arXiv subjects

Kai Gao

Publications and source records attributed to Kai Gao.

At least 19 recordsLinked to original sources

arXiv2026

Reducing Speaker Residual by Considering Pinhole Effect in Voice Anonymization

Voice anonymization aims to protect privacy by suppressing speaker identity while preserving linguistic content and prosody. However, residual speaker attributes in non-identity representations may still increase linkability and weaken privacy protection. To this end, this paper proposes a fine-tuning strategy with a pinhole loss for well-trained voice anonymization frameworks to further reduce residual speaker attributes. Inspired by the pinhole effect, the pinhole loss measures the linkability of anonymized utterances from the same source speaker. By minimizing this loss, linkability is reduced, thereby improving privacy protection. Experiments on multiple anonymization frameworks, pseudo-speaker generation methods, and datasets show improved privacy protection while maintaining utility. Audio samples can be found in https://anonymous.4open.science/r/Pinhole-loss-fine-tunning-4628.

eess.AS↗

arXiv2026

Bridge: Automatically Mining Ecosystem-Scale API Update Mappings and Client Update Instances

Library updates often require adapting client code to API changes. API update mappings that identify relations between legacy and replacement APIs, version transitions that these mappings apply, and client update instances that capture concrete API call changes are essential for developing and evaluating automated library update techniques. Existing library evolution datasets capture only subsets of this information and typically cover few third-party libraries. In this paper, we present Bridge, a client-driven framework for automatically constructing ecosystem-scale library update datasets that connect API update mappings, version transitions, and client update instances. Bridge first mines candidate update instances from client dependency update commits at scale, validates them using library-side evidence, and then derives API update mappings from validated instances. This design grounds each retained mapping in at least one client update instance. On a manually annotated ground truth dataset, Bridge achieves 91.6% precision and 88.7% recall for Java and 90.1% precision and 64.0% recall for Python. Applied to WoC V3, Bridge mines 381,661 Java and 277,259 Python client update instances, representing 18,900 and 4,456 API update mappings across 2,557 and 999 libraries, respectively. The mined mappings exhibit a pronounced long-tail distribution, with most appearing in only a few client update instances. As one application of the dataset, we evaluate four large language models on replacement API recommendation, a key step in library updates. The best recommendation accuracy reaches only 37.1% for Java and 44.4% for Python, and all evaluated models perform substantially better on frequently observed mappings than on mappings observed in only a few client update instances, highlighting the difficulty current LLMs face in recommending replacements for mappings in the long tail.

cs.SE↗

arXiv2026

Unsupervised Multimodal Clustering for Semantics Discovery in Multimodal Utterances

Discovering the semantics of multimodal utterances is essential for understanding human language and enhancing human-machine interactions. Existing methods manifest limitations in leveraging nonverbal information for discerning complex semantics in unsupervised scenarios. This paper introduces a novel unsupervised multimodal clustering method (UMC), making a pioneering contribution to this field. UMC introduces a unique approach to constructing augmentation views for multimodal data, which are then used to perform pre-training to establish well-initialized representations for subsequent clustering. An innovative strategy is proposed to dynamically select high-quality samples as guidance for representation learning, gauged by the density of each sample's nearest neighbors. Besides, it is equipped to automatically determine the optimal value for the top-$K$ parameter in each cluster to refine sample selection. Finally, both high- and low-quality samples are used to learn representations conducive to effective clustering. We build baselines on benchmark multimodal intent and dialogue act datasets. UMC shows remarkable improvements of 2-7\% scores in clustering metrics over state-of-the-art methods, marking the first successful endeavor in this domain. The complete code and data are available at https://github.com/thuiar/UMC.

