arXiv subjects

Song Wang

Publications and source records attributed to Song Wang.

At least 199 records · Page 11Linked to original sources

arXiv2025

Uncertainty-Aware Large Language Models for Explainable Disease Diagnosis

Explainable disease diagnosis, which leverages patient information (e.g., signs and symptoms) and computational models to generate probable diagnoses and reasonings, offers clear clinical values. However, when clinical notes encompass insufficient evidence for a definite diagnosis, such as the absence of definitive symptoms, diagnostic uncertainty usually arises, increasing the risk of misdiagnosis and adverse outcomes. Although explicitly identifying and explaining diagnostic uncertainties is essential for trustworthy diagnostic systems, it remains under-explored. To fill this gap, we introduce ConfiDx, an uncertainty-aware large language model (LLM) created by fine-tuning open-source LLMs with diagnostic criteria. We formalized the task and assembled richly annotated datasets that capture varying degrees of diagnostic ambiguity. Evaluating ConfiDx on real-world datasets demonstrated that it excelled in identifying diagnostic uncertainties, achieving superior diagnostic performance, and generating trustworthy explanations for diagnoses and uncertainties. To our knowledge, this is the first study to jointly address diagnostic uncertainty recognition and explanation, substantially enhancing the reliability of automatic diagnostic systems.

cs.CL↗

arXiv2025

Deep Learning Model Inversion Attacks and Defenses: A Comprehensive Survey

The rapid adoption of deep learning in sensitive domains has brought tremendous benefits. However, this widespread adoption has also given rise to serious vulnerabilities, particularly model inversion (MI) attacks, posing a significant threat to the privacy and integrity of personal data. The increasing prevalence of these attacks in applications such as biometrics, healthcare, and finance has created an urgent need to understand their mechanisms, impacts, and defense methods. This survey aims to fill the gap in the literature by providing a structured and in-depth review of MI attacks and defense strategies. Our contributions include a systematic taxonomy of MI attacks, extensive research on attack techniques and defense mechanisms, and a discussion about the challenges and future research directions in this evolving field. By exploring the technical and ethical implications of MI attacks, this survey aims to offer insights into the impact of AI-powered systems on privacy, security, and trust. In conjunction with this survey, we have developed a comprehensive repository to support research on MI attacks and defenses. The repository includes state-of-the-art research papers, datasets, evaluation metrics, and other resources to meet the needs of both novice and experienced researchers interested in MI attacks and defenses, as well as the broader field of AI security and privacy. The repository will be continuously maintained to ensure its relevance and utility. It is accessible at https://github.com/overgter/Deep-Learning-Model-Inversion-Attacks-and-Defenses.

cs.CR↗

arXiv2025

Acquire and then Adapt: Squeezing out Text-to-Image Model for Image Restoration

Recently, pre-trained text-to-image (T2I) models have been extensively adopted for real-world image restoration because of their powerful generative prior. However, controlling these large models for image restoration usually requires a large number of high-quality images and immense computational resources for training, which is costly and not privacy-friendly. In this paper, we find that the well-trained large T2I model (i.e., Flux) is able to produce a variety of high-quality images aligned with real-world distributions, offering an unlimited supply of training samples to mitigate the above issue. Specifically, we proposed a training data construction pipeline for image restoration, namely FluxGen, which includes unconditional image generation, image selection, and degraded image simulation. A novel light-weighted adapter (FluxIR) with squeeze-and-excitation layers is also carefully designed to control the large Diffusion Transformer (DiT)-based T2I model so that reasonable details can be restored. Experiments demonstrate that our proposed method enables the Flux model to adapt effectively to real-world image restoration tasks, achieving superior scores and visual quality on both synthetic and real-world degradation datasets - at only about 8.5\% of the training cost compared to current approaches.

