arXiv subjects

Ziyu Xiong

Publications and source records attributed to Ziyu Xiong.

11 recordsLinked to original sources

arXiv2026

SRPO: Setwise Relative Policy Optimization for Multi-Agent LLMs

Multi-agent large language models solve complex tasks by coordinating several policies in a shared environment. However, existing reinforcement learning methods usually optimize each response or trajectory separately, even when several outputs jointly cause one state transition. Consequently, the update unit differs from the action executed by the system. To address this problem, we propose SRPO (Setwise Relative Policy Optimization), which treats the active set the minimal set of outputs consumed by one transition, as one multi-agent action. Specifically, SRPO combines member log-ratios into one cardinality-normalized set ratio, assigns one relative advantage, and clips the set once. This formulation unifies division of labor and joint co-evolution as actions with different set sizes. Experiments on mathematical reasoning and multi-turn search demonstrate one training interface for fixed, mixed, and dynamically routed workflows across four model scales, with the strongest macro-average results among the reported comparisons. Optimization diagnostics further characterize its stability under different event reductions and set sizes.

cs.AI↗

arXiv2026

GE-Act 2.0: Pretraining and Scaling a World-Action Model for Robotic Manipulation

World-action models (WAM) predict future states to guide robot actions, enabling learning from both action-free video and action-labeled interaction. Most inherit pretrained video generators, leaving WAM pretraining and scaling underexplored. We introduce Genie Envisioner Act 2.0 (GE-Act 2.0), a world-action model whose trainable generative and action components are all initialized from scratch on manipulation data. It combines a control-oriented autoencoder (CoAE), a single-step visual planner (SVP), and an inverse dynamics model (IDM). CoAE retains action- and instruction-relevant information under aggressive compression, while SVP produces a complete future state in one differentiable pass, so visual planning and inverse dynamics can be pretrained separately on complementary data. The components are then jointly trained with knowledge-aligned selective optimization (KASO), which reduces mismatched supervision by selecting only predicted futures judged behaviorally compatible with the recorded action. We evaluate pretrained checkpoints directly, without per-task fine-tuning, on 100 tasks across 20 manipulation skill groups with held-out scenes, backgrounds, lighting, and object instances. Scaling co-training data from 300 to 30,000 hours raises success from 17.1% to 44.1% on G1-OP and from 13.4% to 31.1% on G2-90D; despite comprising less than 2% of the co-training data, G2-90D improves by 17.7 points, suggesting cross-embodiment transfer. Gains span 19/20 and 18/20 skill groups, and skill-specific coverage strongly correlates with zero-shot out-of-distribution (OOD) success (Pearson r=0.80; Spearman rho=0.85). Under the same protocol, the model grounds object, color, shape, and position references in at least 90% of trials and follows explicit instructions even when they conflict with an already-committed behavior or a conventional scene association.

cs.RO↗

arXiv2026

OmniReasoner: Thinking with Long Audio-Video via Native Tool Use

Long audio-video reasoning is difficult for omnimodal LLMs because the decisive evidence is often sparse, cross-modal, and too expensive to preserve with uniformly high-fidelity inputs. We introduce OmniReasoner, a tool-use post-training framework for Thinking with Long Audio-Video: omni-modal LLMs learn, via supervised fine-tuning and reinforcement learning, to decide whether and where to call a zoom-in tool before answering. OmniReasoner first builds a low-cost global preview of the full stream and then, when needed, calls the zoom-in tool with a requested temporal interval for higher-fidelity visual and audio inspection before answering. Because the model observes different sampling granularities before and after this call -- a sparse global preview and a denser local clip -- we introduce TimeAnchor, which keeps the tool's temporal argument valid and round-trip-consistent across these granularities, rather than tied to frame indices from a particular sampling rate. To make this tool-use behavior trainable without expensive manual interval annotation, we build a Temporal Augmented Data Engine that synthesizes tool-use post-training trajectories by video editing and composition. Experiments across omnimodal and video benchmarks show that OmniReasoner improves both answer accuracy and temporal grounding while concentrating high-fidelity computation on informative regions. Code is available at https://github.com/RockyChen0205/OmniReasoner.

