Searcharxiv⌕ Search

arXiv subjects

Yang Bai

Publications and source records attributed to Yang Bai.

At least 73 records · Page 4Linked to original sources

RynnBrain: Open Embodied Foundation Models

Despite rapid progress in multimodal foundation models, embodied intelligence community still lacks a unified, physically grounded foundation model that integrates perception, reasoning, and planning within real-world spatial-temporal dynamics. We introduce RynnBrain, an open-source spatiotemporal foundation model for embodied intelligence. RynnBrain strengthens four core capabilities in a unified framework: comprehensive egocentric understanding, diverse spatiotemporal localization, physically grounded reasoning, and physics-aware planning. The RynnBrain family comprises three foundation model scales (2B, 8B, and 30B-A3B MoE) and four post-trained variants tailored for downstream embodied tasks (i.e., RynnBrain-Nav, RynnBrain-Plan, and RynnBrain-VLA) or complex spatial reasoning tasks (i.e., RynnBrain-CoP). In terms of extensive evaluations on 20 embodied benchmarks and 8 general vision understanding benchmarks, our RynnBrain foundation models largely outperform existing embodied foundation models by a significant margin. The post-trained model suite further substantiates two key potentials of the RynnBrain foundation model: (i) enabling physically grounded reasoning and planning, and (ii) serving as a strong pretrained backbone that can be efficiently adapted to diverse embodied tasks.

cs.RO↗

A Machine Learning accelerated geophysical fluid solver

Machine learning methods have been successful in many areas, like image classification and natural language processing. However, it still needs to be determined how to apply ML to areas with mathematical constraints, like solving PDEs. Among various approaches to applying ML techniques to solving PDEs, the data-driven discretization method presents a promising way of accelerating and improving existing PDE solver on structured grids where it predicts the coefficients of quasi-linear stencils for computing values or derivatives of a function at given positions. It can improve the accuracy and stability of low-resolution simulation compared with using traditional finite difference or finite volume schemes. Meanwhile, it can also benefit from traditional numerical schemes like achieving conservation law by adapting finite volume type formulations. In this thesis, we have implemented the shallow water equation and Euler equation classic solver under a different framework. Experiments show that our classic solver performs much better than the Pyclaw solver. Then we propose four different deep neural networks for the ML-based solver. The results indicate that two of these approaches could output satisfactory solutions.

cs.CV↗

Classification Under Local Differential Privacy with Model Reversal and Model Averaging

Local differential privacy (LDP) has become a central topic in data privacy research, offering strong privacy guarantees by perturbing user data at the source and removing the need for a trusted curator. However, the noise introduced by LDP often significantly reduces data utility. To address this issue, we reinterpret private learning under LDP as a transfer learning problem, where the noisy data serve as the source domain and the unobserved clean data as the target. We propose novel techniques specifically designed for LDP to improve classification performance without compromising privacy: (1) a noised binary feedback-based evaluation mechanism for estimating dataset utility; (2) model reversal, which salvages underperforming classifiers by inverting their decision boundaries; and (3) model averaging, which assigns weights to multiple reversed classifiers based on their estimated utility. We provide theoretical excess risk bounds under LDP and demonstrate how our methods reduce this risk. Empirical results on both simulated and real-world datasets show substantial improvements in classification accuracy.

cs.LG↗

SRBench: A Comprehensive Benchmark for Sequential Recommendation with Large Language Models

