arXiv subjects

Yang Gao

Publications and source records attributed to Yang Gao.

At least 109 records · Page 6Linked to original sources

arXiv2026

Seer: Language Instructed Video Prediction with Latent Diffusion Models

Imagining the future trajectory is the key for robots to make sound planning and successfully reach their goals. Therefore, text-conditioned video prediction (TVP) is an essential task to facilitate general robot policy learning. To tackle this task and empower robots with the ability to foresee the future, we propose a sample and computation-efficient model, named \textbf{Seer}, by inflating the pretrained text-to-image (T2I) stable diffusion models along the temporal axis. We enhance the U-Net and language conditioning model by incorporating computation-efficient spatial-temporal attention. Furthermore, we introduce a novel Frame Sequential Text Decomposer module that dissects a sentence's global instruction into temporally aligned sub-instructions, ensuring precise integration into each frame of generation. Our framework allows us to effectively leverage the extensive prior knowledge embedded in pretrained T2I models across the frames. With the adaptable-designed architecture, Seer makes it possible to generate high-fidelity, coherent, and instruction-aligned video frames by fine-tuning a few layers on a small amount of data. The experimental results on Something Something V2 (SSv2), Bridgedata and EpicKitchens-100 datasets demonstrate our superior video prediction performance with around 480-GPU hours versus CogVideo with over 12,480-GPU hours: achieving the 31% FVD improvement compared to the current SOTA model on SSv2 and 83.7% average preference in the human evaluation.

cs.CV↗

arXiv2026

Reinforcement Learning with Foundation Priors: Let the Embodied Agent Efficiently Learn on Its Own

Reinforcement learning (RL) is a promising approach for solving robotic manipulation tasks. However, it is challenging to apply the RL algorithms directly in the real world. For one thing, RL is data-intensive and typically requires millions of interactions with environments, which are impractical in real scenarios. For another, it is necessary to make heavy engineering efforts to design reward functions manually. To address these issues, we leverage foundation models in this paper. We propose Reinforcement Learning with Foundation Priors (RLFP) to utilize guidance and feedback from policy, value, and success-reward foundation models. Within this framework, we introduce the Foundation-guided Actor-Critic (FAC) algorithm, which enables embodied agents to explore more efficiently with automatic reward functions. The benefits of our framework are threefold: (1) \textit{sample efficient}; (2) \textit{minimal and effective reward engineering}; (3) \textit{agnostic to foundation model forms and robust to noisy priors}. Our method achieves remarkable performances in various manipulation tasks on both real robots and in simulation. Across 5 dexterous tasks with real robots, FAC achieves an average success rate of 86\% after one hour of real-time learning. Across 8 tasks in the simulated Meta-world, FAC achieves 100\% success rates in 7/8 tasks under less than 100k frames (about 1-hour training), outperforming baseline methods with manual-designed rewards in 1M frames. We believe the RLFP framework can enable future robots to explore and learn autonomously in the physical world for more tasks. Visualizations and code are available at https://yewr.github.io/rlfp.

cs.RO↗

arXiv2026

Adaptive Forensic Feature Refinement via Intrinsic Importance Perception

With the rapid development of generative models and multimodal content editing technologies, the key challenge faced by synthetic image detection (SID) lies in cross-distribution generalization to unknown generation sources. In recent years, visual foundation models (VFM), which acquire rich visual priors through large scale image-text alignment pretraining, have become a promising technical route for improving the generalization ability of SID. However, existing VFM-based methods remain relatively coarse-grained in their adaptation strategies. They typically either directly use the final layer representations of VFM or simply fuse multi layer features, lacking explicit modeling of the optimal representational hierarchy for transferable forgery cues. Meanwhile, although directly fine-tuning VFM can enhance task adaptation, it may also damage the cross-modal pretrained structure that supports open-set generalization. To address this task specific tension, we reformulate VFM adaptation for SID as a joint optimization problem: it is necessary both to identify the critical representational layer that is more suitable for carrying forgery discriminative information and to constrain the disturbance caused by task knowledge injection to the pretrained structure. Based on this, we propose I2P, an SID framework centered on intrinsic importance perception. I2P first adaptively identifies the critical layer representations that are most discriminative for SID, and then constrains task-driven parameter updates within a low sensitivity parameter subspace, thereby improving task specificity while preserving the transferable structure of pretrained representations as much as possible.

