arXiv subjects

Jun Gao

Publications and source records attributed to Jun Gao.

At least 91 records · Page 5Linked to original sources

arXiv2026

Counting cliques without generalized theta graphs

The \textit{generalized Turán number} $\mathrm{ex}(n, T, F)$ is the maximum possible number of copies of $T$ in an $F$-free graph on $n$ vertices for any two graphs $T$ and $F$. For the book graph $B_t$, there is a close connection between $\ex(n,K_3,B_t)$ and the Ruzsa-Szemerédi triangle removal lemma. Motivated by this, in this paper, we study the generalized Turán problem for generalized theta graphs, a natural extension of book graphs. Our main result provides a complete characterization of the magnitude of $\ex(n,K_3,H)$ when $H$ is a generalized theta graph, indicating when it is quadratic, when it is nearly quadratic, and when it is subquadratic. Furthermore, as an application, we obtain the exact value of $\ex(n, K_r, kF)$, where $F$ is an edge-critical generalized theta graph, and $3\le r\le k+1$, extending several recent results.

math.CO↗

arXiv2026

Rethinking Flow and Diffusion Bridge Models for Speech Enhancement

Flow matching and diffusion bridge models have emerged as leading paradigms in generative speech enhancement, modeling stochastic processes between paired noisy and clean speech signals based on principles such as flow matching, score matching, and Schrödinger bridge. In this paper, we present a framework that unifies existing flow and diffusion bridge models by interpreting them as constructions of Gaussian probability paths with varying means and variances between paired data. Furthermore, we investigate the underlying consistency between the training/inference procedures of these generative models and conventional predictive models. Our analysis reveals that each sampling step of a well-trained flow or diffusion bridge model optimized with a data prediction loss is theoretically analogous to executing predictive speech enhancement. Motivated by this insight, we introduce an enhanced bridge model that integrates an effective probability path design with key elements from predictive paradigms, including improved network architecture, tailored loss functions, and optimized training strategies. Experiments on denoising and dereverberation tasks demonstrate that the proposed method outperforms existing flow and diffusion baselines with fewer parameters and reduced computational complexity. The results also highlight that the inherently predictive nature of this generative framework imposes limitations on its achievable upper-bound performance.

eess.AS↗

arXiv2026

Amber-Image: Efficient Compression of Large-Scale Diffusion Transformers

Diffusion Transformer (DiT) architectures have significantly advanced Text-to-Image (T2I) generation but suffer from prohibitive computational costs and deployment barriers. To address these challenges, we propose an efficient compression framework that transforms the 60-layer dual-stream MMDiT-based Qwen-Image into lightweight models without training from scratch. Leveraging this framework, we introduce Amber-Image, a series of streamlined T2I models. We first derive Amber-Image-10B using a timestep-sensitive depth pruning strategy, where retained layers are reinitialized via local weight averaging and optimized through layer-wise distillation and full-parameter fine-tuning. Building on this, we develop Amber-Image-6B by introducing a hybrid-stream architecture that converts deep-layer dual streams into a single stream initialized from the image branch, further refined via progressive distillation and lightweight fine-tuning. Our approach reduces parameters by 70% and eliminates the need for large-scale data engineering. Notably, the entire compression and training pipeline-from the 10B to the 6B variant-requires fewer than 2,000 GPU hours, demonstrating exceptional cost-efficiency compared to training from scratch. Extensive evaluations on benchmarks like DPG-Bench and LongText-Bench show that Amber-Image achieves high-fidelity synthesis and superior text rendering, matching much larger models.

cs.CV↗

arXiv2026

Hello-Chat: Towards Realistic Social Audio Interactions

Recent advancements in Large Audio Language Models (LALMs) have demonstrated exceptional performance in speech recognition and translation. However, existing models often suffer from a disconnect between perception and expression, resulting in a robotic "read-speech" style that lacks the spontaneity and emotional resonance of real human interaction. In this report, we introduce Hello-Chat, an end-to-end audio language model designed for realistic social scenarios. By leveraging a massive dataset of real-life conversations and employing a modality-interleaved training strategy, Hello-Chat achieves a breakthrough in anthropomorphic generation. Experimental results show that our model not only reaches state-of-the-art (SOTA) performance on specific audio understanding tasks but also significantly outperforms existing baselines in prosodic naturalness and emotional alignment, paving the way for the next generation of empathetic AI agents.