cs.CV↗

arXiv2026

ACoT-VLA: Action Chain-of-Thought for Vision-Language-Action Models

Vision-Language-Action models have emerged as essential generalist robot policies for diverse manipulation tasks, conventionally relying on directly translating multimodal inputs into actions via Vision-Language Model embeddings. Recent advancements have introduced explicit intermediary reasoning-such as sub-task prediction (language) or goal image synthesis (vision)-to guide action generation. However, these intermediate reasoning are often indirect and inherently limited in their capacity to convey the full, granular information required for precise action execution. Instead, we posit that the most effective form of reasoning is one that deliberates directly in the action space. We introduce Action Chain-of-Thought (ACoT), a paradigm where the reasoning process itself is formulated as a structured sequence of coarse action intents that guide the final policy. In this paper, we propose ACoT-VLA, a novel architecture that materializes the ACoT paradigm. Specifically, we introduce two complementary components: an Explicit Action Reasoner (EAR) and Implicit Action Reasoner (IAR). The former proposes coarse reference trajectories as explicit action-level reasoning steps, while the latter extracts latent action priors from internal representations of multimodal input, co-forming an ACoT that conditions the downstream action head to enable grounded policy learning. Extensive experiments in real-world and simulation environments demonstrate the superiority of our proposed method. Code is available at: https://github.com/AgibotTech/ACoT-VLA.

cs.RO↗

arXiv2026

SOP: A Scalable Online Post-Training System for Vision-Language-Action Models

Vision-language-action (VLA) models achieve strong generalization through large-scale pre-training, but real-world deployment requires expert-level task proficiency in addition to broad generality. Existing post-training approaches for VLA models are typically offline, single-robot, or task-specific, limiting effective on-policy adaptation and scalable learning from real-world interaction. We introduce a Scalable Online Post-training (SOP) system that enables online, distributed, multi-task post-training of generalist VLA models directly in the physical world. SOP tightly couples execution and learning through a closed-loop architecture in which a fleet of robots continuously streams on-policy experience and human intervention signals to a centralized cloud learner, and asynchronously receives updated policies. This design supports prompt on-policy correction, scales experience collection through parallel deployment, and preserves generality during adaptation. SOP is agnostic to the choice of post-training algorithm; we instantiate it with both interactive imitation learning (HG-DAgger) and reinforcement learning (RECAP). Across a range of real-world manipulation tasks including cloth folding, box assembly, and grocery restocking, we show that SOP substantially improves the performance of large pretrained VLA models while maintaining a single shared policy across tasks. Effective post-training can be achieved within hours of real-world interaction, and performance scales near-linearly with the number of robots in the fleet. These results suggest that tightly coupling online learning with fleet-scale deployment is instrumental to enabling efficient, reliable, and scalable post-training of generalist robot policies in the physical world.

cs.RO↗

arXiv2025

Atomic Diffusion Models for Small Molecule Structure Elucidation from NMR Spectra

Nuclear Magnetic Resonance (NMR) spectroscopy is a cornerstone technique for determining the structures of small molecules and is especially critical in the discovery of novel natural products and clinical therapeutics. Yet, interpreting NMR spectra remains a time-consuming, manual process requiring extensive domain expertise. We introduce ChefNMR (CHemical Elucidation From NMR), an end-to-end framework that directly predicts an unknown molecule's structure solely from its 1D NMR spectra and chemical formula. We frame structure elucidation as conditional generation from an atomic diffusion model built on a non-equivariant transformer architecture. To model the complex chemical groups found in natural products, we generated a dataset of simulated 1D NMR spectra for over 111,000 natural products. ChefNMR predicts the structures of challenging natural product compounds with an unsurpassed accuracy of over 65%. This work takes a significant step toward solving the grand challenge of automating small-molecule structure elucidation and highlights the potential of deep learning in accelerating molecular discovery. Code is available at https://github.com/ml-struct-bio/chefnmr.

cs.LG↗

arXiv2025

Conceptual Design Report of Super Tau-Charm Facility: The Accelerator

Electron-positron colliders operating in the GeV region of center-of-mass energies or the Tau-Charm energy region, have been proven to enable competitive frontier research, due to its several unique features. With the progress of high energy physics in the last two decades, a new-generation Tau-Charm factory, Super Tau Charm Facility (STCF) has been actively promoting by the particle physics community in China. STCF holds great potential to address fundamental questions such as the essence of color confinement and the matter-antimatter asymmetry in the universe in the next decades. The main design goals of STCF are with a center-of-mass energy ranging from 2 to 7 GeV and a peak luminosity surpassing 5*10^34 cm^-2s^-1 that is optimized at a center-of-mass energy of 4 GeV, which is about 50 times that of the currently operating Tau-Charm factory - BEPCII. The STCF accelerator is composed of two main parts: a double-ring collider with the crab-waist collision scheme and an injector that provides top-up injections for both electron and positron beams. As a typical third-generation electron-positron circular collider, the STCF accelerator faces many challenges in both accelerator physics and technology. In this paper, the conceptual design of the STCF accelerator complex is presented, including the ongoing efforts and plans for technological R&D, as well as the required infrastructure. The STCF project aims to secure support from the Chinese central government for its construction during the 15th Five-Year Plan (2026-2030) in China.

