arXiv subjects

Peng Xu

Publications and source records attributed to Peng Xu.

At least 91 records · Page 5Linked to original sources

arXiv2026

PegasusFlow: Parallel Rolling-Denoising Score Sampling for Robot Diffusion Planner Flow Matching

Diffusion models offer powerful generative capabilities for robot trajectory planning, yet their practical deployment on robots is hindered by a critical bottleneck: a reliance on imitation learning from expert demonstrations. This paradigm is often impractical for specialized robots where data is scarce and creates an inefficient, theoretically suboptimal training pipeline. To overcome this, we introduce PegasusFlow, a hierarchical rolling-denoising framework that enables direct and parallel sampling of trajectory score gradients from environmental interaction, completely bypassing the need for expert data. Our core innovation is a novel sampling algorithm, Weighted Basis Function Optimization (WBFO), which leverages spline basis representations to achieve superior sample efficiency and faster convergence compared to traditional methods like MPPI. The framework is embedded within a scalable, asynchronous parallel simulation architecture that supports massively parallel rollouts for efficient data collection. Extensive experiments on trajectory optimization and robotic navigation tasks demonstrate that our approach, particularly Action-Value WBFO (AVWBFO) combined with a reinforcement learning warm-start, significantly outperforms baselines. In a challenging barrier-crossing task, our method achieved a 100% success rate and was 18% faster than the next-best method, validating its effectiveness for complex terrain locomotion planning. https://masteryip.github.io/pegasusflow.github.io/

cs.RO↗

arXiv2026

KCFRC: Kinematic Collision-Aware Foothold Reachability Criteria for Legged Locomotion

Legged robots face significant challenges in navigating complex environments, as they require precise real-time decisions for foothold selection and contact planning. While existing research has explored methods to select footholds based on terrain geometry or kinematics, a critical gap remains: few existing methods efficiently validate the existence of a non-collision swing trajectory. This paper addresses this gap by introducing KCFRC, a novel approach for efficient foothold reachability analysis. We first formally define the foothold reachability problem and establish a sufficient condition for foothold reachability. Based on this condition, we develop the KCFRC algorithm, which enables robots to validate foothold reachability in real time. Our experimental results demonstrate that KCFRC achieves remarkable time efficiency, completing foothold reachability checks for a single leg across 900 potential footholds in an average of 2 ms. Furthermore, we show that KCFRC can accelerate trajectory optimization and is particularly beneficial for contact planning in confined spaces, enhancing the adaptability and robustness of legged robots in challenging environments.

cs.RO↗

arXiv2026

TokenMixer-Large: Scaling Up Large Ranking Models in Industrial Recommenders

While scaling laws for recommendation models have gained significant traction, existing architectures such as Wukong, HiFormer and DHEN, often struggle with sub-optimal designs and hardware under-utilization, limiting their practical scalability. Our previous TokenMixer architecture (introduced in RankMixer paper) addressed effectiveness and efficiency by replacing self-attention with a ightweight token-mixing operator; however, it faced critical bottlenecks in deeper configurations, including sub-optimal residual paths, vanishing gradients, incomplete MoE sparsification and constrained scalability. In this paper, we propose TokenMixer-Large, a systematically evolved architecture designed for extreme-scale recommendation. By introducing a mixing-and-reverting operation, inter-layer residuals and the auxiliary loss, we ensure stable gradient propagation even as model depth increases. Furthermore, we incorporate a Sparse Per-token MoE to enable efficient parameter expansion. TokenMixer-Large successfully scales its parameters to 7-billion and 15-billion on online traffic and offline experiments, respectively. Currently deployed in multiple scenarios at ByteDance, TokenMixer-Large has achieved significant offline and online performance gains, delivering an increase of +1.66\% in orders and +2.98\% in per-capita preview payment GMV for e-commerce, improving ADSS by +2.0\% in advertising and achieving a +1.4\% revenue growth for live streaming.