cs.MM↗

arXiv2026

DuraMark: Duration-Embedded Watermarking in LLM-based TTS

Large language model (LLM)-based text-to-speech (TTS) models have achieved remarkable voice cloning capabilities, raising concerns about potential deepfake misuse. Speech watermarking mitigates this by embedding traceable information into generated speech. Mainstream watermarking methods operate at the signal level (waveform or spectrogram), rendering the watermark vulnerable to generative attacks (e.g., neural codec and vocoder). To address this, we propose DuraMark, a robust information-level watermarking framework. It utilizes syllable duration editing to achieve watermark embedding. Specifically, DuraMark integrates a duration-controllable LLM-based TTS model to edit syllable durations during synthesis, coupled with a duration extractor to extract these durations for detection. Experiments demonstrate DuraMark's superior robustness against generative attacks, significantly outperforming signal-level baselines. Audio samples are available at https://muzw.github.io/duramark_demo/.

eess.AS↗

arXiv2026

The planar Turán number of $\{K_{4},Θ_{6}^{i}\}$

Let $\mathcal{H}$ be a family of graphs. A graph is said to be $\mathcal{H}$-free if it contains no subgraph isomorphic to a graph in $\mathcal{H}$. The planar Turán number $ex_{_\mathcal{P}}(n,\mathcal{H})$ is defined as the maximum number of edges in an $\mathcal{H}$-free planar graph on $n$ vertices. In this paper, we determine the exact value of $ex_{_\mathcal{P}}(n,\{K_{4}, Θ_{6}^{1}\})$ and a tight upper bound of $ex_{_\mathcal{P}}(n,\{K_{4}, Θ_{6}^{2}\})$.

math.CO↗

arXiv2026

A governance horizon for ethical-use constraints in open-weight AI models

Ethical constraints on open-weight AI models are both a reflection of societal concerns and a foundation for AI governance policy. They are expected to propagate to downstream derivatives while implemented as voluntary metadata disclosures that must be restated at each generation of reuse. We audit 2,142,823 model repositories on Hugging Face Hub to test whether this disclosure-based governance infrastructure can sustain traceability across deep model lineages. Restriction evidence decays with a half-life of 1.31 derivation steps ($R^2$=0.98), and beyond seven downstream generations at least 80% of descendant models lack sufficient public evidence for a governance determination, a depth boundary we formalize as the governance horizon. Platform-level interventions to restore missing licence metadata reveal that policy design (not enforcement alone) is the binding factor: inheritance-only designs require near-complete enforcement to move the horizon, whereas a mandatory-declaration design that explicitly resolves orphan lineage components shifts the horizon already at moderate enforcement. The structural bottleneck is lineages with no inheritable upstream intent: such orphan components remain undecidable under any inheritance-only policy regardless of enforcement rate, and unresolved upstream nodes additionally create direct downstream undecidability bottlenecks that inheritance rules alone cannot recover. Comparison with PyPI, where governance signals are carried by explicit machine-readable declarations, corroborates that the collapse is topology-specific to open-weight derivation rather than inherent to open ecosystems. These results establish that disclosure-based governance has a shallow, structurally determined reach in open-weight AI, and that achieving deep supply-chain accountability requires provenance mechanisms propagating governance signals through derivation itself.

cs.AI↗

arXiv2026

Mitigating Multimodal Inconsistency via Cognitive Dual-Pathway Reasoning for Intent Recognition

Multimodal Intent Recognition (MIR) aims to understand complex user intentions by leveraging text, video, and audio signals. However, existing approaches face two key challenges: (1) overlooking intricate cross-modal interactions for distinguishing consistent and inconsistent cues, and (2) ineffectively modeling multimodal conflicts, leading to semantic cancellation. To address these, we propose a novel Cognitive Dual-Pathway Reasoning (CDPR) framework, which constructs a stable semantic foundation via the intuition pathway and mitigates high-level semantic conflicts through the reasoning pathway, cooperatively establishing deep semantic relations. Specifically, we first employ a representation disentanglement strategy to extract modality-invariant and specific features. Subsequently, the intuition pathway aggregates cross-modal consensus using shared features for solid global representations. The reasoning pathway introduces an inconsistency perception mechanism, combining semantic prototype matching with statistical probability calibration to precisely quantify conflict severity, and dynamically adjusting the weights between both pathways. Furthermore, a multi-view loss function is adopted to alleviate modality laziness and learn structured features at different stages. Extensive experiments on two benchmarks show that CDPR achieves SOTA performance and superior robustness in mitigating multimodal inconsistency. The code is available at https://github.com/Hebust-NLP/CDPR.

