Searcharxiv⌕ Search

arXiv subjects

Jing Gao

Publications and source records attributed to Jing Gao.

At least 19 recordsLinked to original sources

Study of $CP$ violation in $Λ_b^0\rightarrow N^*M$ decays with the final-state rescattering mechanism

In this work, we investigate the charmless nonleptonic two-body $Λ_b$ decays within the framework of final-state rescattering mechanism. Using this framework, we evaluate the relevant hadronic triangle diagrams by performing the loop integrations, retaining both their absorptive and dispersive parts. Based on the established formalism, we analyze the $Λ_b \to N^*(1535,1520)M$ decay processes with $M =K_S, K^*_0(700)$, $f_0(500,980), ρ(770), \bar{K}^{*0}$, $ϕ$, and predict various physical observables, such as their branching ratios, direct and partial-wave $CP$ asymmetries, as well as decay asymmetry parameters. These two-body decay processes are expected to contribute primarily to the subsequent four-body decay channels, such as $Λ_b^0 \to p\,π^-\,π^+\,π^-$, whose $CP$ asymmetry measurements will be accessible at the LHCb experiment.

hep-ph↗

MatOS: An Agent Operating System for Materials Research

Materials research relies on scattered and growing data, knowledge, and simulation tools, and a single study often spans multi-day workflows across literature analysis, computation, and experiment. Large language model agents can now retrieve materials data, operate simulation codes, and close design-synthesis-characterization loops with autonomous laboratories. However, extending these agents with new capabilities, keeping long-running tasks on track, and letting them learn from past tasks and user feedback remain difficult. Here we present MatOS, an agent operating system for materials research with three core components: (1) a layered scientific asset graph that organizes data, knowledge, tools, skills, roles, and workflows as reusable assets; (2) a stateful adaptive workflow that runs tasks as explicit execution graphs with scientific checks, so that they can be resumed and traced; and (3) evidence-driven asset evolution, which revises these assets from execution records and user feedback. On ResearchClawBench, MatOS improves on the previously best-performing agent by 11.15% with the same base model and extends the cost-score Pareto frontier. With a common base model, it outperforms Codex CLI and Claude Code, exceeding Codex CLI by 2.13% on MatTools and 4.17% on SciAgentGym. In eight research cases, MatOS narrowed more than 30 million candidate structures to 97 solid-electrolyte candidates, identified 89 halide variants with higher computed conductivity than their parent structures, linked the force errors of 11 pretrained interatomic potentials to local atomic environments, and supported two new methods that reduce errors by 3.1% and 13.6% relative to their baselines. These results show that organizing research capabilities as reusable, evolving assets enables agents to carry out long, multistage materials studies. Code is available at https://github.com/DingyangLyu/MatOS.

cs.AI↗

A Field-Deployable 2U Ultra-Stable Laser with Sub-Hertz Linewidth Based on Optoelectronic Oscillator Locking

Compact, high-performance ultra-stable lasers that can operate outside the laboratory remain a long-standing goal. Here we report, to our knowledge, the most compact field-deployable ultra-stable laser, packaged in a standard 2U rack chassis. It is stabilized to a 50 mm-long, 19 mL optical reference cavity with low vibration sensitivity by optoelectronic oscillator (OEO) locking, an approach particularly suited to compact, transportable systems. The laser delivers an integrated linewidth of about 0.42 Hz and a fractional frequency stability of about 5E-15 at 1 s averaging time (about 2.2E-15 at 0.4 s). Measured phase noise approaches the cavity thermal-noise limit below 100 Hz and the theoretical OEO-locking limit from 100 Hz to 100 kHz. The system volume is roughly one quarter of the volume of previously reported integrated, transportable ultra-stable lasers of comparable performance, making it well suited for field applications in time-frequency metrology, quantum technologies, and environmental sensing.

physics.optics↗

Cattle-CLIP: A Multimodal Framework for Dairy Cattle Behaviour Recognition from Video