cs.CV↗

arXiv2025

Efficient MAP Estimation of LLM Judgment Performance with Prior Transfer

LLM ensembles are widely used for LLM judges. However, how to estimate their accuracy, especially in an efficient way, is unknown. In this paper, we present a principled maximum a posteriori (MAP) framework for an economical and precise estimation of the performance of LLM ensemble judgment. We first propose a mixture of Beta-Binomial distributions to model the judgment distribution, revising from the vanilla Binomial distribution. Next, we introduce a conformal prediction-driven approach that enables adaptive stopping during iterative sampling to balance accuracy with efficiency. Furthermore, we design a prior transfer mechanism that utilizes learned distributions on open-source datasets to improve estimation on a target dataset when only scarce annotations are available. Finally, we present BetaConform, a framework that integrates our distribution assumption, adaptive stopping, and the prior transfer mechanism to deliver a theoretically guaranteed distribution estimation of LLM ensemble judgment with minimum labeled samples. BetaConform is also validated empirically. For instance, with only 10 samples from the TruthfulQA dataset, for a Llama ensembled judge, BetaConform gauges its performance with error margin as small as 3.37%.

cs.LG↗

arXiv2025

Are We Merely Justifying Results ex Post Facto? Quantifying Explanatory Inversion in Post-Hoc Model Explanations

Post-hoc explanation methods provide interpretation by attributing predictions to input features. Natural explanations are expected to interpret how the inputs lead to the predictions. Thus, a fundamental question arises: Do these explanations unintentionally reverse the natural relationship between inputs and outputs? Specifically, are the explanations rationalizing predictions from the output rather than reflecting the true decision process? To investigate such explanatory inversion, we propose Inversion Quantification (IQ), a framework that quantifies the degree to which explanations rely on outputs and deviate from faithful input-output relationships. Using the framework, we demonstrate on synthetic datasets that widely used methods such as LIME and SHAP are prone to such inversion, particularly in the presence of spurious correlations, across tabular, image, and text domains. Finally, we propose Reproduce-by-Poking (RBP), a simple and model-agnostic enhancement to post-hoc explanation methods that integrates forward perturbation checks. We further show that under the IQ framework, RBP theoretically guarantees the mitigation of explanatory inversion. Empirically, for example, on the synthesized data, RBP can reduce the inversion by 1.8% on average across iconic post-hoc explanation approaches and domains.

cs.LG↗

arXiv2025

The Ripple Effect of Vulnerabilities in Maven Central: Prevalence, Propagation, and Mitigation Challenges

The widespread use of package managers like Maven has accelerated software development but has also introduced significant security risks due to vulnerabilities in dependencies. In this study, we analyze the prevalence and impact of vulnerabilities within the Maven Central ecosystem, using Common Vulnerabilities and Exposures (CVE) data from OSV.dev and a subsample enriched with aggregated CVE data (CVE_AGGREGATED), which captures both direct and transitive vulnerabilities. In our subsample of around 4 million releases, we found that while only about 1% of releases have direct vulnerabilities, approximately 46.8% are affected by transitive vulnerabilities. This highlights how a small number of vulnerable yet influential artifacts can impact a vast portion of the ecosystem. Moreover, our analysis shows that vulnerabilities propagate rapidly through dependency networks and that more central artifacts (those with a high number of dependents) are not necessarily less vulnerable. We also observed that the time taken to patch vulnerabilities, including those of high or critical severity, often spans several years. Additionally, we found that dependents of artifacts tend to prefer presumably non-vulnerable versions; however, some continue to use vulnerable versions, indicating challenges in adopting patched releases. These findings highlight the critical need for improved dependency management practices and timely vulnerability remediation to enhance the security of software ecosystems.

cs.SE↗

arXiv2025

The Mini-SiTian Array: Prospects of Searching for Tidal Disruption Events

We assess the detectability of tidal disruption events (TDEs) using mock observations from the Mini-SiTian array. We select 100 host galaxy samples from a simulated galaxy catalog based on specific criteria such as redshift, BH mass, and event rate. Taking into account the site conditions and survey strategy, we simulate observations over a 440 deg$^2$ field. The results indicate that $0.53\pm 0.73$ TDEs can be detected per year when observing in both $g$ and $r$ bands with 300-second exposures every 3 days. Applying this method to the SiTian project, we expect to discover approximately 204 TDEs annually, heralding a new era in TDE science.

astro-ph.HE↗

arXiv2025

Developing a Reliable, Fast, General-Purpose Hallucination Detection and Mitigation Service