cs.MM↗

arXiv2026

Understanding Bugs in Template Engine-Based Applications: Symptoms, Root Causes, and Fix Patterns

Template engines are indispensable components in modern software ecosystems, enabling the generation of structured documents and scripts across domains such as web development, Infrastructure as Code, and data engineering. However, the unique architectural characteristics of template engine-based applications (i.e., TE applications), including multi-language composition, opaque data flow, deferred validation, and complex integration, pose significant challenges for diagnosing and resolving bugs in TE applications. While prior research has primarily focused on template engine security, bugs in TE applications remain under-investigated. To bridge this gap, we present the first comprehensive study of TE application bugs. By analyzing 1,004 application bugs across 15 template engines in five programming languages, we identify the symptoms and root causes of TE application bugs and common patterns to fix them. Our findings reveal that Abnormal Rendering Result (e.g., unexpected or blank output) is the most prevalent symptom (48.61%), often manifesting as silent failures that are difficult to diagnose. We identify 17 root causes, with Syntax Misuse, Mismatched Data Context, and Incompatible Integration as the dominant categories. Furthermore, we find that while 67.92% of the bugs are fixed within the template, over 20% require modifications in the host-side logic to resolve data context issues. Based on these findings, we derive actionable implications for tool designers, practitioners, and researchers. To demonstrate the practical utility of our findings, we further develop two prototype tools for the Jinja engine to facilitate the development and debugging of TE applications.

cs.SE↗

arXiv2026

Imitation Game for Adversarial Disillusion with Chain-of-Thought Reasoning in Generative AI

As the cornerstone of artificial intelligence, machine perception confronts a fundamental threat posed by adversarial illusions. These adversarial attacks manifest in two primary forms: deductive illusion, where specific stimuli are crafted based on the victim model's general decision logic, and inductive illusion, where the victim model's general decision logic is shaped by specific stimuli. The former exploits the model's decision boundaries to create a stimulus that, when applied, interferes with its decision-making process. The latter reinforces a conditioned reflex in the model, embedding a backdoor during its learning phase that, when triggered by a stimulus, causes aberrant behaviours. The multifaceted nature of adversarial illusions calls for a unified defence framework, addressing vulnerabilities across various forms of attack. In this study, we propose a disillusion paradigm based on the concept of an imitation game. At the heart of the imitation game lies a multimodal generative agent, steered by chain-of-thought reasoning, which observes, internalises and reconstructs the semantic essence of a sample, liberated from the classic pursuit of reversing the sample to its original state. As a proof of concept, we conduct experimental simulations using a multimodal generative dialogue agent and evaluates the methodology under a variety of attack scenarios. Experimental results demonstrate that the proposed framework consistently neutralises both deductive and inductive adversarial illusions across diverse white-box and black-box attack scenarios.

cs.AI↗

arXiv2026

Hypnopaedia-Aware Machine Unlearning via Psychometrics of Artificial Mental Imagery

Neural backdoors represent insidious cybersecurity loopholes that render learning machinery vulnerable to unauthorised manipulations, potentially enabling the weaponisation of artificial intelligence with catastrophic consequences. A backdoor attack involves the clandestine infiltration of a trigger during the learning process, metaphorically analogous to hypnopaedia, where ideas are implanted into a subject's subconscious mind under the state of hypnosis or unconsciousness. When activated by a sensory stimulus, the trigger evokes a conditioned reflex that directs a machine to mount a predetermined response. In this study, we propose a cybernetic framework for constant surveillance of backdoor threats, driven by the dynamic nature of untrustworthy data sources. We develop a self-aware unlearning mechanism to autonomously detach a machine's behaviour from the backdoor trigger. Through reverse engineering and statistical inference, we detect deceptive patterns and estimate the likelihood of backdoor infection. We employ model inversion to elicit artificial mental imagery, using stochastic processes to disrupt optimisation pathways and avoid convergent but potentially flawed patterns. This is followed by hypothesis analysis, which estimates the likelihood of each potentially malicious pattern as the true trigger and infers the probability of infection. The primary objective of this study is to maintain a stable state of equilibrium between knowledge fidelity and backdoor vulnerability.