cs.IR↗

arXiv2026

View-Centric Multi-Object Tracking with Homographic Matching in Moving UAV

In this paper, we address the challenge of Multi-Object Tracking (MOT) in moving Unmanned Aerial Vehicle (UAV) scenarios, where irregular flight trajectories, such as hovering, turning left/right, and moving up/down, lead to significantly greater complexity compared to fixed-camera MOT. Specifically, changes in the scene background not only render traditional frame-to-frame object IoU association methods ineffective but also introduce significant view shifts in the objects, which complicates tracking. To overcome these issues, we propose a novel HomView-MOT framework, which for the first time, harnesses the view homography inherent in changing scenes to solve MOT challenges in moving environments, incorporating homographic matching and view-centric concepts. We introduce a Fast Homography Estimation (FHE) algorithm for rapid computation of homography matrices between video frames, enabling object View-Centric ID Learning (VCIL) and leveraging multi-view homography to learn cross-view ID features. Concurrently, our Homographic Matching Filter (HMF) maps object bounding boxes from different frames onto a common view plane for a more realistic physical IoU association. Extensive experiments have proven that these innovations allow HomView-MOT to achieve state-of-the-art performance on prominent UAV MOT datasets VisDrone and UAVDT.

cs.CV↗

arXiv2026

Generative quantum machine learning via denoising diffusion probabilistic models

Deep generative models are key-enabling technology to computer vision, text generation, and large language models. Denoising diffusion probabilistic models (DDPMs) have recently gained much attention due to their ability to generate diverse and high-quality samples in many computer vision tasks, as well as to incorporate flexible model architectures and a relatively simple training scheme. Quantum generative models, empowered by entanglement and superposition, have brought new insight to learning classical and quantum data. Inspired by the classical counterpart, we propose the quantum denoising diffusion probabilistic model (QuDDPM) to enable efficiently trainable generative learning of quantum data. QuDDPM adopts sufficient layers of circuits to guarantee expressivity, while it introduces multiple intermediate training tasks as interpolation between the target distribution and noise to avoid barren plateau and guarantee efficient training. We provide bounds on the learning error and demonstrate QuDDPM's capability in learning correlated quantum noise model, quantum many-body phases, and topological structure of quantum data. The results provide a paradigm for versatile and efficient quantum generative learning.

quant-ph↗

arXiv2026

Scalable native multiqubit gates via engineered noncomputational-state interactions in superconducting fluxonium qubits

Native multiqubit gates could be essential for bridging the gap from current noisy devices to future utility-scale quantum computers, as they can substantially reduce circuit depth for near-term applications on noisy devices and may also lower the physical overhead of fault-tolerant quantum computation. Here we introduce a scalable protocol for implementing native multi-controlled gates on fluxonium qubits, supporting an arbitrary number of control qubits ($N > 1$) while remaining compatible with existing single- and two-qubit gate realizations. Our approach leverages engineered interactions in noncomputational state manifolds to enable qubit-state selective transitions, which is activated for the direct implementation of $(C^{\otimes N})Z$ gates. We show that in square lattices with fluxonium qubits, $CCZ$, $CCCZ$, and $CCCCZ$ gates with errors around 0.01 (0.001) are achievable, with gate lengths of $50\,(100)\,\text{ns}$, $100\,(250)\,\text{ns}$, and $150\,(300)\,\text{ns}$, respectively. Looking forward, integrating these native multi-controlled gates with primitive single- and two-qubit gate sets within a single quantum processor could significantly enhance flexibility in circuit synthesis and offer a promising alternative pathway toward utility-scale quantum computing.

quant-ph↗

arXiv2026

Calibration Method of Spacecraft-Inertial Sensor Center-of-Mass Offset for the Taiji Gravitational Wave Detection Mission under Science Mode

Accurately calibrating the center-of-mass (CoM) offset between the spacecraft (SC) and the inertial sensor test mass (TM) is crucial for space-based gravitational-wave (GW) antennas, such as LISA and Taiji. Current calibration methods require additional spacecraft maneuvers that disrupt science data continuity and inter-satellite links, compromising the coherence of gravitational wave signals. Here, we present a maneuver-free calibration scheme that directly estimates the CoM offset vector using only standard science-mode measurements from inertial sensors, interferometers, and differential wavefront sensors. By embedding the CoM offset induced coupling acceleration as an extended state in a model-based adaptive Kalman filter, we achieve estimation accuracy of 0.01-1.5 mm across all axes with a maximum error below 1%. This approach enables continuous, high-precision calibration during nominal observation runs, ensuring continuous and coherent gravitational wave data collection while maintaining the required precision, and also facilitating advanced DFACS functions such as performance evaluations and fault diagnosis. For LISA-like missions, where data continuity is paramount for detecting faint gravitational wave signals, this method will enhance scientific output and reliability.