cs.SD↗

arXiv2026

SoulX-FlashHead: Oracle-guided Generation of Infinite Real-time Streaming Talking Heads

Achieving a balance between high-fidelity visual quality and low-latency streaming remains a formidable challenge in audio-driven portrait generation. Existing large-scale models often suffer from prohibitive computational costs, while lightweight alternatives typically compromise on holistic facial representations and temporal stability. In this paper, we propose SoulX-FlashHead, a unified 1.3B-parameter framework designed for real-time, infinite-length, and high-fidelity streaming video generation. To address the instability of audio features in streaming scenarios, we introduce Streaming-Aware Spatiotemporal Pre-training equipped with a Temporal Audio Context Cache mechanism, which ensures robust feature extraction from short audio fragments. Furthermore, to mitigate the error accumulation and identity drift inherent in long-sequence autoregressive generation, we propose Oracle-Guided Bidirectional Distillation, leveraging ground-truth motion priors to provide precise physical guidance. We also present VividHead, a large-scale, high-quality dataset containing 782 hours of strictly aligned footage to support robust training. Extensive experiments demonstrate that SoulX-FlashHead achieves state-of-the-art performance on HDTF and VFHQ benchmarks. Notably, our Lite variant achieves an inference speed of 96 FPS on a single NVIDIA RTX 4090, facilitating ultra-fast interaction without sacrificing visual coherence.

cs.CV↗

arXiv2026

Improving Medical Visual Reinforcement Fine-Tuning via Perception and Reasoning Augmentation

While recent advances in Reinforcement Fine-Tuning (RFT) have shown that rule-based reward schemes can enable effective post-training for large language models, their extension to cross-modal, vision-centric domains remains largely underexplored. This limitation is especially pronounced in the medical imaging domain, where effective performance requires both robust visual perception and structured reasoning. In this work, we address this gap by proposing VRFT-Aug, a visual reinforcement fine-tuning framework tailored for the medical domain. VRFT-Aug introduces a series of training strategies designed to augment both perception and reasoning, including prior knowledge injection, perception-driven policy refinement, medically informed reward shaping, and behavioral imitation. Together, these methods aim to stabilize and improve the RFT process. Through extensive experiments across multiple medical datasets, we show that our approaches consistently outperform both standard supervised fine-tuning and RFT baselines. Moreover, we provide empirically grounded insights and practical training heuristics that can be generalized to other medical image tasks. We hope this work contributes actionable guidance and fresh inspiration for the ongoing effort to develop reliable, reasoning-capable models for high-stakes medical applications.

cs.CV↗

arXiv2026

DB3 Team's Solution For Meta KDD Cup' 25

This paper presents the db3 team's winning solution for the Meta CRAG-MM Challenge 2025 at KDD Cup'25. Addressing the challenge's unique multi-modal, multi-turn question answering benchmark (CRAG-MM), we developed a comprehensive framework that integrates tailored retrieval pipelines for different tasks with a unified LLM-tuning approach for hallucination control. Our solution features (1) domain-specific retrieval pipelines handling image-indexed knowledge graphs, web sources, and multi-turn conversations; and (2) advanced refusal training using SFT, DPO, and RL. The system achieved 2nd place in Task 1, 2nd place in Task 2, and 1st place in Task 3, securing the grand prize for excellence in ego-centric queries through superior handling of first-person perspective challenges.

cs.IR↗

arXiv2026

Balancing Fidelity and Plasticity: Aligning Mixed-Precision Fine-Tuning with Linguistic Hierarchies