cs.CR↗

arXiv2026

The First Issue Matters: Linking Task-Level Characteristics to Long-Term Newcomer Retention in OSS

Sustaining newcomer participation is critical for the long-term health of open-source communities. Although prior research has explored various task recommendation approaches to help newcomers resolve their first-issue, these methods overlook how characteristics of first-issues may influence newcomers' long-term retention, limiting our understanding of whether initial success leads to sustained participation and hindering effective onboarding design. In this paper, we conduct a large-scale empirical study to examine how first-issue characteristics affect newcomer retention. We combine predictive analysis, interpretability techniques, and causal inference to estimate the causal effects of issue characteristics on retention outcomes. The prediction task supports the interpretation and shows that interaction-related characteristics exhibit stronger associations with retention than intrinsic issue attributes. The causal analysis further reveals that issues reported by moderately experienced contributors, accompanied by moderate discussion intensity and participation from project members, and neutral or slightly negative comment sentiment, have higher retention potential. These findings provide actionable insights for OSS maintainers on designing issue management practices that better support long-term newcomer retention.

cs.SE↗

arXiv2026

DeepFAN, a transformer-based deep learning model for human-artificial intelligence collaborative assessment of incidental pulmonary nodules in CT scans: a multi-reader, multi-case trial

The widespread adoption of CT has notably increased the number of detected lung nodules. However, current deep learning methods for classifying benign and malignant nodules often fail to comprehensively integrate global and local features, and most of them have not been validated through clinical trials. To address this, we developed DeepFAN, a transformer-based model trained on over 10K pathology-confirmed nodules and further conducted a multi-reader, multi-case clinical trial to evaluate its efficacy in assisting junior radiologists. DeepFAN achieved diagnostic area under the curve (AUC) of 0.939 (95% CI 0.930-0.948) on an internal test set and 0.954 (95% CI 0.934-0.973) on the clinical trial dataset involving 400 cases across three independent medical institutions. Explainability analysis indicated higher contributions from global than local features. Twelve readers' average performance significantly improved by 10.9% (95% CI 8.3%-13.5%) in AUC, 10.0% (95% CI 8.9%-11.1%) in accuracy, 7.6% (95% CI 6.1%-9.2%) in sensitivity, and 12.6% (95% CI 10.9%-14.3%) in specificity (P<0.001 for all). Nodule-level inter-reader diagnostic consistency improved from fair to moderate (overall k: 0.313 vs. 0.421; P=0.019). In conclusion, DeepFAN effectively assisted junior radiologists and may help homogenize diagnostic quality and reduce unnecessary follow-up of indeterminate pulmonary nodules. Chinese Clinical Trial Registry: ChiCTR2400084624.