gr-qc↗

arXiv2026

Tensor Abelian Geometry of VI-modules

In this short note, we study the spectrum of prime Serre ideals of global representations for noetherian families. In particular, we prove that the spectrum of prime Serre ideals of finitely generated VI-modules is homeomorphic to N^{*}, the one-point compactification of N, which differs from the Balmer spectrum of derived VI-modules. Our method could also be applied to the category of finitely generated FI-modules and the category of global representations for the family of cyclic p-groups.

math.CT↗

arXiv2026

Extending the micro-Hertz detection horizons via orbital resonance effect for geocentric gravitational wave antennas

The $μ$Hz gravitational wave band holds crucial insights into coalescing supermassive black hole binaries and stochastic backgrounds but remains inaccessible due to technical challenges. We demonstrate that geocentric space-based GW detectors (e.g., TianQin, gLISA, GADFLI) can bridge this gap by considering orbital resonance effects, circumventing the need for prohibitively long baselines. When GW frequencies match with integer multiples of a satellite's orbital frequency, sustained tidal forces induce cumulative orbital deviations through resonant effects, which, combined with orbital modulation, improve detector sensitivity by 1-2 orders of magnitude in the $μ$Hz band. Consequently, geocentric missions can detect SMBHBs across significantly expanded mass-redshift parameter space. Crucially, such observations could synergize with pulsar timing array data of the same binaries at earlier inspiral stages, enabling unprecedented joint tests of strong-field gravity and binary evolution. Our findings establish geocentric antennas as a cost-effective, near-term precursor for unlocking the $μ$Hz GW astronomy.

gr-qc↗

arXiv2025

SPECTRE: Spectral Pre-training Embeddings with Cylindrical Temporal Rotary Position Encoding for Fine-Grained sEMG-Based Movement Decoding

Decoding fine-grained movement from non-invasive surface Electromyography (sEMG) is a challenge for prosthetic control due to signal non-stationarity and low signal-to-noise ratios. Generic self-supervised learning (SSL) frameworks often yield suboptimal results on sEMG as they attempt to reconstruct noisy raw signals and lack the inductive bias to model the cylindrical topology of electrode arrays. To overcome these limitations, we introduce SPECTRE, a domain-specific SSL framework. SPECTRE features two primary contributions: a physiologically-grounded pre-training task and a novel positional encoding. The pre-training involves masked prediction of discrete pseudo-labels from clustered Short-Time Fourier Transform (STFT) representations, compelling the model to learn robust, physiologically relevant frequency patterns. Additionally, our Cylindrical Rotary Position Embedding (CyRoPE) factorizes embeddings along linear temporal and annular spatial dimensions, explicitly modeling the forearm sensor topology to capture muscle synergies. Evaluations on multiple datasets, including challenging data from individuals with amputation, demonstrate that SPECTRE establishes a new state-of-the-art for movement decoding, significantly outperforming both supervised baselines and generic SSL approaches. Ablation studies validate the critical roles of both spectral pre-training and CyRoPE. SPECTRE provides a robust foundation for practical myoelectric interfaces capable of handling real-world sEMG complexities.

cs.HC↗

arXiv2025

DeepKnown-Guard: A Proprietary Model-Based Safety Response Framework for AI Agents

With the widespread application of Large Language Models (LLMs), their associated security issues have become increasingly prominent, severely constraining their trustworthy deployment in critical domains. This paper proposes a novel safety response framework designed to systematically safeguard LLMs at both the input and output levels. At the input level, the framework employs a supervised fine-tuning-based safety classification model. Through a fine-grained four-tier taxonomy (Safe, Unsafe, Conditionally Safe, Focused Attention), it performs precise risk identification and differentiated handling of user queries, significantly enhancing risk coverage and business scenario adaptability, and achieving a risk recall rate of 99.3%. At the output level, the framework integrates Retrieval-Augmented Generation (RAG) with a specifically fine-tuned interpretation model, ensuring all responses are grounded in a real-time, trustworthy knowledge base. This approach eliminates information fabrication and enables result traceability. Experimental results demonstrate that our proposed safety control model achieves a significantly higher safety score on public safety evaluation benchmarks compared to the baseline model, TinyR1-Safety-8B. Furthermore, on our proprietary high-risk test set, the framework's components attained a perfect 100% safety score, validating their exceptional protective capabilities in complex risk scenarios. This research provides an effective engineering pathway for building high-security, high-trust LLM applications.