cs.CV↗

arXiv2026

Cylindrical Metasurface for Efficient Traveling-wave MRI at 7 T

This research focuses on the design and evaluation of an ultrathin cylindrical metasurface for improving the transmit efficiency of traveling-wave magnetic resonance imaging (MRI) of the human brain. To improve efficiency, we matched a travelling waveguide mode to an electrically large, lossy dielectric load using a thin cylindrical metasurface, which occurs to be a task closely related to impedance matching in waveguide circuits in the microwave. This metasurface was designed as a compact and lightweight replacement for a high-permittivity dielectric waveguide previously proposed for the same purpose. The dispersion analysis showed that both structures (waveguide and metasurface) support a similar type of slow-wave propagation, characterized by a uniform magnetic field profile close to the cylinder axis. At the Larmor frequency, the longitudinal wavenumbers showed close agreement. Based on the optimized unit cell geometry of the periodic copper strip grid loaded with PCB capacitors, full numerical model of the cylindrical metasurface in the presence of a voxel human body model was constructed. We also compared the proposed metasurface with the dielectric waveguide in the traveling-wave setup experimentally, including in vivo measurements performed on a healthy volunteer. The proposed metasurface showed improved B1 + homogeneity (by 17.3%), transmit efficiency (by 27.4%), and SAR-efficiency (by 23%) compared to the dielectric waveguide. The proposed cylindrical metasurface, optimized for field enhancement in the human brain at 7 T in the traveling-wave excitation regime, can further improve the transmit efficiency and homogeneity in the region of interest compared to state-of-the-art structures for traveling-wave MRI, at the same time, granting the advantages of light weight and compactness.

physics.app-ph↗

arXiv2026

Instinct vs. Reflection: Unifying Token and Verbalized Confidence in Multimodal Large Models

Multimodal Large Language Models (MLLMs) have demonstrated exceptional capabilities in various perception and reasoning tasks. Despite this success, ensuring their reliability in practical deployment necessitates robust confidence estimation. Prior works have predominantly focused on text-only LLMs, often relying on computationally expensive self-consistency sampling. In this paper, we extend this to multimodal settings and conduct a comprehensive evaluation of MLLMs' response confidence estimation. Our analysis reveals a significant instinct-reflection misalignment: the model's implicit token-level support frequently diverges from its verbal self-assessment confidence. To address this misalignment, we propose a monotone confidence fusion framework to merge dual-channel signals and cross-channel consistency to estimate correctness. Subsequently, an order-preserving mean alignment step is applied to correct global bias, which improves calibration while preserving the risk-coverage trade-off for selective prediction. Experiments on diverse open-source and closed-source MLLMs show that our method consistently yields more reliable confidence estimates and improves both calibration and failure prediction. Code will be available at https://github.com/Yunkaidang/Instinct-vs.-Reflection.

cs.CV↗

arXiv2026

Persona-E$^2$: A Human-Grounded Dataset for Personality-Shaped Emotional Responses to Textual Events

Most affective computing research treats emotion as a static property of text, focusing on the writer's sentiment while overlooking the reader's perspective. This approach ignores how individual personalities lead to diverse emotional appraisals of the same event. Although role-playing Large Language Models (LLMs) attempt to simulate such nuanced reactions, they often suffer from "personality illusion'' -- relying on surface-level stereotypes rather than authentic cognitive logic. A critical bottleneck is the absence of ground-truth human data to link personality traits to emotional shifts. To bridge the gap, we introduce Persona-E$^2$ (Persona-Event2Emotion), a large-scale dataset grounded in annotated MBTI and Big Five traits to capture reader-based emotional variations across news, social media, and life narratives. Extensive experiments reveal that state-of-the-art LLMs struggle to capture precise appraisal shifts, particularly in social media domains. Crucially, we find that personality information significantly improves comprehension, with the Big Five traits alleviating "personality illusion.'

cs.CL↗

arXiv2026

UHR-BAT: Budget-Aware Token Compression Vision-Language model for Ultra-High-Resolution Remote Sensing

