arXiv subjects

Guang Yang

Publications and source records attributed to Guang Yang.

At least 19 recordsLinked to original sources

arXiv2026

Taming the Entropy Cliff: Variable Codebook Size Quantization for Autoregressive Visual Generation

Most discrete visual tokenizers rely on a default design: every position in the sequence shares the same codebook. Such a constant-codebook design can consume the finite-data entropy budget very quickly. We observe that the per-position empirical conditional entropy decays so quickly along the sequence that, after a few positions, the conditional distribution becomes essentially deterministic. On ImageNet with $K=16384$, this happens after only about 2 out of 256 positions, when most prefixes already identify individual images in the dataset. We call this phenomenon the Entropy Cliff. Under full conditional utilization, larger codebooks exhaust the finite-data entropy budget in fewer positions. The empirical entropy decreases more gradually in our language comparison, where the effective entropy per position stays below the codebook capacity. To address this, we propose Variable Codebook Size Quantization (VCQ), where the codebook size $K_t$ grows monotonically along the sequence from $K_{\min}=2$ to $K_{\max}$, leaving the loss function, parameter count, and AR training procedure unchanged. With a vanilla autoregressive Transformer and standard next-token prediction, VCQ-Base achieves gFID w/o CFG 14.80 on ImageNet $256\times256$, compared with 27.98 and 18.16 for constant codebooks of size 16384 and 8192, respectively. Scaled up, it reaches gFID 1.71 with 684M autoregressive parameters, without any extra training techniques such as semantic regularization or causal alignment. The extreme information bottleneck at $K_{\min}=2$ encourages semantic concentration in early encoder features: a linear probe on features from only the first 10 positions reaches 43.8% top-1 accuracy on ImageNet, compared to 27.1% for uniform codebooks. These results show that what matters is not only the total capacity of the codebook, but also how that capacity is distributed and organized.

cs.CV↗

arXiv2026

Agent-Driven Generative Priors for Pathology Image Synthesis and Segmentation

Deep learning methods for pathology are constrained by the limited availability and high cost of large-scale pixel-level annotations. Synthetic data can alleviate this limitation, but existing pipelines often lack reliable quality control and remain weakly connected to downstream analysis. To address these issues, we propose a three-stage framework that integrates automated data screening, generative representation learning, and efficient adaptation to downstream tasks. First, we develop a pathology agent system to assess image quality and filter samples, resulting in TCGA-200K, a dataset containing 217,030 pathology patches selected from the public TCGA repository. Second, we train a generative model on the screened data. In addition to image synthesis, its backbone learns structural and semantic representations that can be transferred to downstream tasks. Third, we adapt the pretrained generative backbone to semantic segmentation using a lightweight task-specific head. We further introduce PathSeg-380K, a semantic segmentation dataset with expert annotations for approximately 380,000 cells across 12 cell types. Experiments on multiple pathology datasets show that the proposed framework achieves state-of-the-art segmentation performance. Our code is publicly available at https://github.com/script-Yang/Agent-GenPath.

cs.CV↗

arXiv2026

Comprehensive study of timing resolution in plastic scintillator detectors with wavelength-shifting fiber and silicon photomultiplier readout

We present a comprehensive study of the timing resolution achievable in plastic scintillator detectors read out through wavelength-shifting (WLS) fibers coupled to silicon photomultipliers (SiPMs), combining a semi-analytical framework, toy Monte Carlo validation, and full Geant4 optical photon simulation. The analytical model traces the complete photon detection chain: scintillation emission, WLS fiber re-emission, optical transit time dispersion, SiPM single-photon time resolution, and electronics quantization. It expresses the timing resolution $\sigt$ as a function of the detected photoelectron yield $\Npe$, scintillator decay constants ($\taur$, $\taud$), WLS re-emission time ($\tauwls$), fiber numerical aperture, detector geometry, and readout electronics parameters. The analytical predictions are validated at two levels. First, toy Monte Carlo simulations ($2\times 10^5$ events per parameter point across 80 grid points spanning 8 fiber types and $\Npe$ from 5 to 200) achieve analytical-to-MC agreement of \add{$1.0000 \pm 0.0018$}. Second, full Geant4 optical photon simulations track the entire scintillation, wavelength-shifting, and fiber transport chain in realistic detector geometries, confirming the analytical timing predictions\add{ to within 10--20\% for $\Npe \gtrsim 10$} and providing first-principles photoelectron yield calibration. A comprehensive parameter scan covering 7 scintillator materials, 8 WLS fiber types, 5 SiPM models, 5 electronics configurations, 3 readout topologies, and 3 boundary conditions produces quantitative design maps and lookup tables for detector optimization.