Monitoring cattle behaviour provides valuable insights into animal health, productivity and welfare. However, robust behaviour recognition in real-world farm environments remains challenging because well-annotated livestock video datasets are scarce and agricultural surveillance footage differs substantially from large-scale pretraining data. To address these challenges, we propose Cattle-CLIP, a domain-adaptive vision-language framework based on Contrastive Language-Image Pretraining (CLIP), which reformulates cattle behaviour recognition as a cross-modal semantic alignment task. Cattle-CLIP incorporates a temporal integration module that extends image-level contrastive pretraining to video-based behaviour understanding and promotes consistent semantic alignment across time. To reduce domain discrepancy, customised data augmentation and behaviour-oriented prompts are employed. We also construct CattleBehaviours6, a curated dataset containing 1,905 annotated clips from six indoor behaviours: feeding, drinking, standing-self-grooming, standing-ruminating, lying-self-grooming and lying-ruminating. The dataset provides standardised ethogram definitions for model training, evaluation and future livestock behaviour research. Cattle-CLIP was evaluated under supervised and few-shot learning settings, with particular emphasis on data-constrained behaviour recognition. In supervised experiments, the model achieved 96.1% accuracy across six behaviours and showed high precision and recall for feeding, drinking and standing-ruminating. It also maintained promising few-shot performance with limited labelled data, demonstrating the potential of multimodal learning for cattle behaviour recognition in data-constrained farm environments.

cs.CV↗

Auditing Failure Attribution: Label Predictability and Judge Behavior

Failure-attribution benchmarks ask an LLM judge which step or agent caused a failure. A high score should mean the judge checked the faulty step against the run (a cross-step check). However, spurious signals (surface cues) that predict the annotation can produce a high score. In this paper, we propose four tests to separate checking from cue use. Input restriction asks whether partial input predicts the annotation. Cue following checks whether moved cues change answers. Residual accuracy scores judges where cue-only predictors fail. Aggregation asks whether the failure mix or metric changes rankings. Real observations motivate these tests: on three benchmarks, simple predictors match or beat GPT-4.1. In two Who&When Pro systems, a classifier reading one step's reasoning finds 95-100% of faulty steps, motivating input restriction. In one of them, error injection leaves no reasoning after the faulty step; judges fine-tuned on its labels beat GPT-4.1 by 41 points. Adding task-related reasoning after that step moves this cue: fine-tuned judges lose 31-99 points, but zero-shot judges lose at most 9, motivating cue following. On Aegis, blaming every agent matches GPT-4.1 on micro-F1 but trails by 19 points in exact-set accuracy, motivating aggregation. We recommend reporting these tests when developing failure-attribution benchmarks.

cs.CL↗

Can Language Models Learn to Reject Their Own Bad Reasoning Steps?

Verifier-guided decoding can prevent harmful reasoning steps from contaminating subsequent generation, but typically relies on an external learned verifier. We ask whether a language model can instead reject its own bad reasoning steps. We define a prefix's recoverability as the probability that the frozen generator can complete it correctly. Diagnostics show that adjacent recoverability changes are often difficult to resolve with practical Monte Carlo budgets, while same-prefix candidates exhibit a sparse low-recoverability tail. We introduce Self-Step Rejection (SSR), which trains a lightweight LoRA acceptance gate on the generator backbone while keeping the base model frozen. SSR uses confidence-qualified first-passage supervision: steps before the first resolved crossing of a root-relative recoverability barrier are accepted, the crossing step is rejected, and unresolved steps and suffixes are excluded. Training combines pointwise classification, same-prefix pairwise learning, and group-relative policy refinement using final-answer correctness. At inference, SSR accepts candidates or resamples from the unchanged prefix under rejection budgets, without an external learned verifier. Across three reasoning models and five mathematical reasoning benchmarks, SSR improves macro-average accuracy over single-pass decoding by 5.4--10.1 points using 1.21--1.40x as many generated tokens, and achieves the highest macro-average accuracy among evaluated step-level methods. Full-solution scaling methods require 4.47--8.27x the single-pass token cost for comparable performance.