cs.CV↗

arXiv2025

Angular dependence of third-order law in anisotropic MHD turbulence

In solar wind turbulence, the energy transfer/dissipation rate is typically estimated using MHD third-order structure functions calculated using spacecraft observations. However, the inherent anisotropy of solar wind turbulence leads to significant variations in structure functions along different observational directions, thereby affecting the accuracy of energy-dissipation rate estimation. An unresolved issue is how to optimise the selection of observation angles under limited directional sampling to improve estimation precision. We conduct a series of MHD turbulence simulations with different mean magnetic field strengths, $ B_0 $. Our analysis of the third-order structure functions reveals that the global energy dissipation rate estimated around a polar angle of $ θ= 60^\circ$ agrees reasonably with the exact one for $ 0 \le B_0/b_{rms} \le 5 $, where $b_{rms}$ denotes the root-mean-square magnetic field fluctuation. The speciality of $60^\circ$ polar angle can be understood by the Mean Value Theorem of Integrals, since the spherical integral of the polar-angle component ($\widetilde{T_θ}$) of the divergence of Yaglom flux is zero, and $\widetilde{T_θ}$ changes sign around 60$^\circ$. Existing theory on the energy flux vector as a function of the polar angle is assessed, and supports the speciality of $60^\circ$ polar angle. The angular dependence of the third-order structure functions is further assessed with virtual spacecraft data analysis. The present results can be applied to measure the turbulent dissipation rates of energy in the solar wind, which are of potential importance to other areas in which turbulence takes place, such as laboratory plasmas and astrophysics.

physics.space-ph↗

arXiv2025

Small Changes, Big Trouble: Demystifying and Parsing License Variants for Incompatibility Detection in the PyPI Ecosystem

Open-source licenses establish the legal foundation for software reuse, yet license variants, including both modified standard licenses and custom-created alternatives, introduce significant compliance complexities. Despite their prevalence and potential impact, these variants are poorly understood in modern software systems, and existing tools do not account for their existence, leading to significant challenges in both effectiveness and efficiency of license analysis. To fill this knowledge gap, we conduct a comprehensive empirical study of license variants in the PyPI ecosystem. Our findings show that textual variations in licenses are common, yet only 2% involve substantive modifications. However, these license variants lead to significant compliance issues, with 10.7% of their downstream dependencies found to be license-incompatible. Inspired by our findings, we introduce LV-Parser, a novel approach for efficient license variant analysis leveraging diff-based techniques and large language models, along with LV-Compat, an automated pipeline for detecting license incompatibilities in software dependency networks. Our evaluation demonstrates that LV-Parser achieves an accuracy of 0.936 while reducing computational costs by 30%, and LV-Compat identifies 5.2 times more incompatible packages than existing methods with a precision of 0.98. This work not only provides the first empirical study into license variants in software packaging ecosystem but also equips developers and organizations with practical tools for navigating the complex landscape of open-source licensing.

cs.SE↗

arXiv2025

A Foundation Model for Material Fracture Prediction

Accurately predicting when and how materials fail is critical to designing safe, reliable structures, mechanical systems, and engineered components that operate under stress. Yet, fracture behavior remains difficult to model across the diversity of materials, geometries, and loading conditions in real-world applications. While machine learning (ML) methods show promise, most models are trained on narrow datasets, lack robustness, and struggle to generalize. Meanwhile, physics-based simulators offer high-fidelity predictions but are fragmented across specialized methods and require substantial high-performance computing resources to explore the input space. To address these limitations, we present a data-driven foundation model for fracture prediction, a transformer-based architecture that operates across simulators, a wide range of materials (including plastic-bonded explosives, steel, aluminum, shale, and tungsten), and diverse loading conditions. The model supports both structured and unstructured meshes, combining them with large language model embeddings of textual input decks specifying material properties, boundary conditions, and solver settings. This multimodal input design enables flexible adaptation across simulation scenarios without changes to the model architecture. The trained model can be fine-tuned with minimal data on diverse downstream tasks, including time-to-failure estimation, modeling fracture evolution, and adapting to combined finite-discrete element method simulations. It also generalizes to unseen materials such as titanium and concrete, requiring as few as a single sample, dramatically reducing data needs compared to standard ML. Our results show that fracture prediction can be unified under a single model architecture, offering a scalable, extensible alternative to simulator-specific workflows.

cs.LG↗

arXiv2025

Analyzing C/C++ Library Migrations at the Package-level: Prevalence, Domains, Targets and Rationals across Seven Package Management Tools