Ultra-high-resolution (UHR) remote sensing imagery couples kilometer-scale context with query-critical evidence that may occupy only a few pixels. Such vast spatial scale leads to a quadratic explosion of visual tokens and hinders the extraction of information from small objects. Previous works utilize direct downsampling, dense tiling, or global top-k pruning, which either compromise query-critical image details or incur unpredictable compute. In this paper, we propose UHR-BAT, a query-guided and region-faithful token compression framework to efficiently select visual tokens under a strict context budget. Specifically, we leverage text-guided, multi-scale importance estimation for visual tokens, effectively tackling the challenge of achieving precise yet low-cost feature extraction. Furthermore, by introducing region-wise preserve and merge strategies, we mitigate visual token redundancy, further driving down the computational budget. Experimental results show that UHR-BAT achieves state-of-the-art performance across various benchmarks. Code will be available at https://github.com/Yunkaidang/UHR.

cs.CV↗

arXiv2026

CLASP: Class-Adaptive Layer Fusion and Dual-Stage Pruning for Multimodal Large Language Models

Multimodal Large Language Models (MLLMs) suffer from substantial computational overhead due to the high redundancy in visual token sequences. Existing approaches typically address this issue using single-layer Vision Transformer (ViT) features and static pruning strategies. However, such fixed configurations are often brittle under diverse instructions. To overcome these limitations, we propose CLASP, a plug-and-play token reduction framework based on class-adaptive layer fusion and dual-stage pruning. Specifically, CLASP first constructs category-specific visual representations through multi-layer vision feature fusion. It then performs dual-stage pruning, allocating the token budget between attention-salient pivot tokens for relevance and redundancy-aware completion tokens for coverage. Through class-adaptive pruning, CLASP enables prompt-conditioned feature fusion and budget allocation, allowing aggressive yet robust visual token reduction. Extensive experiments demonstrate that CLASP consistently outperforms existing methods across a wide range of benchmarks, pruning ratios, and MLLM architectures. Code will be available at https://github.com/Yunkaidang/CLASP.

cs.CV↗

arXiv2026

Improved measurement of the branching fraction of $D_s^+\toμ^+ν_μ$

Using $e^+e^-$ collision data with an integrated luminosity of $7.33~\mathrm{fb}^{-1}$ collected at center-of-mass energies between 4.128 and 4.226 GeV with the BESIII detector operating at the BEPCII collider, the branching fraction of the leptonic decay $D_s^+\toμ^+ν_μ$ is measured to be $(0.5294\pm0.0108_{\rm stat}\pm0.0085_{\rm syst})$\%. Based on this, the product of the $D_s^+$ decay constant $f_{D_s^+}$ and the magnitude of the $c\to s$ quark mixing matrix element $|V_{cs}|$ is determined to be $f_{D_s^+}|V_{cs}|=241.8\pm2.5_{\rm stat}\pm2.2_{\rm syst}~\mathrm{MeV}$. Using the value of $|V_{cs}|$ given by the global standard model fit, $f_{D_s^+}$ is found to be $248.4\pm2.5_{\rm stat}\pm2.2_{\rm syst}$\,MeV. Alternatively, using the value of $f_{D_s^+}$ from a recent lattice quantum chromodynamics calculation, $|V_{cs}|$ is determined to be $0.968\pm0.010_{\rm stat}\pm0.009_{\rm syst}$.

hep-ex↗

arXiv2026

Accelerating Quantum Tensor Network Simulations with Unified Path Variations and Non-Degenerate Batched Sampling

Quantum trajectory methods reduce the computational overhead of simulating noisy quantum systems, approximating them with $m$ stochastically sampled $2^n$-entry quantum statevectors rather than exact $2^{2n}$-entry density matrices. Recently, Pre-Trajectory Sampling with Batched Execution (PTSBE) has dramatically increased the data collection rate of these methods. While statevector PTSBE has demonstrated data collection speedups of over $10^6 \times$, tensor network implementations only achieved $\sim 15 \times$ speedup. This comparatively modest tensor network advantage stemmed from 1) contraction path recalculations, 2) sequential tensor network sampling, and 3) inflexible/unoptimized contraction hyperparameters. In this manuscript, we increase PTSBE's tensor network data collection rate to more than $10^8\times$ that of traditional trajectories methods by developing 1) error-independent unified path variation, 2) non-degenerate tensor network sampling, and 3) a flexible/optimized contraction framework. While our methods are particularly powerful for accelerating non-proportional sampling, we also demonstrate a more than $1000\times$ speedup for more general quantum simulations.