physics.ins-det↗

arXiv2026

Performance Benchmarks for 2-View and 3-View Fiber-Projection Fine-Grained Particle Detectors

Fine-grained scintillator detectors are critical for precision measurements in nuclear and particle physics, where accurate reconstruction of interaction vertices and secondary particle directions enables separation of signal from background events. A well-known design choice is the fiber readout geometry: traditional 2-View systems use orthogonal X and Y fibers, while next-generation 3-View designs add a third Z-fiber layer that provides unambiguous 3D voxel identification. The 2-View approach suffers from combinatorial ''ghost hits'', which are false 3D candidates arising from fiber projection ambiguities, that degrade reconstruction performance in high-multiplicity events. This paper presents comprehensive simulation benchmarks quantifying the performance difference between 2-View and 3-View geometries across key metrics. We find that the 3-View geometry reduces ghost hits by 30 to 90% depending on event topology, provides robust vertex resolution across complex topologies, and maintains superior angular resolution for shower direction reconstruction. These benchmarks inform the design optimization of future detectors and provide quantitative guidance for reconstruction algorithm development across a broad range of experiments including neutrino physics, rare kaon/pion decays, and collider calorimetry.

physics.ins-det↗

arXiv2026

Chain-of-Thought in Neural Code Generation: From and For Lightweight Language Models

Large Language Models (LLMs) have demonstrated remarkable potential in code generation. The integration of Chain of Thought (CoT) reasoning can further boost their performance. However, current CoT methods often require manual writing or LLMs with over 100 billion parameters to generate, impeding their applicability in resource-constrained scenarios. In this study, we investigate lightweight Language Models (lLMs), which are defined to have fewer than 10 billion parameters. Empirically, we find that most lLMs cannot generate high-quality CoTs when prompted by the few-shot method, but can take advantage of high-quality CoTs generated elsewhere to improve their performance in code generation. Based on these findings, we design a novel approach COTTON which can leverage lLMs to automatically generate CoTs for code generation. We synthesize new datasets and conduct extensive experiments on various benchmarks. The results show that the CoTs generated by COTTON outperform the baselines in terms of automated and human evaluation metrics. In particular, the CoTs generated by COTTON boost various lLMs to achieve higher performance gains than those generated by LLMs such as ChatGLM (130B), and are competitive with those generated by gpt-3.5-turbo (175B). Our study also showcases the potential of lLMs in software engineering applications.

cs.SE↗

arXiv2026

Stable topology and stable compactness

To solve a large class of noncompact problems ubiquitous in random functional analysis and its applications, most of compactness-based fundamental theorems established for normed spaces or locally convex spaces are required to be extended for random normed modules and random locally convex modules in an appropriate way of weakening compactness. To achieve this goal, in this paper we first develop stable set theory by introducing some new set-theoretic tools, then propose a class of topological spaces with a stable structure, called stable topological spaces, and in particular propose and systematically develop the theory of stably compact spaces, which is a natural generalization of that of compact spaces. The resulting theory is of independent interest from the perspective of general topology since its investigation requires an analysis of stratification structures coordinated with partitions of unity in a complete Boolean algebra. Finally, we show the power of the theory of stable compactness by establishing random Banach-Alaoglu, random James and random Banach-Bourbaki-Kakutani-Šmulian theorems.

math.FA↗

arXiv2026

PhyEvolve: Multi-Agent State-Evolution Planning for Physics-Aware Video Generation

Recent advances in text-to-video generation have substantially improved visual fidelity, yet existing models still struggle to understand the physical evolution implicit in user prompts, making it difficult to generate videos with continuous physical state evolution. Here, we propose PhyEvolve, a multi-agent framework that enhances natural-language prompts through structured state-evolution planning. PhyEvolve consists of three agents: a Director, a Physical Evolver, and a Composer. Given a user prompt and an optional first frame, the Director grounds the scene and disentangles physical events from observation changes. The Physical Evolver uses scene evidence to identify physical mechanisms then constructs causally linked state transitions that unfold throughout the video, whether visible or not. The Composer utilizes these transitions to enhance the original prompts. Without retraining or model modification, PhyEvolve works directly with existing generators through prompt replacement. Extensive experiments on four major benchmarks show that PhyEvolve achieves state-of-the-art performance, outperforming the cutting-edge prior agentic baseline by 7.9 percentage points on average. These results demonstrate that explicit causal state-evolution planning improves physical consistency and state continuity in generated videos through natural-language prompt enhancement.