LLM development has aroused great interest in Sequential Recommendation (SR) applications. However, comprehensive evaluation of SR models remains lacking due to the limitations of the existing benchmarks: 1) an overemphasis on accuracy, ignoring other real-world demands (e.g., fairness); 2) existing datasets fail to unleash LLMs' potential, leading to unfair comparison between Neural-Network-based SR (NN-SR) models and LLM-based SR (LLM-SR) models; and 3) no reliable mechanism for extracting task-specific answers from unstructured LLM outputs. To address these limitations, we propose SRBench, a comprehensive SR benchmark with three core designs: 1) a multi-dimensional framework covering accuracy, fairness, stability and efficiency, aligned with practical demands; 2) a unified input paradigm via prompt engineering to boost LLM-SR performance and enable fair comparisons between models; 3) a novel prompt-extractor-coupled extraction mechanism, which captures answers from LLM outputs through prompt-enforced output formatting and a numeric-oriented extractor. We have used SRBench to evaluate 13 mainstream models and discovered some meaningful insights (e.g., LLM-SR models overfocus on item popularity but lack deep understanding of item quality). Concisely, SRBench enables fair and comprehensive assessments for SR models, underpinning future research and practical application.

cs.IR↗

Note2Chat: Improving LLMs for Multi-Turn Clinical History Taking Using Medical Notes

Effective clinical history taking is a foundational yet underexplored component of clinical reasoning. While large language models (LLMs) have shown promise on static benchmarks, they often fall short in dynamic, multi-turn diagnostic settings that require iterative questioning and hypothesis refinement. To address this gap, we propose \method{}, a note-driven framework that trains LLMs to conduct structured history taking and diagnosis by learning from widely available medical notes. Instead of relying on scarce and sensitive dialogue data, we convert real-world medical notes into high-quality doctor-patient dialogues using a decision tree-guided generation and refinement pipeline. We then propose a three-stage fine-tuning strategy combining supervised learning, simulated data augmentation, and preference learning. Furthermore, we propose a novel single-turn reasoning paradigm that reframes history taking as a sequence of single-turn reasoning problems. This design enhances interpretability and enables local supervision, dynamic adaptation, and greater sample efficiency. Experimental results show that our method substantially improves clinical reasoning, achieving gains of +16.9 F1 and +21.0 Top-1 diagnostic accuracy over GPT-4o. Our code and dataset can be found at https://github.com/zhentingsheng/Note2Chat.

cs.CL↗

Approaching Stable Quark Matter

The determination of whether the ground state of baryon matter in Quantum Chromodynamics (QCD) is the ordinary nucleus or a quark matter state remains a long-standing question in physics. A critical parameter in this investigation is the bag parameter $B$, which quantifies the QCD vacuum energy and can be computed using nonperturbative methods such as Lattice QCD (LQCD). By combining the equation of state derived from perturbative QCD (pQCD) with the bag parameter to fit the LQCD-simulated data for isospin-dense matter, we address the stability of quark matter within the LQCD+pQCD framework. Our findings suggest that the current data imposes an upper bound on $B^{1/4} \lesssim 160$ MeV, approaching a conclusive statement on quark matter stability. Given the lower bound on $B$ from the quark condensate contribution to the vacuum energy, the stable 2-flavor quark matter remains possible, whereas the stable 2+1-flavor quark matter is excluded, assuming complete deconfinement and chiral-symmetry restoration and the reliability of pQCD at baryon chemical potentials around the proton mass. Additionally, we derive more general thermodynamic bounds on the quark matter energy-per-baryon and $B$, which, while weaker, provide complementary insights.

hep-ph↗

Multi-transport Distributional Regression

We study distribution-on-distribution regression problems in which a response distribution depends on multiple distributional predictors. Such settings arise naturally in applications where the outcome distribution is driven by several heterogeneous distributional sources, yet remain challenging due to the nonlinear geometry of the Wasserstein space. We propose an intrinsic regression framework that aggregates predictor-specific transported distributions through a weighted Fréchet mean in the Wasserstein space. The resulting model admits multiple distributional predictors, assigns interpretable weights quantifying their relative contributions, and defines a flexible regression operator that is invariant to auxiliary construction choices, such as the selection of a reference distribution. From a theoretical perspective, we establish identifiability of the induced regression operator and derive asymptotic guarantees for its estimation under a predictive Wasserstein semi-norm, which directly characterizes convergence of the composite prediction map. Extensive simulation studies and a real data application demonstrate the improved predictive performance and interpretability of the proposed approach compared with existing Wasserstein regression methods.