physics.acc-ph↗

arXiv2024

Statistical Inference for Regression with Imputed Binary Covariates with Application to Emotion Recognition

In the flourishing live streaming industry, accurate recognition of streamers' emotions has become a critical research focus, with profound implications for audience engagement and content optimization. However, precise emotion coding typically requires manual annotation by trained experts, making it extremely expensive and time-consuming to obtain complete observational data for large-scale studies. Motivated by this challenge in streamer emotion recognition, we develop here a novel imputation method together with a principled statistical inference procedure for analyzing partially observed binary data. Specifically, we assume for each observation an auxiliary feature vector, which is sufficiently cheap to be fully collected for the whole sample. We next assume a small pilot sample with both the target binary covariates (i.e., the emotion status) and the auxiliary features fully observed, of which the size could be considerably smaller than that of the whole sample. Thereafter, a regression model can be constructed for the target binary covariates and the auxiliary features. This enables us to impute the missing binary features using the fully observed auxiliary features for the entire sample. We establish the associated asymptotic theory for principled statistical inference and present extensive simulation experiments, demonstrating the effectiveness and theoretical soundness of our proposed method. Furthermore, we validate our approach using a comprehensive dataset on emotion recognition in live streaming, demonstrating that our imputation method yields smaller standard errors and is more statistically efficient than using pilot data only. Our findings have significant implications for enhancing user experience and optimizing engagement on streaming platforms.

stat.ME↗

arXiv2021

State-level Racially Motivated Hate Crimes Contrast Public Opinion on the #StopAsianHate and #StopAAPIHate Movement

#StopAsianHate and #StopAAPIHate are two of the most commonly used hashtags that represent the current movement to end hate crimes against the Asian American and Pacific Islander community. We conduct a social media study of public opinion on the #StopAsianHate and #StopAAPIHate movement based on 46,058 Twitter users across 30 states in the United States ranging from March 18 to April 11, 2021. The movement attracts more participation from women, younger adults, Asian and Black communities. 51.56% of the Twitter users show direct support, 18.38% are news about anti-Asian hate crimes, while 5.43% show a negative attitude towards the movement. Public opinion varies across user characteristics. Furthermore, among the states with most racial bias motivated hate crimes, the negative attitude towards the #StopAsianHate and #StopAAPIHate movement is the weakest. To our best knowledge, this is the first large-scale social media-based study to understand public opinion on the #StopAsianHate and #StopAAPIHate movement. We hope our study can provide insights and promote research on anti-Asian hate crimes, and ultimately help address such a serious societal issue for the common benefits of all communities.

cs.SI↗

arXiv2021

Observational Study of Working from Home during the COVID-19 Pandemic Using Social Media Data

Since March 2020, companies nationwide have started work from home (WFH) due to the rapid increase of confirmed COVID-19 cases in an attempt to help prevent the coronavirus from spreading and rescue the economy from the pandemic. Many organizations have conducted surveys to understand people's opinions towards WFH. However, the findings are limited due to small sample size and the dynamic topics over time. This study aims to understand the U.S. public opinions on working from home during the COVID-19 pandemic. We conduct a large-scale social media study using Twitter data to portrait different groups who have positive/negative opinions about WFH. We perform an ordinary least squares regression to investigate the relationship between the sentiment about WFH and user characteristics including gender, age, ethnicity, median household income, and population density. To better understand public opinion, we use latent Dirichlet allocation to extract topics and discover how tweet contents relate to people's attitudes. These findings provide evidence that sentiment about WFH varies across user characteristics. Furthermore, the content analysis sheds light on the nuanced differences in sentiment and reveals disparities relate to WFH.

cs.SI↗

arXiv2018

The Trajectory of Voice Onset Time with Vocal Aging

Vocal aging, a universal process of human aging, can largely affect one's language use, possibly including some subtle acoustic features of one's utterances like Voice Onset Time. To figure out the time effects, Queen Elizabeth's Christmas speeches are documented and analyzed in the long-term trend. We build statistical models of time dependence in Voice Onset Time, controlling a wide range of other fixed factors, to present annual variations and the simulated trajectory. It is revealed that the variation range of Voice Onset Time has been narrowing over fifty years with a slight reduction in the mean value, which, possibly, is an effect of diminishing exertion, resulting from subdued muscle contraction, transcending other non-linguistic factors in forming Voice Onset Time patterns over a long time.

cs.SD↗