cs.AI↗

arXiv2025

LEMUR: Large scale End-to-end MUltimodal Recommendation

Traditional ID-based recommender systems often struggle with cold-start and generalization challenges. Multimodal recommendation systems, which leverage textual and visual data, offer a promising solution to mitigate these issues. However, existing industrial approaches typically adopt a two-stage training paradigm: first pretraining a multimodal model, then applying its frozen representations to train the recommendation model. This decoupled framework suffers from misalignment between multimodal learning and recommendation objectives, as well as an inability to adapt dynamically to new data. To address these limitations, we propose LEMUR, the first large-scale multimodal recommender system trained end-to-end from raw data. By jointly optimizing both the multimodal and recommendation components, LEMUR ensures tighter alignment with downstream objectives while enabling real-time parameter updates. Constructing multimodal sequential representations from user history often entails prohibitively high computational costs. To alleviate this bottleneck, we propose a novel memory bank mechanism that incrementally accumulates historical multimodal representations throughout the training process. After one month of deployment in Douyin Search, LEMUR has led to a 0.843% reduction in query change rate decay and a 0.81% improvement in QAUC. Additionally, LEMUR has shown significant gains across key offline metrics for Douyin Advertisement. Our results validate the superiority of end-to-end multimodal recommendation in real-world industrial scenarios.

cs.IR↗

arXiv2025

Source Confusion of Massive Black Hole Binaries for the Taiji Mission

We systematically investigate the source confusion of massive black hole binaries (MBHBs) for the Taiji space-based gravitational wave mission. Source confusion, arising from the overlap of signals in both time and frequency domains, can degrade parameter recovery accuracy. To assess this effect, we simulate three representative models MBHB populations to estimate event overlap events. Assuming 100 detections per year, only 0.31-4.2 overlaps are expected annually. Based on Fisher information matrix with the $\texttt{IMRPhenomD}$ and $\texttt{IMRPhenomHM}$ waveform models, we find that overlap significantly enlarges parameter uncertainties, while the inclusion of higher-order modes (HMs) effectively mitigates this effect. Severe confusion ($Δ\mathcal{M}_z / \mathcal{M}_z<$ 0.2%) occurs in fewer than 0.14% across the three population models. The full Bayesian analysis further corroborates the Fisher predictions, and also reveals that HMs help break key parameter degeneracies, with or without signal overlap. These findings underscore the importance of incorporating HMs for accurate inference in future space-based observations.

gr-qc↗

arXiv2025

Gravitomagnetic-Hydrodynamics and Turbulence in Early Universe

The nonlinear coupling between spacetime geometry and matter in the early Universe remains a frontier in theoretical cosmology. By introducing a novel gravitomagnetic-hydrodynamic framework, we reveal a fundamental analogy between magnetohydrodynamics and the co-evolution of spacetime geometry and relativistic plasma. We demonstrate that, in high-energy environments such as the electroweak phase transition, the (newly defined) gravitomagnetic Reynolds number becomes large, signifying a strongly coupled system where the gravitomagnetic field could be frozen into the fluid. This coupling inevitably leads to the emergence of gravitational Alfvén waves and could drive a transition to turbulence involving the dynamics of spacetime itself. Our findings suggest that gravitomagnetic-hydrodynamic turbulence may leave imprints on the stochastic gravitational wave background, offering a new window into the nonlinear dynamics of the primordial Universe.

gr-qc↗

arXiv2025

From Feedback to Checklists: Grounded Evaluation of AI-Generated Clinical Notes

AI-generated clinical notes are increasingly used in healthcare, but evaluating their quality remains a challenge due to high subjectivity and limited scalability of expert review. Existing automated metrics often fail to align with real-world physician preferences. To address this, we propose a pipeline that systematically distills real user feedback into structured checklists for note evaluation. These checklists are designed to be interpretable, grounded in human feedback, and enforceable by LLM-based evaluators. Using deidentified data from over 21,000 clinical encounters (prepared in accordance with the HIPAA safe harbor standard) from a deployed AI medical scribe system, we show that our feedback-derived checklist outperforms a baseline approach in our offline evaluations in coverage, diversity, and predictive power for human ratings. Extensive experiments confirm the checklist's robustness to quality-degrading perturbations, significant alignment with clinician preferences, and practical value as an evaluation methodology. In offline research settings, our checklist offers a practical tool for flagging notes that may fall short of our defined quality standards.