Deploying and fine-tuning Large Language Models (LLMs) on resource-constrained edge devices requires navigating a strict trade-off between memory footprint and task performance. While Quantization-Aware Fine-tuning has emerged as a viable solution, existing paradigms typically decouple quantization and adapter optimization. This separation overlooks a fundamental theoretical constraint we identify as the \textit{Fidelity-Plasticity Trade-off}: a layer's capacity to adapt to new tasks (Plasticity) is inherently constrained by the information capacity of its frozen weights (Fidelity). Aggressively quantizing semantically critical layers creates an information bottleneck that no amount of adapter rank can recover, while high precision in robust syntactic layers wastes valuable memory. To address this, we introduce \textbf{QR-Adaptor}, a unified framework that jointly optimizes per-layer quantization bit-width and LoRA rank. By formulating resource allocation as a multi-objective search aligned with the model's linguistic hierarchy, our method systematically liberates memory from redundancy-heavy layers to reinvest in capacity-critical ones. Extensive experiments demonstrate that QR-Adaptor establishes a new Pareto frontier: notably, a model fine-tuned under a strict 4-bit memory budget achieves performance rivaling 16-bit baselines, demonstrating that precise resource alignment is as critical as model size.

cs.LG↗

arXiv2026

Rational codegree Turán density of hypergraphs

Let $H$ be a $k$-graph (i.e. a $k$-uniform hypergraph). Its minimum codegree $δ_{k-1}(H)$ is the largest integer $t$ such that every $(k-1)$-subset of $V(H)$ is contained in at least $t$ edges of~$H$. The \emph{codegree Turán density} $γ(\mathcal{F})$ of a family $\mathcal{F}$ of $k$-graphs is the infimum of $γ> 0$ such that every $k$-graph $H$ on $n\to\infty$ vertices with $δ_{k-1}(H) \ge (γ+o(1))\, n$ contains some member of $\mathcal{F}$ as a subgraph. We prove that, for every integer $k\ge3$ and every rational number $α\in [0,1)$, there exists a finite family of $k$-graphs $\mathcal{F}$ such that $γ(\mathcal{F})=α$. Also, for every $k \ge 3$, we establish a strong version of non-principality, namely that there are two $k$-graphs $F_1$ and $F_2$ such that the codegree Turán density of $\{F_1,F_2\}$ is strictly smaller than that of each $F_i$. This answers a question of Mubayi and Zhao [J Comb Theory (A) 114 (2007) 1118--1132].

math.CO↗

arXiv2025

Non-Hermitian Exceptional Topology on a Klein Bottle Photonic Circuit

Non-Hermitian physics has unlocked a wealth of unconventional wave phenomena beyond the reach of Hermitian systems, with exceptional points (EPs) driving enhanced sensitivity, nonreciprocal transport, and topological behavior unique to non-Hermitian degeneracies. Here, we present a scalable and reconfigurable silicon photonic integrated circuit capable of emulating arbitrary non-Hermitian time evolution with high precision. Using this programmable platform, we implement a two-band non-Hermitian Hamiltonian defined on a Klein-bottle topology a nonorientable parameter space that enables exceptional phases forbidden on orientable manifolds. Through an on-chip amplitude-and-phase reconstruction protocol, we retrieve the full complex Hamiltonian at multiple points in parameter space and experimentally map the associated Fermi arc where the imaginary eigenvalue gap closes. The orientation of the measured Fermi arc reveals a nontrivial exceptional topology: it implies the presence of same-charge EPs (or an EP monopole) that cannot annihilate locally on the Klein bottle. Our results demonstrate the first photonic realization of exceptional topology on a nonorientable manifold and establish a versatile platform for exploring exotic non-Hermitian and topological models relevant to classical and quantum photonics.

physics.optics↗

arXiv2025

The Semantic Architect: How FEAML Bridges Structured Data and LLMs for Multi-Label Tasks