cs.CV↗

arXiv2026

From Debate to Decision: Conformal Social Choice for Safe Multi-Agent Deliberation

Multi-agent debate improves LLM reasoning, yet agreement is not evidence of correctness. When agents converge on a wrong answer through social reinforcement, consensus stopping commits that error to automated action with no recourse. We introduce Conformal Social Choice, a post-hoc layer that converts debate outputs into calibrated act-versus-escalate decisions. Verbalized probability distributions from heterogeneous agents are aggregated via a linear opinion pool and calibrated with split conformal prediction, yielding fixed-round marginal coverage: at any pre-specified round, the correct answer is included with probability ${\geq}\,1{-}α$, without assumptions on individual model calibration. A hierarchical action policy maps singletons to autonomous action and larger sets to human escalation. Across eight MMLU-Pro domains and two additional benchmarks, fixed-round coverage stays close to target. The key finding is not that debate becomes more accurate, but that the conformal layer makes its failures actionable: at the final round, 81.9\% of wrong-consensus cases are flagged by non-singleton sets at $α{=}0.05$. Separately, among cases that reach a singleton before the final round, accuracy reaches 90.0--96.8\% on MMLU-Pro (up to 22.1pp above consensus stopping): a selection effect, not a reasoning improvement. Simple probability- and entropy-threshold rules attain comparable interception; conformal's distinct value is a finite-sample fixed-round coverage guarantee together with explicit candidate sets for escalation. Adaptive first-singleton stopping is evaluated empirically and not covered by this guarantee.

cs.AI↗

arXiv2026

Yuvion VL: A Multimodal Foundation Model for Adversarial Content and AI Safety

General-purpose models often struggle to reliably identify and understand real-world multimodal risks, largely due to the inherent multimodal adversarial nature of content and AI safety. We present Yuvion VL, a family of multimodal large language models purpose-built for content and AI safety, with both instruction-tuned and reasoning-oriented variants. Yuvion VL addresses this gap by treating safety as an inherently adversarial and multimodal problem and designing the entire pipeline around adversarial robustness. For data construction, we develop an automated pipeline integrating adversarial-aware data synthesis with multi-stage quality control, producing large-scale, high-quality multimodal samples augmented with domain knowledge and reasoning annotations. For training, we adopt a three-stage pipeline that includes continued pretraining for risk-concept cross-modal alignment, instruct post-training for production-grade safety tasks, and reasoning post-training for enhanced interpretability and performance in complex tasks. We further introduce Confuse-then-Contrast Fine-Tuning, a contrastive framework that mines model-specific confusions and constructs multi-image contrastive groups to enforce explicit discrimination of fine-grained visual-semantic elements, enabling the model to distinguish between visually similar cases with different safety implications in adversarial safety tasks. To support rigorous evaluation, we further introduce Yuvion VL RiskEval (YVRE), a collection of benchmarks covering diverse open and internal evaluations, with a focus on content and AI safety, adversarial robustness, and real-world capability requirements. Experiments show that Yuvion VL-32B achieves industry-leading safety performance, surpassing comparably sized open-source models and best closed-source commercial models, while maintaining comparable general capabilities.

cs.CV↗

arXiv2026

MemT-1.0: Towards Memory-Native Evolution

Memory-native evolution requires making memory an intrinsic capability of foundation models, enabling them to continuously accumulate and utilize information. To this end, we introduce MemT-1.0, a new architecture of memory foundation models that internalize memory as part of model computation. Building on Metis, MemT-1.0 employs Gated DeltaNet-2 with a stable token-level memory update mechanism within each turn. We further introduce Mixture-of-Storage to scale native memory capacity. Following a deliberate backbone preparation, we develop a two-stage training pipeline consisting of continued pre-training and post-training, supported by 34B tokens of training data and efficient training and inference infrastructure. We release a model family comprising MemT-1.0-Nano and MemT-1.0-Flash, and conduct a comprehensive evaluation of memory and general capabilities. The results show that they outperform the parametric memory baselines on average and remain competitive with retrieval-augmented methods, while largely preserving general capabilities. At a history length of 128K tokens, MemT-1.0-Nano achieves approximately $1.86\times$ end-to-end inference speedup and a $66\times$ reduction in persistent storage relative to full context KV cache. By internalizing retrieved procedural experience, MemT-1.0 improves task success on agentic tasks, providing a practical step toward memory-native evolution.

