation literature, we observe that not only weather type but also weather severity introduce multi-domain within each weather type domain, which is ignored by previous methods, and further limit their performance. To this end, we propose a degradation type and severity aware model, called UtilityIR, for blind all-in-one bad weather image restoration. To extract weather information from single image, we propose a novel Marginal Quality Ranking Loss (MQRL) and utilize Contrastive Loss (CL) to guide weather severity and type extraction, and leverage a bag of novel techniques such as Multi-Head Cross Attention (MHCA) and Local-Global Adaptive Instance Normalization (LG-AdaIN) to efficiently restore spatial varying weather degradation. The proposed method can outperform the state-of-the-art methods subjectively and objectively on different weather removal tasks with a large margin, and enjoy less model parameters. Proposed method even can restore unseen combined multiple degradation images, and modulate restoration level. Implementation code and pre-trained weights will be available at \url{https://github.com/fordevoted/UtilityIR}

cs.CV↗

arXiv2022

Domain Adaptation for Underwater Image Enhancement via Content and Style Separation

Underwater image suffer from color cast, low contrast and hazy effect due to light absorption, refraction and scattering, which degraded the high-level application, e.g, object detection and object tracking. Recent learning-based methods demonstrate astonishing performance on underwater image enhancement, however, most of these works use synthetic pair data for supervised learning and ignore the domain gap to real-world data. To solve this problem, we propose a domain adaptation framework for underwater image enhancement via content and style separation, different from prior works of domain adaptation for underwater image enhancement, which target to minimize the latent discrepancy of synthesis and real-world data, we aim to separate encoded feature into content and style latent and distinguish style latent from different domains, i.e. synthesis, real-world underwater and clean domain, and process domain adaptation and image enhancement in latent space. By latent manipulation, our model provide a user interact interface to adjust different enhanced level for continuous change. Experiment on various public real-world underwater benchmarks demonstrate that the proposed framework is capable to perform domain adaptation for underwater image enhancement and outperform various state-of-the-art underwater image enhancement algorithms in quantity and quality. The model and source code will be available at https://github.com/fordevoted/UIESS

cs.CV↗