Existing feature engineering methods based on large language models (LLMs) have not yet been applied to multi-label learning tasks. They lack the ability to model complex label dependencies and are not specifically adapted to the characteristics of multi-label tasks. To address the above issues, we propose Feature Engineering Automation for Multi-Label Learning (FEAML), an automated feature engineering method for multi-label classification which leverages the code generation capabilities of LLMs. By utilizing metadata and label co-occurrence matrices, LLMs are guided to understand the relationships between data features and task objectives, based on which high-quality features are generated. The newly generated features are evaluated in terms of model accuracy to assess their effectiveness, while Pearson correlation coefficients are used to detect redundancy. FEAML further incorporates the evaluation results as feedback to drive LLMs to continuously optimize code generation in subsequent iterations. By integrating LLMs with a feedback mechanism, FEAML realizes an efficient, interpretable and self-improving feature engineering paradigm. Empirical results on various multi-label datasets demonstrate that our FEAML outperforms other feature engineering methods.

cs.LG↗

arXiv2025

The Dynamic Prior: Understanding 3D Structures for Casual Dynamic Videos

Estimating accurate camera poses, 3D scene geometry, and object motion from in-the-wild videos is a long-standing challenge for classical structure from motion pipelines due to the presence of dynamic objects. Recent learning-based methods attempt to overcome this challenge by training motion estimators to filter dynamic objects and focus on the static background. However, their performance is largely limited by the availability of large-scale motion segmentation datasets, resulting in inaccurate segmentation and, therefore, inferior structural 3D understanding. In this work, we introduce the Dynamic Prior (\ourmodel) to robustly identify dynamic objects without task-specific training, leveraging the powerful reasoning capabilities of Vision-Language Models (VLMs) and the fine-grained spatial segmentation capacity of SAM2. \ourmodel can be seamlessly integrated into state-of-the-art pipelines for camera pose optimization, depth reconstruction, and 4D trajectory estimation. Extensive experiments on both synthetic and real-world videos demonstrate that \ourmodel not only achieves state-of-the-art performance on motion segmentation, but also significantly improves accuracy and robustness for structural 3D understanding.

cs.CV↗

arXiv2025

HW-GNN: Homophily-Aware Gaussian-Window Constrained Graph Spectral Network for Social Network Bot Detection

Social bots are increasingly polluting online platforms by spreading misinformation and engaging in coordinated manipulation, posing severe threats to cybersecurity. Graph Neural Networks (GNNs) have become mainstream for social bot detection due to their ability to integrate structural and attribute features, with spectral-based approaches demonstrating particular efficacy due to discriminative patterns in the spectral domain. However, current spectral GNN methods face two limitations: (1) their broad-spectrum fitting mechanisms degrade the focus on bot-specific spectral features, and (2) certain domain knowledge valuable for bot detection, e.g., low homophily correlates with high-frequency features, has not been fully incorporated into existing methods. To address these challenges, we propose HW-GNN, a novel homophily-aware graph spectral network with Gaussian window constraints. Our framework introduces two key innovations: (i) a Gaussian-window constrained spectral network that employs learnable Gaussian windows to highlight bot-related spectral features, and (ii) a homophily-aware adaptation mechanism that injects domain knowledge between homophily ratios and frequency features into the Gaussian window optimization process. Through extensive experimentation on multiple benchmark datasets, we demonstrate that HW-GNN achieves state-of-the-art bot detection performance, outperforming existing methods with an average improvement of 4.3% in F1-score, while exhibiting strong plug-in compatibility with existing spectral GNNs.

cs.SI↗

arXiv2025

Bipartite Turán problems via graph gluing

For graphs $H_1$ and $H_2$, if we glue them by identifying a given pair of vertices $u \in V(H_1)$ and $v \in V(H_2)$, what is the extremal number of the resulting graph $H_1^u \odot H_2^v$? In this paper, we study this problem and show that interestingly it is equivalent to an old question of Erdős and Simonovits on the Zarankiewicz problem. When $H_1, H_2$ are copies of a same bipartite graph $H$ and $u, v$ come from a same part, we prove that $\operatorname{ex}(n, H_1^u \odot H_2^v) = Θ\bigl( \operatorname{ex}(n, H) \bigr)$. As a corollary, we provide a short self-contained disproof of a conjecture of Erdős, which was recently disproved by Janzer.