cs.CL↗

arXiv2025

ColLab: A Collaborative Spatial Progressive Data Engine for Referring Expression Comprehension and Generation

Referring Expression Comprehension (REC) and Referring Expression Generation (REG) are fundamental tasks in multimodal understanding, supporting precise object localization through natural language. However, existing REC and REG datasets rely heavily on manual annotation, which is labor-intensive and difficult to scale. In this paper, we propose ColLab, a collaborative spatial progressive data engine that enables fully automated REC and REG data generation without human supervision. Specifically, our method introduces a Collaborative Multimodal Model Interaction (CMMI) strategy, which leverages the semantic understanding of multimodal large language models (MLLMs) and large language models (LLMs) to generate descriptions. Furthermore, we design a module termed Spatial Progressive Augmentation (SPA) to enhance spatial expressiveness among duplicate instances. Experiments demonstrate that ColLab significantly accelerates the annotation process of REC and REG while improving the quality and discriminability of the generated expressions. In addition to the core methodological contribution, our framework was partially adopted in the data generation pipeline of the ICCV 2025 MARS2 Challenge on Multimodal Reasoning, enriching the dataset with diverse and challenging samples that better reflect real-world reasoning demands.

cs.CV↗

arXiv2025

Inter-Spacecraft Tilt-to-Length Noise Reduction Algorithm for Taiji Mission

The Taiji mission for space-based gravitational wave (GW) detection employs laser interferometry to measure picometer-scale distance variations induced by GWs. The tilt-to-length (TTL) coupling noise in the inter-spacecraft interferometers, which originates from the angular jitters of the spacecrafts and the movable optical subassemblies, is predicted to be one of the main noise sources that might reduce Taiji's sensitivity to GWs. Since these angular jitters can be read out through the differential wavefront sensors, it is possible to suppress TTL noise during the data processing stage by fitting and subtracting it. This paper proposes an improved algorithm for TTL noise suppression, which addresses the issue of unknown noise floor required for optimal estimation in the practical detection scenario, and the design of this algorithm takes into account the presence of GW signals. The algorithm is validated via numerical simulation, which is built on a spacecraft dynamics simulation incorporating Taiji's drag-free and attitude control system. We also demonstrate the robustness of this algorithm by varying TTL coefficients at different levels, indicating that our algorithm is applicable to a range of payload statuses, and ultimately providing a critical advancement toward realizing Taiji's full sensitivity.

gr-qc↗

arXiv2025

MARS2 2025 Challenge on Multimodal Reasoning: Datasets, Methods, Results, Discussion, and Outlook

This paper reviews the MARS2 2025 Challenge on Multimodal Reasoning. We aim to bring together different approaches in multimodal machine learning and LLMs via a large benchmark. We hope it better allows researchers to follow the state-of-the-art in this very dynamic area. Meanwhile, a growing number of testbeds have boosted the evolution of general-purpose large language models. Thus, this year's MARS2 focuses on real-world and specialized scenarios to broaden the multimodal reasoning applications of MLLMs. Our organizing team released two tailored datasets Lens and AdsQA as test sets, which support general reasoning in 12 daily scenarios and domain-specific reasoning in advertisement videos, respectively. We evaluated 40+ baselines that include both generalist MLLMs and task-specific models, and opened up three competition tracks, i.e., Visual Grounding in Real-world Scenarios (VG-RS), Visual Question Answering with Spatial Awareness (VQA-SA), and Visual Reasoning in Creative Advertisement Videos (VR-Ads). Finally, 76 teams from the renowned academic and industrial institutions have registered and 40+ valid submissions (out of 1200+) have been included in our ranking lists. Our datasets, code sets (40+ baselines and 15+ participants' methods), and rankings are publicly available on the MARS2 workshop website and our GitHub organization page https://github.com/mars2workshop/, where our updates and announcements of upcoming events will be continuously provided.

cs.CV↗