cs.CL↗

Spectral extremal graphs for $W_5$-free graphs with odd size

For a fixed integer $k\ge 2$, let $W_{2k+1}=K_1\vee C_{2k}$ be an odd wheel graph. The fixed-size spectral extremal problem aims to determine \[ \operatorname{spex}(m,W_{2k+1}):=\max\{ρ(G): e(G)=m,\ G \text{ is } W_{2k+1}\text{-free}\}, \] where $ρ(G)$ denotes the adjacency spectral radius. Based on this problem, Yu, Li, and Peng [12] proposed the following conjecture: When $m-\binom{k}{2}$ is divisible by $k$ and $m$ is large, every $W_{2k+1}$-free graph of size $m$ satisfies \( ρ(G)^2-(k-1)ρ(G)\le m-\binom{k}{2} \) with equality precisely for $K_k\vee qK_1$. For nonzero residue class, Yu, Zhang, and Zhang [13] proposed the following conjecture: Let $r$ be a nonzero remainder when $m-\binom{k}{2}$ is divided by $k$ and $m$ is large. Then $S_{k,m}$ is the unique graph among $W_{2k+1}$-free graphs of size $m$ having maximum spectral radius, where $S_{k,m}$ is obtained from $K_k\vee qK_1$ by adding a vertex $z$ and joining it to exactly $r$ vertices of the $K_k$. Very recently, Fang, Zhai and Zhang [4] confirmed the Yu--Li--Peng conjecture for $k\ge 2$. Chen, Gao and Li [2] confirmed Yu-Zhang-Zhang conjecture for $k\ge 3$. When $k=2$, then $W_5=K_1\vee C_4$. For large odd $m$, determining $\operatorname{spex}(m,W_5)$ is still open. In this paper we address the odd-size problem. Our result disproved Yu-Zhang-Zhang conjecture for $k= 2$. In our proof, a universal defect bound shows that only $O(1)$ edges can lie outside the dense core. Perron localization then reduces this to at most one edge. A discrete defect inequality forces the complete bipartite crossing and quantizes the two matching deficiencies. Exact quotient-polynomial comparisons eliminate the remaining cross-edge and odd--odd candidates.

math.CO↗

The edge spectral extremal problem for $kK_3$ in nonzero residue classes

For a fixed integer $k\ge 2$, let $kK_3$ denote the vertex-disjoint union of $k$ triangles. A recent fixed-size spectral theorem of Das and Yamini asserts that, for all sufficiently large $m$, every $kK_3$-free graph $G$ of size $m$ satisfies $λ(G)\le (k-1)+\sqrt{m-k(k-1)},$ and equality holds if and only if $(2k-1)\mid m$ and $ G\cong \bigl(K_{2k-1}\vee qK_1\bigr)\cup tK_1,\, q=\frac{m}{2k-1}-(k-1) $ for some $t\ge 0$. They explicitly posed the open problem: Let $k\ge 2$ be fixed and $\ell$ be a residue in $\{1,\dots,2k-2\}$. For all sufficiently large integers $m\equiv \ell\pmod{2k-1}$, determine the exact value of $\max\bigl\{λ(G): e(G)=m,\ G\text{ is }kK_3\text{-free}\bigr\}, $ and characterize all graphs attaining this maximum. In this paper, using the positive-defect version of the bounded-core method for divisible sizes together with several new ideas developed in this paper, we give a complete solution to the aforementioned open problem.

math.CO↗

Vision Wormhole: Latent-Space Communication in Heterogeneous Multi-Agent Systems

Heterogeneous multi-agent systems combine models with different capabilities through a common communication interface. Exchanging internal states directly requires translating between model-specific representations and controlling intermediate computation. We introduce the Vision Wormhole, which repurposes the visual input interface of Vision-Language Models (VLMs) for continuous communication between frozen heterogeneous agents. A Universal Visual Codec encodes each sender's latent rollout into a fixed-size message, maps it through a shared reference space, and decodes received messages into the receiver's image-token span. Per-model codecs and affine reference maps form a hub-and-spoke architecture with $O(N)$ components for $N$ models. Each model learns its codec independently through self-distillation on anchor texts, and shared-anchor alignment enables reuse across communication partners. Across four VLM families, six team configurations, and nine reasoning benchmarks, Vision Wormhole improves accuracy by 6.0 percentage points on average over text-mediated MAS and achieves a 1.69$\times$ geometric-mean speedup in batch-normalized end-to-end runtime.

cs.CL↗

Do Proactive Agents Need an LLM to Decide When to Act?