Hallucination, a phenomenon where large language models (LLMs) produce output that is factually incorrect or unrelated to the input, is a major challenge for LLM applications that require accuracy and dependability. In this paper, we introduce a reliable and high-speed production system aimed at detecting and rectifying the hallucination issue within LLMs. Our system encompasses named entity recognition (NER), natural language inference (NLI), span-based detection (SBD), and an intricate decision tree-based process to reliably detect a wide range of hallucinations in LLM responses. Furthermore, we have crafted a rewriting mechanism that maintains an optimal mix of precision, response time, and cost-effectiveness. We detail the core elements of our framework and underscore the paramount challenges tied to response time, availability, and performance metrics, which are crucial for real-world deployment of these technologies. Our extensive evaluation, utilizing offline data and live production traffic, confirms the efficacy of our proposed framework and service.

cs.CL↗

arXiv2025

Question-Aware Knowledge Graph Prompting for Enhancing Large Language Models

Large Language Models (LLMs) often struggle with tasks requiring external knowledge, such as knowledge-intensive Multiple Choice Question Answering (MCQA). Integrating Knowledge Graphs (KGs) can enhance reasoning; however, existing methods typically demand costly fine-tuning or retrieve noisy KG information. Recent approaches leverage Graph Neural Networks (GNNs) to generate KG-based input embedding prefixes as soft prompts for LLMs but fail to account for question relevance, resulting in noisy prompts. Moreover, in MCQA tasks, the absence of relevant KG knowledge for certain answer options remains a significant challenge. To address these issues, we propose Question-Aware Knowledge Graph Prompting (QAP), which incorporates question embeddings into GNN aggregation to dynamically assess KG relevance. QAP employs global attention to capture inter-option relationships, enriching soft prompts with inferred knowledge. Experimental results demonstrate that QAP outperforms state-of-the-art methods across multiple datasets, highlighting its effectiveness.

cs.CL↗

arXiv2025

Multimodal Feature-Driven Deep Learning for the Prediction of Duck Body Dimensions and Weight

Accurate body dimension and weight measurements are critical for optimizing poultry management, health assessment, and economic efficiency. This study introduces an innovative deep learning-based model leveraging multimodal data-2D RGB images from different views, depth images, and 3D point clouds-for the non-invasive estimation of duck body dimensions and weight. A dataset of 1,023 Linwu ducks, comprising over 5,000 samples with diverse postures and conditions, was collected to support model training. The proposed method innovatively employs PointNet++ to extract key feature points from point clouds, extracts and computes corresponding 3D geometric features, and fuses them with multi-view convolutional 2D features. A Transformer encoder is then utilized to capture long-range dependencies and refine feature interactions, thereby enhancing prediction robustness. The model achieved a mean absolute percentage error (MAPE) of 6.33% and an R2 of 0.953 across eight morphometric parameters, demonstrating strong predictive capability. Unlike conventional manual measurements, the proposed model enables high-precision estimation while eliminating the necessity for physical handling, thereby reducing animal stress and broadening its application scope. This study marks the first application of deep learning techniques to poultry body dimension and weight estimation, providing a valuable reference for the intelligent and precise management of the livestock industry with far-reaching practical significance.

cs.CV↗

arXiv2025

Uncertainty-Instructed Structure Injection for Generalizable HD Map Construction

Reliable high-definition (HD) map construction is crucial for the driving safety of autonomous vehicles. Although recent studies demonstrate improved performance, their generalization capability across unfamiliar driving scenes remains unexplored. To tackle this issue, we propose UIGenMap, an uncertainty-instructed structure injection approach for generalizable HD map vectorization, which concerns the uncertainty resampling in statistical distribution and employs explicit instance features to reduce excessive reliance on training data. Specifically, we introduce the perspective-view (PV) detection branch to obtain explicit structural features, in which the uncertainty-aware decoder is designed to dynamically sample probability distributions considering the difference in scenes. With probabilistic embedding and selection, UI2DPrompt is proposed to construct PV-learnable prompts. These PV prompts are integrated into the map decoder by designed hybrid injection to compensate for neglected instance structures. To ensure real-time inference, a lightweight Mimic Query Distillation is designed to learn from PV prompts, which can serve as an efficient alternative to the flow of PV branches. Extensive experiments on challenging geographically disjoint (geo-based) data splits demonstrate that our UIGenMap achieves superior performance, with +5.7 mAP improvement on the nuScenes dataset. Source code will be available at https://github.com/xiaolul2/UIGenMap.

cs.CV↗

arXiv2025

A Survey of Deep Graph Learning under Distribution Shifts: from Graph Out-of-Distribution Generalization to Adaptation