stat.ME↗

Efficient Context Scaling with LongCat ZigZag Attention

We introduce LongCat ZigZag Attention (LoZA), which is a sparse attention scheme designed to transform any existing full-attention models into sparse versions with rather limited compute budget. In long-context scenarios, LoZA can achieve significant speed-ups both for prefill-intensive (e.g., retrieval-augmented generation) and decode-intensive (e.g., tool-integrated reasoning) cases. Specifically, by applying LoZA to LongCat-Flash during mid-training, we serve LongCat-Flash-Exp as a long-context foundation model that can swiftly process up to 1 million tokens, enabling efficient long-term reasoning and long-horizon agentic capabilities.

cs.CL↗

Using LLMs for Late Multimodal Sensor Fusion for Activity Recognition

Sensor data streams provide valuable information around activities and context for downstream applications, though integrating complementary information can be challenging. We show that large language models (LLMs) can be used for late fusion for activity classification from audio and motion time series data. We curated a subset of data for diverse activity recognition across contexts (e.g., household activities, sports) from the Ego4D dataset. Evaluated LLMs achieved 12-class zero- and one-shot classification F1-scores significantly above chance, with no task-specific training. Zero-shot classification via LLM-based fusion from modality-specific models can enable multimodal temporal applications where there is limited aligned training data for learning a shared embedding space. Additionally, LLM-based fusion can enable model deploying without requiring additional memory and computation for targeted application-specific multimodal models.

cs.LG↗

CoVAR: Co-generation of Video and Action for Robotic Manipulation via Multi-Modal Diffusion

We present a method to generate video-action pairs that follow text instructions, starting from an initial image observation and the robot's joint states. Our approach automatically provides action labels for video diffusion models, overcoming the common lack of action annotations and enabling their full use for robotic policy learning. Existing methods either adopt two-stage pipelines, which limit tightly coupled cross-modal information sharing, or rely on adapting a single-modal diffusion model for a joint distribution that cannot fully leverage pretrained video knowledge. To overcome these limitations, we (1) extend a pretrained video diffusion model with a parallel, dedicated action diffusion model that preserves pretrained knowledge, (2) introduce a Bridge Attention mechanism to enable effective cross-modal interaction, and (3) design an action refinement module to convert coarse actions into precise controls for low-resolution datasets. Extensive evaluations on multiple public benchmarks and real-world datasets demonstrate that our method generates higher-quality videos, more accurate actions, and significantly outperforms existing baselines, offering a scalable framework for leveraging large-scale video data for robotic learning.

cs.CV↗

DRAW2ACT: Turning Depth-Encoded Trajectories into Robotic Demonstration Videos

Video diffusion models provide powerful real-world simulators for embodied AI but remain limited in controllability for robotic manipulation. Recent works on trajectory-conditioned video generation address this gap but often rely on 2D trajectories or single modality conditioning, which restricts their ability to produce controllable and consistent robotic demonstrations. We present DRAW2ACT, a depth-aware trajectory-conditioned video generation framework that extracts multiple orthogonal representations from the input trajectory, capturing depth, semantics, shape and motion, and injects them into the diffusion model. Moreover, we propose to jointly generate spatially aligned RGB and depth videos, leveraging cross-modality attention mechanisms and depth supervision to enhance the spatio-temporal consistency. Finally, we introduce a multimodal policy model conditioned on the generated RGB and depth sequences to regress the robot's joint angles. Experiments on Bridge V2, Berkeley Autolab, and simulation benchmarks show that DRAW2ACT achieves superior visual fidelity and consistency while yielding higher manipulation success rates compared to existing baselines.