Proactive assistants continuously decide when to intervene and what context should support the intervention. Large language model (LLM) pipelines repeatedly interpret activity histories to make these decisions, paying an inference cost even when the assistant remains silent. We show that a small graph model can handle both decisions and improve the language agents it controls. Our key insight is that user activity has a native graph structure: events involve persistent entities whose recurrence connects interactions over time. Triggering and context selection map directly to predictions on event and entity nodes. We introduce a temporal-graph-learning (TGL) controller that learns these predictions jointly and supplies both outputs in one forward pass. The downstream language agent generates suggestions on triggered events using the activity history and scored entities. At 11.13 ms per event on a GPU server, TGL achieves the highest AUCs among nine trigger architectures and gives approximately $4$--$7\times$ trigger-stage speedups over the two single-forward LLM triggers. A shared TGL model improves F1 across all 14 downstream backbones by a mean of 16.7 points. The controller also runs at 13.99 ms on a consumer laptop with an approximately 220 MiB BF16 resident footprint, bringing effective proactive control to on-device deployment.

cs.CL↗

HeurEvo: Agentic Evolution of Hybrid Solver-Augmented Heuristics for Time-Critical Mathematical Optimization

Recent advances in agentic heuristic design use AI agents and execution feedback to automate algorithm discovery for challenging optimization problems. In many practical settings, high-quality solutions must be obtained under strict runtime constraints, motivating hybrid approaches that combine problem-specific heuristics with powerful mathematical programming solvers. However, existing approaches typically improve heuristic components within predefined procedures or tune solver configurations in isolation. This limits holistic adaptation of where to allocate computation, how to leverage solvers, and how to refine the overall algorithmic structure. To address these limitations, we propose HeurEvo, an automated plan--code--component co-evolution framework that jointly evolves the high-level algorithmic structures, their implementations, and a shared pool of reusable components. A planner determines which algorithmic components to use, how to combine them, and how to allocate runtime across stages, a coder realizes the resulting plan as executable code, while a component evolver updates the shared component pool. Within an island-based evolutionary framework, plans and implementations co-evolve with feedback from an interpreter agent that analyzes execution results and identifies opportunities for improvement. Across diverse combinatorial optimization benchmarks and challenging MIPLIB instances, HeurEvo finds high-quality solutions within tight runtime budgets, often matching or surpassing state-of-the-art optimization solvers given hours or days of computation. On several nonlinear geometry problems such as hexagon packing, it also improves upon the best previously reported results. These results highlight the value of jointly searching over algorithmic structure and implementation for agentic heuristic design.

cs.CL↗

The edge spectral extremal problem for odd wheels in nonzero residue classes

For a fixed integer $k\ge 2$, let $W_{2k+1}=K_1\vee C_{2k}$ be an odd wheel graph. The fixed-size spectral extremal problem aims to determine \[ \operatorname{spex}(m,W_{2k+1}):=\max\{ρ(G): e(G)=m,\ G \text{ is } W_{2k+1}\text{-free}\}, \] where $ρ(G)$ denotes the adjacency spectral radius. Based on this problem, Yu, Li, and Peng [18] proposed the following conjecture: For large $m$, every $W_{2k+1}$-free graph of size $m$ satisfies $ρ(G)^2-(k-1)ρ(G)\le m-\binom{k}{2}$ with equality precisely for $K_k\vee qK_1$ and $m-\binom{k}{2}=kq$. Very recently, Fang, Zhai and Zhang [7] confirmed the Yu--Li--Peng conjecture. When $m$ is large, $k\ge 2$ and $m-\binom{k}{2}$ is not divisible by $k$, the exact solution for the above problem is still open. Regarding this problem, Yu, Zhang, and Zhang [19] proposed the following conjecture: Let $r$ be a nonzero remainder when $m-\binom{k}{2}$ is divided by $k$. Then $S_{k,m}$ is the unique graph among $W_{2k+1}$-free graphs of size $m$ having maximum spectral radius, where $S_{k,m}$ is obtained from $K_k\vee qK_1$ by adding a vertex $z$ and joining it to exactly $r$ vertices of the $K_k$. In this paper we address this problem in each nonzero residue class. Our result completely settles the Yu-Zhang-Zhang conjecture for $k\ge 3$.