math.CO↗

arXiv2025

\texttt{ReMind}: Understanding Deductive Code Reasoning in LLMs

Large Language Models (LLMs) have achieved remarkable progress in code-related tasks. Despite their advancement, empirical evidence reveals that they still struggle with \emph{deductive code reasoning}, the ability to reason about the program execution process. While prior studies have recognized this limitation, the underlying causes remain largely underexplored. In this paper, we begin by presenting a comprehensive empirical study that reveals three key challenges undermining deductive code reasoning: (1) an intrinsic gap between generation and reasoning abilities, (2) a consistent bias towards code sources, and (3) weak zero-shot generalization on complex benchmarks. In light of these challenges, we propose \texttt{ReMind}, a multi-agent framework composed of \texttt{Mutator}, \texttt{Executor}, and \texttt{Inspector}. The \texttt{Mutator} generates code variants to mitigate bias towards code sources, the \texttt{Executor} traces variable states step-by-step to expose inconsistency, and the \texttt{Inspector} identifies problematic reasoning steps and provides control-flow refinement to bridge the intrinsic reasoning gap. Through their coordinated collaboration, \texttt{ReMind} systematically identifies and refines reasoning flaws, achieving outstanding performance and enabling robust zero-shot generalization. Extensive experiments on two benchmarks with five LLMs demonstrate the superior advantages of \texttt{ReMind} compared to baseline approaches in deductive code reasoning.

cs.PL↗

arXiv2025

On cliques in hypergraphs

We prove that for any $k \ge 3$, every $k$-uniform hypergraph on $n$ vertices contains at most $n - ω(1)$ different sizes of cliques (maximal complete subgraphs). In particular, the 3-uniform case answers a question of Erdős.

math.CO↗

arXiv2025

NNLO QCD calculation of hadron multiplicities in light-quark jets at lepton colliders

We present the calculation of next-to-next-to-leading-order (NNLO) QCD corrections to hadron multiplicities in light-quark jets at lepton colliders, employing the ``projection-to-Born" (P2B) method implemented in the FMNLO program. Taking the next-to-leading-order result as an example, we rigorously establish the validity of our P2B-based calculation. We then present NNLO predictions for the normalized asymmetry $D_{K^{-}}$ between hadron and antihadron production in light-quark jets and compare them with SLD data. We find that a suppression of these SLD measurements relative to NPC23 predictions for $D_{K^{-}}$ emerges in the intermediate $z_h$ domain ($0.2 \lesssim z_h \lesssim 0.7$). We expect that incorporating these SLD data into global QCD fits will enable improved determination of fragmentation functions.

hep-ph↗

arXiv2025

SI-Bench: Benchmarking Social Intelligence of Large Language Models in Human-to-Human Conversations

As large language models (LLMs) develop anthropomorphic abilities, they are increasingly being deployed as autonomous agents to interact with humans. However, evaluating their performance in realistic and complex social interactions remains a significant challenge. Most previous research built datasets through simulated agent-to-agent interactions, which fails to capture the authentic linguistic styles and relational dynamics found in real human conversations. To address this gap, we introduce SI-Bench, a novel benchmark designed to evaluate aspects of social intelligence in LLMs. Grounded in broad social science theories, SI-Bench contains 2,221 authentic multi-turn dialogues collected from a social networking application. We further selected a subset of 312 dialogues for manual annotation across 8 major models. The experiments show that SOTA models have surpassed the human expert in process reasoning under complex social situations, yet they still fall behind humans in reply quality. Moreover, introducing Chain-of-Thought (CoT) reasoning may degrade the performance of LLMs in social dialogue tasks. All datasets are openly available at https://github.com/SI-Bench/SI-Bench.git.

cs.CL↗