cs.CV↗

Improving Generative Ad Text on Facebook using Reinforcement Learning

Generative artificial intelligence (AI), in particular large language models (LLMs), is poised to drive transformative economic change. LLMs are pre-trained on vast text data to learn general language patterns, but a subsequent post-training phase is critical to align them for specific real-world tasks. Reinforcement learning (RL) is the leading post-training technique, yet its economic impact remains largely underexplored and unquantified. We examine this question through the lens of the first deployment of an RL-trained LLM for generative advertising on Facebook. Integrated into Meta's Text Generation feature, our model, "AdLlama," powers an AI tool that helps advertisers create new variations of human-written ad text. To train this model, we introduce reinforcement learning with performance feedback (RLPF), a post-training method that uses historical ad performance data as a reward signal. In a large-scale 10-week A/B test on Facebook spanning nearly 35,000 advertisers and 640,000 ad variations, we find that AdLlama improves click-through rates by 6.7% (p=0.0296) compared to a supervised imitation model trained on curated ads. This represents a substantial improvement in advertiser return on investment on Facebook. We also find that advertisers who used AdLlama generated more ad variations, indicating higher satisfaction with the model's outputs. To our knowledge, this is the largest study to date on the use of generative AI in an ecologically valid setting, offering an important data point quantifying the tangible impact of RL post-training. Furthermore, the results show that RLPF is a promising and generalizable approach for metric-driven post-training that bridges the gap between highly capable language models and tangible outcomes.

cs.LG↗

Exploring Community-Powered Conversational Agent for Health Knowledge Acquisition: A Case Study in Colorectal Cancer

Online communities have become key platforms where young adults, actively seek and share information, including health knowledge. However, these users often face challenges when browsing these communities, such as fragmented content, varying information quality and unfamiliar terminology. Based on a survey with 56 participants and follow-up interviews, we identify common challenges and expected features for learning health knowledge. In this paper, we develop a computational workflow that integrates community content into a conversational agent named CanAnswer to facilitate health knowledge acquisition. Using colorectal cancer as a case study, we evaluate CanAnswer through a lab study with 24 participants and interviews with six medical experts. Results show that CanAnswer improves the recalled gained knowledge and reduces the task workload of the learning session. Our expert interviews (N=6) further confirm the reliability and usefulness of CanAnswer. We discuss the generality of CanAnswer and provide design considerations for enhancing the usefulness and credibility of community-powered learning tools.

cs.HC↗

Heterogeneous Cosmological Phase Transitions: Seeded by Domain Walls and Junctions

Heterogeneous nucleation is central to many familiar first-order phase transitions such as the freezing of water and the solidification of metals, and it can also play a crucial role in cosmology. We examine nucleation seeded by preexisting domain walls and demonstrate its strong impact on the dynamics of cosmological phase transitions. The bubble solutions take the form of spherical caps, and the contact angle is fixed by the ratio of the domain-wall tension to the bubble-wall tension. A larger domain-wall tension, or equivalently a smaller contact angle, reduces the wall-seeded bubble volume and lowers the critical nucleation action. For theories with $\mathbb{Z}_{n\geq 3}$ symmetry, domain-wall junctions naturally appear and we find that they seed nucleation even more efficiently than the walls themselves. Using a two-scalar-field model as an illustration, we compute nucleation temperatures for both homogeneous and heterogeneous channels and show that junction-seeded nucleation occurs at a higher temperature and is the dominant mechanism that completes the first-order cosmological phase transition.

hep-ph↗

Toward generic control for soft robotic systems

Soft robotics has advanced rapidly, yet its control methods remain fragmented: different morphologies and actuation schemes still require task-specific controllers, hindering theoretical integration and large-scale deployment. A generic control framework is therefore essential, and a key obstacle lies in the persistent use of rigid-body control logic, which relies on precise models and strict low-level execution. Such a paradigm is effective for rigid robots but fails for soft robots, where the ability to tolerate and exploit approximate action representations, i.e., control compliance, is the basis of robustness and adaptability rather than a disturbance to be eliminated. Control should thus shift from suppressing compliance to explicitly exploiting it. Human motor control exemplifies this principle: instead of computing exact dynamics or issuing detailed muscle-level commands, it expresses intention through high-level movement tendencies, while reflexes and biomechanical mechanisms autonomously resolve local details. This architecture enables robustness, flexibility, and cross-task generalization. Motivated by this insight, we propose a generic soft-robot control framework grounded in control compliance and validate it across robots with diverse morphologies and actuation mechanisms. The results demonstrate stable, safe, and cross-platform transferable behavior, indicating that embracing control compliance, rather than resisting it, may provide a widely applicable foundation for unified soft-robot control.