math.CO↗

On the Regularization Landscape for the Linear Recommendation Models

Recently, a wide range of recommendation algorithms inspired by deep learning techniques have emerged as the performance leaders on several standard recommendation benchmarks. While these algorithms were built on different DL techniques (e.g., dropouts, autoencoder), they have similar performance and even similar cost functions. This paper studies whether the models' comparable performance are sheer coincidence, or they can be unified under a single framework. We find that all linear performance leaders effectively add only a nuclear-norm based regularizer, or a Frobenius-norm based regularizer. The former ones possess a (surprising) rigid structure that limits the models' predictive power but their solutions are low rank and have closed form. The latter ones are more expressive and more efficient for recommendation but their solutions are either full-rank or require executing hard-to-tune numeric procedures such as ADMM. Along this line of finding, we further propose two low-rank, closed-form solutions, derived from carefully generalizing Frobenius-norm based regularizers. The new solutions get the best of both nuclear-norm and Frobenius-norm world.

cs.AI↗

Post-training Quantization for Hybrid Iterative Generative Models

Iterative Generative Models (IGMs) span autoregressive and diffusion paradigms, and hybrid variants that couple them can achieve remarkable image-generation fidelity. However, their iterative inference incurs substantial computational overhead, making Post-training Quantization (PTQ) appealing for acceleration, while directly applying vanilla PTQ to hybrid IGMs can trigger model collapse. By analyzing these failures, we identify two critical challenges: Excessive Outliers (EOs) in the activations create an irreconcilable trade-off between preserving normal precision and covering EOs, resulting in severe degradation in generation quality; Amplified Anomalies (AAs) arising unpredictably from minor quantization errors, create a mismatch between calibration and inference, thus iteratively triggering model collapse. To address these challenges, we introduce HyGenQ, a PTQ framework for hybrid IGMs. HyGenQ comprises Hierarchical Cluster Decoupling (HCD) and Scaling Recalibration (SR). HCD identifies and decouples outlier channels via a multi-stage clustering process, effectively isolating EOs while maintaining normal value precision, thereby alleviating performance degradation. SR scales AAs beyond Gaussian Bound, thereby avoiding model collapse caused by aggressive truncation. Extensive experiments demonstrate that HyGenQ successfully quantizes representative hybrid IGMs to 8-bit precision (W8A8), significantly outperforming existing baselines and validating its robustness across different model families.

cs.LG↗

TraVEL: Trajectory-Guided Video Embedding Learning for Driving-Video Retrieval

Efficiently retrieving relevant clips from large-scale driving logs is essential for data curation, model development, and safety analysis. Structured and rule-based retrieval systems can explicitly target driving events, but typically require expert-defined rules, auxiliary data, and multi-stage perception pipelines. Multimodal embedding models offer a simpler and more efficient alternative by representing each video with a single searchable vector. However, general-purpose models often rely on shortcuts from static scene context and struggle to distinguish motion-centric events, such as turning left versus right or accelerating versus decelerating. In this work, we study how to adapt a general-purpose multimodal embedding model to driving-video retrieval. We first fine-tune Qwen3-VL-Embedding on paired clips and reasoning traces from nuReasoning using an InfoNCE objective. While this stage substantially improves overall retrieval, caption supervision alone remains insufficient for fine-grained motion understanding. We therefore introduce TraVEL (Trajectory-Guided Video Embedding Learning), a motion-aware fine-tuning framework that uses ego-trajectory similarity as a reward within Group Relative Policy Optimization. Trajectories serve only as privileged training supervision; retrieval still operates on single-vector video embeddings without ego poses, expert rules, or auxiliary perception outputs. We further construct a driving-video retrieval benchmark from nuReasoning. Experiments show that TraVEL improves motion-centric retrieval across model scales: relative to SFT, it raises longitudinal and lateral mAP by 9.8 and 4.7 points at 2B, with corresponding gains of 7.2 and 1.5 points at 8B. TraVEL thus combines physically grounded supervision with efficient embedding-based search.