Distribution shifts on graphs -- the discrepancies in data distribution between training and employing a graph machine learning model -- are ubiquitous and often unavoidable in real-world scenarios. These shifts may severely deteriorate model performance, posing significant challenges for reliable graph machine learning. Consequently, there has been a surge in research on graph machine learning under distribution shifts, aiming to train models to achieve satisfactory performance on out-of-distribution (OOD) test data. In our survey, we provide an up-to-date and forward-looking review of deep graph learning under distribution shifts. Specifically, we cover three primary scenarios: graph OOD generalization, training-time graph OOD adaptation, and test-time graph OOD adaptation. We begin by formally formulating the problems and discussing various types of distribution shifts that can affect graph learning, such as covariate shifts and concept shifts. To provide a better understanding of the literature, we introduce a systematic taxonomy that classifies existing methods into model-centric and data-centric approaches, investigating the techniques used in each category. We also summarize commonly used datasets in this research area to facilitate further investigation. Finally, we point out promising research directions and the corresponding challenges to encourage further study in this vital domain. We also provide a continuously updated reading list at https://github.com/kaize0409/Awesome-Graph-OOD.

cs.LG↗

arXiv2025

Reasoning of Large Language Models over Knowledge Graphs with Super-Relations

While large language models (LLMs) have made significant progress in processing and reasoning over knowledge graphs, current methods suffer from a high non-retrieval rate. This limitation reduces the accuracy of answering questions based on these graphs. Our analysis reveals that the combination of greedy search and forward reasoning is a major contributor to this issue. To overcome these challenges, we introduce the concept of super-relations, which enables both forward and backward reasoning by summarizing and connecting various relational paths within the graph. This holistic approach not only expands the search space, but also significantly improves retrieval efficiency. In this paper, we propose the ReKnoS framework, which aims to Reason over Knowledge Graphs with Super-Relations. Our framework's key advantages include the inclusion of multiple relation paths through super-relations, enhanced forward and backward reasoning capabilities, and increased efficiency in querying LLMs. These enhancements collectively lead to a substantial improvement in the successful retrieval rate and overall reasoning performance. We conduct extensive experiments on nine real-world datasets to evaluate ReKnoS, and the results demonstrate the superior performance of ReKnoS over existing state-of-the-art baselines, with an average accuracy gain of 2.92%.

cs.LG↗

arXiv2025

Impact of Spectral Resolution on S-index and Its Application to Spectroscopic Surveys

Utilizing the PHOENIX synthetic spectra, we investigated the impact of spectral resolution on the calculation of $S$-indices. We found that for spectra with a resolution lower than $\approx$30,000, it is crucial to calibrate $S$-indices for accurate estimations. This is especially essential for low-resolution spectral observations. We provided calibrations for several ongoing or upcoming spectroscopic surveys such as the LAMOST low-resolution survey, the SEGUE survey, the SDSS-V/BOSS survey, the DESI survey, the MSE survey, and the MUST survey. Using common targets between the HARPS and MWO observations, we established conversions from spectral $S$-indices to the well-known $S_{\rm MWO}$ values, applicable to stars with [Fe/H] values greater than $-$1. These calibrations offer a reliable approach to convert $S$-indices obtained from various spectroscopic surveys into $S_{\rm{MWO}}$ values and can be widely applied in studies on chromospheric activity.

astro-ph.SR↗

arXiv2025

Searching for accreting compact binary systems from spectroscopy and photometry: Application to LAMOST spectra

Compact objects undergoing mass transfer exhibit significant (and double-peaked) $H_α$ emission lines. Recently, new methods have been developed to identify black hole X-ray binaries (BHXBs) and calculate their systematic parameters using $H_α$ line parameters, such as the full-width at half maximum (FWHM), equivalent width (EW), and separation of double peaks. In addition, the FWHM-EW plane from spectroscopy and the $H_α$ color-color diagram from photometry can be used for rapid stellar classification. We measure the $H_α$ and $H_β$ profiles (e.g., FWHM and EW) using the LAMOST DR9 low- and medium-resolution spectra, and calculate the systematic parameters (e.g., velocity semi-amplitude of the donor star, mass ratio, inclination angle, and mass of the accretor). A new correlation between FWHM and $K_{\rm 2}$, $K_{\rm 2} = 0.205(18)\ \rm{FWHM}$, is obtained for cataclysmic variables (CVs) in our sample. Both the FWHM-EW plane and the $H_α$ color-color diagram can distinguish CVs with FWHM $\gtrsim$ 1000 km/s from Be stars and young stellar objects (YSOs) to some extent. To improve classification accuracy and enhance the selection of compact objects, we propose a new set of idealized filters with effective widths of 30 Å, 130 Å, and 400 Å\ for the narrow $H_α$ filter, broad $H_α$ filter, and $r$-band filter, respectively.