cs.RO↗

JW-Flare: Accurate Solar Flare Forecasting Method Based on Multimodal Large Language Models

Solar flares, the most powerful explosive phenomena in the solar system, may pose significant hazards to spaceborne satellites and ground-based infrastructure. Despite decades of intensive research, reliable flare prediction remains a challenging task. Large Language Models, as a milestone in artificial intelligence, exhibit exceptional general knowledge and next-token prediction capabilities. Here we introduce JW-Flare, the first Multimodal Large Language Models (MLLMs) explicitly trained for solar flare forecasting through fine-tuning on textual physic parameters of solar active regions and magnetic field images. This method demonstrates state-of-the-art (SOTA) performance for large flares prediction on the test dataset. It effectively identifies all 79 X-class flares from 18,949 test samples, yielding a True Skill Statistic (TSS) of 0.95 and a True Positive Rate (TPR) of 1.00, outperforming traditional predictive models. We further investigate the capability origins of JW-Flare through explainability experiments, revealing that solar physics knowledge acquired during pre-training contributes to flare forecasting performance. Additionally, we evaluate models of different parameter scales, confirming the Scaling_Law of Large Language Models in domain-specific applications, such as solar physics. This study marks a substantial advance in both the scale and accuracy of solar flare forecasting and opens a promising avenue for AI-driven methodologies in broader scientific domains.

astro-ph.SR↗

JPRO: Automated Multimodal Jailbreaking via Multi-Agent Collaboration Framework

The widespread application of large VLMs makes ensuring their secure deployment critical. While recent studies have demonstrated jailbreak attacks on VLMs, existing approaches are limited: they require either white-box access, restricting practicality, or rely on manually crafted patterns, leading to poor sample diversity and scalability. To address these gaps, we propose JPRO, a novel multi-agent collaborative framework designed for automated VLM jailbreaking. It effectively overcomes the shortcomings of prior methods in attack diversity and scalability. Through the coordinated action of four specialized agents and its two core modules: Tactic-Driven Seed Generation and Adaptive Optimization Loop, JPRO generates effective and diverse attack samples. Experimental results show that JPRO achieves over a 60\% attack success rate on multiple advanced VLMs, including GPT-4o, significantly outperforming existing methods. As a black-box attack approach, JPRO not only uncovers critical security vulnerabilities in multimodal models but also offers valuable insights for evaluating and enhancing VLM robustness.

cs.CR↗

Explore Data Left Behind in Reinforcement Learning for Reasoning Language Models

Reinforcement Learning with Verifiable Rewards (RLVR) has emerged as an effective approach for improving the reasoning abilities of large language models (LLMs). The Group Relative Policy Optimization (GRPO) family has demonstrated strong performance in training LLMs with RLVR. However, as models train longer and scale larger, more training prompts become residual prompts, those with zero variance rewards that provide no training signal. Consequently, fewer prompts contribute to training, reducing diversity and hindering effectiveness. To fully exploit these residual prompts, we propose the Explore Residual Prompts in Policy Optimization (ERPO) framework, which encourages exploration on residual prompts and reactivates their training signals. ERPO maintains a history tracker for each prompt and adaptively increases the sampling temperature for residual prompts that previously produced all correct responses. This encourages the model to generate more diverse reasoning traces, introducing incorrect responses that revive training signals. Empirical results on the Qwen2.5 series demonstrate that ERPO consistently surpasses strong baselines across multiple mathematical reasoning benchmarks.

cs.CL↗