Library migration happens when a library can not meet the project's requirements and is non-trivial to accomplish. To mitigate the problem, substantial efforts have been devoted to understanding its characteristics and recommending alternative libraries, especially for programming language (PL) ecosystems with a central package hosting platform, such as Python (PyPI). However, to the best of our knowledge, understanding of C/C++ library migrations is still lacking, possibly due to challenges resulting from the fragmented and complicated dependency management practices in the C/C++ ecosystem. To bridge this knowledge gap, this paper analyzes 19,943 C/C++ projects that utilize different package management tools and establishes the first C/C++ library migration dataset. Based on the dataset, we investigate the prevalence, domains, target library, and rationale of C/C++ library migrations and compare the results with three widely investigated PLs: Python, JavaScript, and Java. We find that the overall trend in the number of C/C++ library migrations is similar to Java. Migrations across different package management tools are also observed. In C/C++, library migrations mainly occur in GUI, Build, and OS development, but are rare in domains (e.g., Testing and Logging) that dominate library migrations in the three compared PLs. 83.46\% of C/C++ source libraries only have one migration target, suggesting that our library migration dataset could be used directly to recommend migration targets. We find four C/C++-specific migration reasons, such as less compile time and unification of dependency management, revealing the unique dependency management requirements in C/C++ projects. We believe our findings can help C/C++ developers make more informed library migration decisions and shed light on the design of C/C++ library migration tools.

cs.SE↗

arXiv2025

Infinite-Instruct: Synthesizing Scaling Code instruction Data with Bidirectional Synthesis and Static Verification

Traditional code instruction data synthesis methods suffer from limited diversity and poor logic. We introduce Infinite-Instruct, an automated framework for synthesizing high-quality question-answer pairs, designed to enhance the code generation capabilities of large language models (LLMs). The framework focuses on improving the internal logic of synthesized problems and the quality of synthesized code. First, "Reverse Construction" transforms code snippets into diverse programming problems. Then, through "Backfeeding Construction," keywords in programming problems are structured into a knowledge graph to reconstruct them into programming problems with stronger internal logic. Finally, a cross-lingual static code analysis pipeline filters invalid samples to ensure data quality. Experiments show that on mainstream code generation benchmarks, our fine-tuned models achieve an average performance improvement of 21.70% on 7B-parameter models and 36.95% on 32B-parameter models. Using less than one-tenth of the instruction fine-tuning data, we achieved performance comparable to the Qwen-2.5-Coder-Instruct. Infinite-Instruct provides a scalable solution for LLM training in programming. We open-source the datasets used in the experiments, including both unfiltered versions and filtered versions via static analysis. The data are available at https://github.com/xingwenjing417/Infinite-Instruct-dataset

cs.CL↗

arXiv2025

Multimodal Classification and Out-of-distribution Detection for Multimodal Intent Understanding

Multimodal intent understanding is a significant research area that requires effective leveraging of multiple modalities to analyze human language. Existing methods face two main challenges in this domain. Firstly, they have limitations in capturing the nuanced and high-level semantics underlying complex in-distribution (ID) multimodal intents. Secondly, they exhibit poor generalization when confronted with unseen out-of-distribution (OOD) data in real-world scenarios. To address these issues, we propose a novel method for both ID classification and OOD detection (MIntOOD). We first introduce a weighted feature fusion network that models multimodal representations. This network dynamically learns the importance of each modality, adapting to multimodal contexts. To develop discriminative representations for both tasks, we synthesize pseudo-OOD data from convex combinations of ID data and engage in multimodal representation learning from both coarse-grained and fine-grained perspectives. The coarse-grained perspective focuses on distinguishing between ID and OOD binary classes, while the fine-grained perspective not only enhances the discrimination between different ID classes but also captures instance-level interactions between ID and OOD samples, promoting proximity among similar instances and separation from dissimilar ones. We establish baselines for three multimodal intent datasets and build an OOD benchmark. Extensive experiments on these datasets demonstrate that our method significantly improves OOD detection performance with a 3~10% increase in AUROC scores while achieving new state-of-the-art results in ID classification. Data and codes are available at https://github.com/thuiar/MIntOOD.

cs.MM↗