astro-ph.HE↗

arXiv2025

Bias Unveiled: Investigating Social Bias in LLM-Generated Code

Large language models (LLMs) have significantly advanced the field of automated code generation. However, a notable research gap exists in evaluating social biases that may be present in the code produced by LLMs. To solve this issue, we propose a novel fairness framework, i.e., Solar, to assess and mitigate the social biases of LLM-generated code. Specifically, Solar can automatically generate test cases for quantitatively uncovering social biases of the auto-generated code by LLMs. To quantify the severity of social biases in generated code, we develop a dataset that covers a diverse set of social problems. We applied Solar and the crafted dataset to four state-of-the-art LLMs for code generation. Our evaluation reveals severe bias in the LLM-generated code from all the subject LLMs. Furthermore, we explore several prompting strategies for mitigating bias, including Chain-of-Thought (CoT) prompting, combining positive role-playing with CoT prompting and dialogue with Solar. Our experiments show that dialogue with Solar can effectively reduce social bias in LLM-generated code by up to 90%. Last, we make the code and data publicly available is highly extensible to evaluate new social problems.

cs.SE↗

arXiv2025

PianoMotion10M: Dataset and Benchmark for Hand Motion Generation in Piano Performance

Recently, artificial intelligence techniques for education have been received increasing attentions, while it still remains an open problem to design the effective music instrument instructing systems. Although key presses can be directly derived from sheet music, the transitional movements among key presses require more extensive guidance in piano performance. In this work, we construct a piano-hand motion generation benchmark to guide hand movements and fingerings for piano playing. To this end, we collect an annotated dataset, PianoMotion10M, consisting of 116 hours of piano playing videos from a bird's-eye view with 10 million annotated hand poses. We also introduce a powerful baseline model that generates hand motions from piano audios through a position predictor and a position-guided gesture generator. Furthermore, a series of evaluation metrics are designed to assess the performance of the baseline model, including motion similarity, smoothness, positional accuracy of left and right hands, and overall fidelity of movement distribution. Despite that piano key presses with respect to music scores or audios are already accessible, PianoMotion10M aims to provide guidance on piano fingering for instruction purposes. The source code and dataset can be accessed at https://github.com/agnJason/PianoMotion10M.

cs.SD↗

arXiv2025

"Frog-eyes" in Astronomy: Monitoring Binary Radial Velocity Variations Through A Pair of Narrow-Band Filters

Spectroscopic observations are a crucial step in driving major discoveries in the era of time-domain surveys. However, the pace of current spectroscopic surveys is increasingly unable to meet the demands of rapidly advancing large-scale time-domain surveys. To address this issue, we propose the ``Frog-eyes" system, which employs a pair of narrow-band filters: one positioned near a strong absorption line to capture signals from Doppler shifts, and the other placed on the adjacent continuum to monitor intrinsic variations. The combination of observations from the two filters enables the extraction of radial velocity (RV) curves from a large sample of binary stars, and is particularly efficient for single-lined binaries (SB1), using photometric techniques. Comprehensive mock simulations on SB1 demonstrate that the binary orbital parameters can be precisely measured from the extracted RV curves for binary systems where the primary star has an effective temperature greater than 6000 K. With a typical ground-based photometric precision of approximately 0.3%, the uncertainties in the derived semi-amplitude K and eccentricity e are less than 10% and 0.1, respectively, for binary systems with K $\ge$ 30 km/s. These encouraging results are further validated by real observations of the hot subdwarf-white dwarf binary system HD 265435, using a non-specialized ``Frog-eyes" system installed on the Chinese 2.16m telescope. Once this system is properly installed on large-field-of-view survey telescopes, the rate of acquiring RV curves for binaries will approach their detection rate in leading time-domain photometric surveys.

astro-ph.SR↗