quant-ph↗

arXiv2026

The Python Simulations of Chemistry Framework: 10 years of an open-source quantum chemistry project

Over the past decade, the Python-based Simulations of Chemistry Framework (PySCF) has developed into a widely used open-source platform for electronic structure theory and quantum chemical method development. This article reviews the major advances since the previous overview in 2020, covering new modules and methodology, infrastructure changes, and performance benchmarks.

physics.chem-ph↗

arXiv2026

AICA-Bench: Holistically Examining the Capabilities of VLMs in Affective Image Content Analysis

Vision-Language Models (VLMs) have demonstrated strong capabilities in perception, yet holistic Affective Image Content Analysis (AICA), which integrates perception, reasoning, and generation into a unified framework, remains underexplored. To address this gap, we introduce AICA-Bench, a comprehensive benchmark with three core tasks: Emotion Understanding (EU), Emotion Reasoning (ER), and Emotion-Guided Content Generation (EGCG). We evaluate 23 VLMs and identify two major limitations: weak intensity calibration and shallow open-ended descriptions. To address these issues, we propose Grounded Affective Tree (GAT) Prompting, a training-free framework that combines visual scaffolding with hierarchical reasoning. Experiments show that GAT reduces intensity errors and improves descriptive depth, providing a strong baseline for future research on affective multimodal understanding and generation.

cs.CV↗

arXiv2026

How Far Are We? Systematic Evaluation of LLMs vs. Human Experts in Mathematical Contest in Modeling

Large language models (LLMs) have achieved strong performance on reasoning benchmarks, yet their ability to solve real-world problems requiring end-to-end workflows remains unclear. Mathematical modeling competitions provide a stringent testbed for evaluating such end-to-end problem-solving capability. We propose a problem-oriented, stage-wise evaluation framework that assesses LLM performance across modeling stages using expert-verified criteria. We validate the framework's reliability by comparing automatic scores with independent human expert judgments on problems from the China Postgraduate Mathematical Contest in Modeling, demonstrating substantially stronger alignment than existing evaluation schemes. Using this framework, we reveal a comprehension-execution gap in state-of-the-art LLMs: while they perform well in early stages such as problem identification and formulation, they exhibit persistent deficiencies in execution-oriented stages including model solving, code implementation, and result analysis. These gaps persist even with increased model scale. We further trace these failures to insufficient specification, missing verification, and lack of validation, with errors propagating across stages without correction. Our findings suggest that bridging this gap requires approaches beyond model scaling, offering insights for applying LLMs to complex real-world problem solving.

cs.CL↗

arXiv2026

Improving MPI Error Detection and Repair with Large Language Models and Bug References

Message Passing Interface (MPI) is a foundational technology in high-performance computing (HPC), widely used for large-scale simulations and distributed training (e.g., in machine learning frameworks such as PyTorch and TensorFlow). However, maintaining MPI programs remains challenging due to their complex interplay among processes and the intricacies of message passing and synchronization. With the advancement of large language models like ChatGPT, it is tempting to adopt such technology for automated error detection and repair. Yet, our studies reveal that directly applying large language models (LLMs) yields suboptimal results, largely because these models lack essential knowledge about correct and incorrect usage, particularly the bugs found in MPI programs. In this paper, we design a bug detection and repair technique alongside Few-Shot Learning (FSL), Chain-of-Thought (CoT) reasoning, and Retrieval Augmented Generation (RAG) techniques in LLMs to enhance the large language model's ability to detect and repair errors. Surprisingly, such enhancements lead to a significant improvement, from 44% to 77%, in error detection accuracy compared to baseline methods that use ChatGPT directly. Additionally, our experiments demonstrate our bug referencing technique generalizes well to other large language models.

cs.SE↗

arXiv2026

Constraining the presence of exotrojans in hot Jupiter systems using TTV observations from TESS