cs.CV↗

Personalization as Inverse Planning: Learning Latent Design Intents for Agentic Slide Generation via Structural Denoising

Slide design requires personalizing both deck themes and page layouts. Yet, current AI agent-based methods struggle with fine-grained, page-level design. Solely relying on prespecified templates or user verbose instructions, they fail to capture latent design intents, leaving Page-level Slide Personalization (PSP) unresolved. To close this gap, this work formulates PSP as an inverse planning problem. We propose to learn a design intent without assuming any knowledge of the specific executing tools (e.g., PowerPoint, Beamer) being used. However, relinquishing control over these tools makes the problem intractable to optimize end-to-end. To overcome this, we propose SPIRE, a principled framework to solve PSP approximately. By intentionally corrupting the visual structures of clean slides, SPIRE creates a verifiable task to denoise the corruption, whereby two agents learn to collaboratively refine executable designs via reinforcement learning (RL). We present a proof that structural denoising is a consistent surrogate for PSP, and that the multi-agent formulation strictly reduces policy gradient variance in RL. Extensive experiments demonstrate the superiority of SPIRE.

cs.AI↗

Hybrid-Policy Self-Editing for Composable Unstructured Knowledge Editing

Large language models (LLMs) achieve remarkable performance across natural language tasks, yet they are trained on static corpora and their knowledge quickly becomes outdated in a fast-changing world. This motivates knowledge editing (KE), which updates specific knowledge in an LLM without changing unrelated others. Recent works move from structured knowledge triples toward unstructured KE (UKE), where the edit is a free-form passage that may state multiple facts at once. Nonetheless, existing editors inject such a passage yet fail to use it: the edited model can recall the passage, but can neither answer atomic questions about its facts nor compose them into multi-hop reasoning. We attribute this missing property, which we term composability, to editors' passive reliance on the fixed passage as the sole learning source. In response, we cast editing as a proactive self-distillation from a privileged in-context state of the same model, which requires no external supervision. We further reveal that due to the novelty of the injected knowledge, the pre-edited model's own rollouts rarely cover it, which limits the effectiveness of pure on-policy distillation. To close this gap, we propose HPSE, which builds a hybrid rollout that steps in to place missing facts onto the student's own trajectory precisely where its coverage fails, while staying on-policy elsewhere. We theoretically analyze HPSE's advantage over pure on-policy distillation, and empirically establish its plug-and-play improvements across four LLM backbones and two KE editors under various scenarios.

cs.CL↗

τ: Learning Touch-Augmented Vision-Language-Action Models from Future Visual Supervision

Incorporating tactile sensing into Vision-Language-Action (VLA) models holds promise for contact-rich manipulation, where visual observations alone often fail to capture critical cues about physical interactions. However, learning informative tactile representation while effectively adapting it to pretrained VLA models remains challenging under limited task-specific data. Existing methods either focus on instantaneous contact states or model temporal interaction dynamics using 6D wrench sequences, leaving high-dimensional tactile signals underexplored. To address these challenges, we present τ, a touch-augmented VLA framework that learns an action-conditioned spatiotemporal tactile representation from future visual supervision inspired by the Joint-Embedding Predictive Architecture (JEPA), and fuses it with vision-language features for action generation. This supervision operates in latent space and is used only during training, adding no deployment overhead. We also introduce TacAura, a dataset of synchronized vision, proprioception, and vision-based tactile signals across four representative contact-rich manipulation tasks. Experiments show that τ outperforms existing models and generalizes to unseen objects and scenes, delivering improved manipulation performance and robustness. Project Page: https://cocacola-lab.github.io/tau-Page/.

cs.RO↗