cs.CL↗

arXiv2026

Move First, Commit Later: Selective LiDAR-to-BIM Global Initialization by Verifying Proposals Against Competing Explanations

Global LiDAR-to-BIM initialization must place a robot within an as-designed building model without a prior pose. In repetitive interiors, the principal failure mode is not low-confidence registration but confident aliasing: a submap can match several BIM regions with comparable scores, producing a high-scoring pose displaced by symmetry. We present Move First, Commit Later, a selective layer that treats a registration front-end as an evidence source and decides whether to commit. Candidates from multiple submaps are mapped to a common $\mathrm{SE}(2)$ anchor; a top-1 consensus $B_m$, invariant to non-champion multiplicity, aggregates cross-submap evidence; and topology serves only as a binary feasibility gate. The decision is typed--COMMIT, DEFER, or AMBIGUOUS($τ$), reporting the detected symmetry period--and reversible: symmetry-breaking motion upgrades AMBIGUOUS to COMMIT. On a real multi-room building and a controlled symmetric simulation, the layer commits correctly in every tested trajectory-scale configuration and otherwise abstains with a typed state, whereas forced-choice policies on the same front-end select wrong rooms in most cases. Committed anchors are within 0.02--0.36 m of an independent laser-tracker position reference and within $1.9^\circ$ of a BIM-registration orientation proxy. The evaluation covers one building and one front-end; the layer is designed to be front-end modular.

cs.RO↗

arXiv2026

Smooth Hard-Thresholding for Singular Values with Stein's Unbiased Risk Estimate

Low-rank matrix denoising is a central primitive in patch-based image restoration and many other inverse problems. Classical SVD-based image denoising methods often choose a truncation rank by matching residual singular-value energy with an estimated noise energy, but this rule is not a finite-sample risk principle because a fitted low-rank approximation inevitably absorbs part of the noise. This paper develops a mathematically rigorous alternative based on Stein's unbiased risk estimate (SURE). Since singular value hard thresholding is discontinuous and does not satisfy the hypotheses of Stein's lemma, we introduce a logistic smooth hard-threshold spectral estimator. We prove that the smooth shrinker satisfies the regularity conditions required by a spectral-estimator version of Stein's lemma, and therefore admits an exactly unbiased fixed-threshold risk estimate under Gaussian noise. For a fixed observed matrix and a finite set of candidate thresholds separated from the observed singular values, the ordering of the fixed-threshold smooth SURE objective eventually agrees with a simple limiting score. The limiting score has the same algebraic form as the biased hard-threshold SURE formula, but here it is used only as a computational device for ranking finite candidates. Selecting the minimizing threshold is a data-adaptive tuning step; the selected SURE value should not be interpreted as an unbiased risk estimate of the finally selected estimator.

math.ST↗

arXiv2026

Data-driven Smooth-to-hard Spectral Regularization for General Linear Inverse Problems

The logistic smooth-to-hard shrinker is the proximal map of an explicit saturating spectral penalty for general linear inverse problems. Its exact weak convexity yields a unique scaled update and proximal-gradient stationarity guarantees, while independent validation provides finite-grid parameter-selection control. We use MRI as an illustrative example.

math.OC↗

arXiv2026

First achievement of 100 million degree electron temperatures in a spherical torus: the EXL-50U

Here,we report an ST plasma in which Thomson scattering measures a central electron temperature exceeding 8.6 keV (100 million degrees kelvin) on the EXL-50U spherical torus device.The high-temperature phase is obtained at low density by sequential operation of two 50 GHz ECRH systems: ordinary-mode (O-mode) ECRH3 followed by extraordinary-mode (X-mode) ECRH4. In contrast to neutral-beam-heated high-ion-temperature ST scenarios, the discharge is sustained by ECRH only.