Co-orbital bodies (Trojans) share a 1:1 mean-motion resonance with a planet. Although Trojans are common in the Solar System, none has yet been confirmed in an exoplanetary system. Hot Jupiters are not expected to retain primordial co-orbitals efficiently, but their deep and frequent transits make them favorable targets for observational constraints using transit timing variations (TTVs). As part of the ExoEcho project, we analyze TESS photometry for 260 confirmed hot Jupiters with published RV-based masses to search for TTV signals compatible with Trojan companions. We derive transit times and compare the observed residuals with co-orbital models computed with REBOUND N-body simulations. Accounting for the degeneracy between Trojan mass and libration amplitude, we place upper mass limits on possible companions over a range of typical libration amplitudes. For a representative libration amplitude of 15 deg, we rule out exotrojans more massive than 1 Earth mass in 130 systems, corresponding to about 50% of the sample. A more conservative chi-square analysis that incorporates observational uncertainties raises this threshold to 3 Earth masses. We further combine these limits with dynamical-stability constraints for the 1:1 resonance to exclude unstable configurations. Our results provide population-level constraints on massive exotrojans in short-period systems and establish a framework for future high-precision searches with missions such as PLATO and ET (Earth 2.0).

astro-ph.EP↗

arXiv2026

Point What You Mean: Visually Grounded Instruction Policy

Vision-Language-Action (VLA) models align vision and language with embodied control, but their object referring ability remains limited when relying solely on text prompt, especially in cluttered or out-of-distribution (OOD) scenes. In this study, we introduce the Point-VLA, a plug-and-play policy that augments language instructions with explicit visual cues (e.g., bounding boxes) to resolve referential ambiguity and enable precise object-level grounding. To efficiently scale visually grounded datasets, we further develop an automatic data annotation pipeline requiring minimal human effort. We evaluate Point-VLA on diverse real-world referring tasks and observe consistently stronger performance than text-only instruction VLAs, particularly in cluttered or unseen-object scenarios, with robust generalization. These results demonstrate that Point-VLA effectively resolves object referring ambiguity through pixel-level visual grounding, achieving more generalizable embodied control.

cs.CV↗

arXiv2026

UniDex: A Robot Foundation Suite for Universal Dexterous Hand Control from Egocentric Human Videos

Dexterous manipulation remains challenging due to the cost of collecting real-robot teleoperation data, the heterogeneity of hand embodiments, and the high dimensionality of control. We present UniDex, a robot foundation suite that couples a large-scale robot-centric dataset with a unified vision-language-action (VLA) policy and a practical human-data capture setup for universal dexterous hand control. First, we construct UniDex-Dataset, a robot-centric dataset over 50K trajectories across eight dexterous hands (6--24 DoFs), derived from egocentric human video datasets. To transform human data into robot-executable trajectories, we employ a human-in-the-loop retargeting procedure to align fingertip trajectories while preserving plausible hand-object contacts, and we operate on explicit 3D pointclouds with human hands masked to narrow kinematic and visual gaps. Second, we introduce the Function-Actuator-Aligned Space (FAAS), a unified action space that maps functionally similar actuators to shared coordinates, enabling cross-hand transfer. Leveraging FAAS as the action parameterization, we train UniDex-VLA, a 3D VLA policy pretrained on UniDex-Dataset and finetuned with task demonstrations. In addition, we build UniDex-Cap, a simple portable capture setup that records synchronized RGB-D streams and human hand poses and converts them into robot-executable trajectories to enable human-robot data co-training that reduces reliance on costly robot demonstrations. On challenging tool-use tasks across two different hands, UniDex-VLA achieves 81% average task progress and outperforms prior VLA baselines by a large margin, while exhibiting strong spatial, object, and zero-shot cross-hand generalization. Together, UniDex-Dataset, UniDex-VLA, and UniDex-Cap provide a scalable foundation suite for universal dexterous manipulation.

cs.RO↗

arXiv2026

MOSS-TTS Technical Report

This technical report presents MOSS-TTS, a speech generation foundation model built on a scalable recipe: discrete audio tokens, autoregressive modeling, and large-scale pretraining. Built on MOSS-Audio-Tokenizer, a causal Transformer tokenizer that compresses 24 kHz audio to 12.5 fps with variable-bitrate RVQ and unified semantic-acoustic representations, we release two complementary generators: MOSS-TTS, which emphasizes structural simplicity, scalability, and long-context/control-oriented deployment, and MOSS-TTS-Local-Transformer, which introduces a frame-local autoregressive module for higher modeling efficiency, stronger speaker preservation, and a shorter time to first audio. Across multilingual and open-domain settings, MOSS-TTS supports zero-shot voice cloning, token-level duration control, phoneme-/pinyin-level pronunciation control, smooth code-switching, and stable long-form generation. This report summarizes the design, training recipe, and empirical characteristics of the released models.

cs.SD↗