physics.plasm-ph↗

arXiv2026

Rubric-Based Optimization for Text-to-Music Generation

Post-training text-to-music generation requires reward signals that capture multiple aspects of musical quality beyond what any single automatic metric can measure. We study structured, rubric-based rewards from pretrained audio-language models (ALMs) as training signals for both autoregressive and diffusion-based music generators. An ALM scores each generated clip against the rubric; we rank candidates generated for the same text prompt by their scores and convert these rankings into preference pairs for DPO on both MusicGen-small and ACE-Step v1, and additionally use the rubric scores directly as scalar rewards for DiffusionNFT on ACE-Step v1. On MusicCaps, rubric-based optimization improves CLAP, SongEval, and Audiobox-Aesthetics simultaneously, with the strongest gains obtained by DiffusionNFT on ACE-Step. By contrast, on MusicGen-small, building preferences from any one of these automatic evaluators produces clear cross-metric trade-offs: the targeted evaluator improves while other independent evaluators deteriorate. We further study tempo, key, and instrumentation, where precise objective rewards are available. Directly optimizing these specialized rewards reliably improves the target attributes, whereas ALM rubrics provide only partial transfer for tempo and instrumentation and no measurable improvement for key. Together, these results suggest a practical division of labor: ALM rubrics are effective for broad perceptual qualities that are difficult to formalize, while specialized objective rewards remain preferable when reliable measurements are available.

cs.SD↗

arXiv2026

Adaptive Utilization of Low-Rank Adaptation via Conditioned Gating

Low-Rank Adaptation (LoRA) achieves parameter-efficient fine-tuning by constraining model updates to a low-rank subspace and has been widely used in practice. However, LoRA typically employs a shared low-rank update across tokens, which limits its ability to fully exploit the adaptation subspace for tokens from different sequences. To address this issue, we propose an adaptive utilization of Low-Rank Adaptation (U-LoRA), which employs conditioned gating to explicitly learn effective token-level utilization of the limited low-rank adaptation subspace. Specifically, U-LoRA generates utilization coefficients along low-rank directions for each token and jointly coordinates and constrains them using sequence-level contextual information, thereby inducing more consistent adaptive patterns within a sentence. To further enhance training stability, we introduce a bias-corrected exponential moving average (EMA) historical prior that calibrates utilization signals across optimization steps, suppressing noise caused by batch-to-batch fluctuations. The effectiveness of our method arises from a better utilization of the existing low-rank subspace via input-conditioned strategies, rather than from expanding the subspace. Experiments on mathematical reasoning and natural language understanding benchmarks demonstrate that U-LoRA achieves competitive performance under comparable parameter budgets when with strong LoRA baselines and recent variants.

cs.CL↗

arXiv2026

HERA: Harness-Environment Co-Evolution for Reliable Agentic Abstention

Large language model (LLM) agents are increasingly capable of acting in complex tool-use environments, yet they often fail to recognize when tasks are infeasible and no valid solution exists. Recent work has formalized this reliability gap as the problem of agentic abstention, and existing approaches typically optimize a model or agent harness against a fixed set of tasks, leading to limited generalization to unseen failure modes. We introduce HERA, a framework for harness-environment co-evolution for agentic abstention. HERA consists of (i) a pipeline to automatically construct verifiable pairs of feasible and infeasible tasks by applying controlled environment mutations that transform solvable tasks into cases requiring abstention, and (ii) a co-evolution procedure in which performance failures on previous tasks are used to drive harness adaptation and generate new execution environments and tasks geared towards previous weaknesses. On held-out evaluation tasks, an evolved harness from HERA improves abstention accuracy from 61.7% to 83.3% while improving feasible-task completion from 68.3% to 76.7%, achieving the highest abstention and feasible-task completion among the compared methods. The resulting best harness transfers across 19 other LLMs, improving abstention accuracy by 15.3 percentage points on average without any model-specific optimization, and enabling smaller models to match the performance of more powerful models at an estimated 85% lower cost.

cs.AI↗

arXiv2026

GenoTrace: Inheritable Watermarks for Genome Foundation Model Distillation

Can a genome model retain a detectable record of the synthetic sequences used to train it? We study watermark inheritance through distillation with GenoTrace, a codon-aware extension of green-list watermarking. Two token-level factors modulate the teacher's generation bias using codon position and organism-specific codon usage. The resulting sequences train a smaller student, whose outputs are audited without an active watermark processor. In a three-seed GenomeOcean-500M-to-100M experiment, the joint configuration achieves a mean audit score of 17.88 and 94.5% detection at a fixed threshold. It retains 49.0% detection after key-aware token substitution, compared with 0% for the available single-seed plain-watermark comparator, and 47.0% after combined mechanism-targeted nucleotide edits. Additional experiments establish inherited signal across five organism-conditioned datasets and teacher-student size ratios up to 40. Component ablations and computational sequence-quality assays reveal distinct operating points for detection strength and coding coverage. GenoTrace provides a practical token-level construction and an empirical account of how genomic structure shapes inherited watermark signals. The findings concern shared-tokenizer distillation and the tested editing procedures, with calibration and biological utility treated as separate evaluation requirements.

cs.CR↗