arXiv subjects

Wei He

Publications and source records attributed to Wei He.

At least 109 records · Page 6Linked to original sources

arXiv2025

A New Partial State-Feedback IDA-PBC for Two-Dimensional Nonlinear Systems: Application to Power Converters with Experimental Results

In this paper we propose a variation of the widely popular Interconnection-and-Damping-Assigment Passivity-Based Control (IDA-PBC) based on Poincare's Lemma to design output feedback globally stabilizing controllers for two dimensional systems. The procedure is constructive and, in comparison with the classical IDA-PBC, whose application is often stymied by the need to solve the (infamous) matching partial differential equation (PDE), in this new method the PDE is replaced by an ordinary differential equation, whose solution is far simpler. The procedure is then applied for the design of voltage-feedback controllers for the three most typical DC-to-DC power converter topologies: the Buck, Boost and Buck-Boost. It is assumed that these converters feed an uncertain load, which is characterized by a static relation between its voltage and current. In the case when the load consists of the parallel connection of a resistive term and a constant power load we propose an adaptive version of the design, adding an identification scheme for the load parameters. This allows the controller to regulate the converter output when the load varies-that is a typical scenario in these applications. Extensive numerical simulations and experimental results validate the approach.

eess.SY↗

arXiv2025

AgentGym-RL: Training LLM Agents for Long-Horizon Decision Making through Multi-Turn Reinforcement Learning

Developing autonomous LLM agents capable of making a series of intelligent decisions to solve complex, real-world tasks is a fast-evolving frontier. Like human cognitive development, agents are expected to acquire knowledge and skills through exploration and interaction with the environment. Despite advances, the community still lacks a unified, interactive reinforcement learning (RL) framework that can effectively train such agents from scratch -- without relying on supervised fine-tuning (SFT) -- across diverse and realistic environments. To bridge this gap, we introduce AgentGym-RL, a new framework to train LLM agents for multi-turn interactive decision-making through RL. The framework features a modular and decoupled architecture, ensuring high flexibility and extensibility. It encompasses a wide variety of real-world scenarios, and supports mainstream RL algorithms. Furthermore, we propose ScalingInter-RL, a training approach designed for exploration-exploitation balance and stable RL optimization. In early stages, it emphasizes exploitation by restricting the number of interactions, and gradually shifts towards exploration with larger horizons to encourage diverse problem-solving strategies. In this way, the agent develops more diverse behaviors and is less prone to collapse under long horizons. We perform extensive experiments to validate the stability and effectiveness of both the AgentGym-RL framework and the ScalingInter-RL approach. Our agents match or surpass commercial models on 27 tasks across diverse environments. We offer key insights and will open-source the complete AgentGym-RL framework -- including code and datasets -- to empower the research community in developing the next generation of intelligent agents.

cs.LG↗

arXiv2025

Distill Visual Chart Reasoning Ability from LLMs to MLLMs

Solving complex chart Q&A tasks requires advanced visual reasoning abilities in multimodal large language models (MLLMs), including recognizing key information from visual inputs and conducting reasoning over it. While fine-tuning MLLMs for reasoning is critical, collecting and annotating charts and questions is expensive, hard to scale, and often results in low-quality annotations. To address this, we propose Code-as-Intermediary Translation (CIT), a cost-effective, efficient and scalable data synthesis method for distilling visual reasoning abilities from LLMs to MLLMs. The code serves as an intermediary that translates visual chart representations into textual representations, enabling language models to understand cross-modal information and generate reasoning chains accordingly. In this way, we can employ text-based synthesizing techniques to expand chart-plotting code and generate high-quality Q&A pairs for training models. This produces ReachQA, a dataset containing 3k reasoning-intensive charts and 20k Q&A pairs to enhance both recognition and reasoning abilities of MLLMs. Experiments show that models fine-tuned with ReachQA not only perform well on chart-related tasks but also show performance gains on general reasoning benchmarks. The code and dataset are publicly available at https://github.com/hewei2001/ReachQA.

cs.CL↗

arXiv2025

Quadratic BSDEs with double constraints driven by G-Brownian motion

In this paper, we investigate the well-posedness of quadratic backward stochastic differential equations driven by G-Brownian motion (referred to as G-BSDEs) with double mean reflections. By employing a representation of the solution via G-BMO martingale techniques, along with fixed point arguments, the Skorokhod problem, the backward Skorokhod problem, and the θ-method, we establish existence and uniqueness results for such G-BSDEs under both bounded and unbounded terminal conditions.

math.PR↗

arXiv2025

Maximum principle for discrete-time robust stochastic optimal control problem

This paper firstly presents the necessary and sufficient conditions for a kind of discrete-time robust stochastic optimal control problem with convex control domains. As it is an "inf sup problem", the classical variational method is invalid. We obtain the variational inequality with a common reference probability by systematically using weak convergence approach and the minimax theorem. Moreover, a discrete-time robust investment problem is also studied where the explicit optimal control is given.

math.OC↗

arXiv2025

Enhancing LLM Reasoning with Iterative DPO: A Comprehensive Empirical Investigation

Recent advancements in post-training methodologies for large language models (LLMs) have highlighted reinforcement learning (RL) as a critical component for enhancing reasoning. However, the substantial computational costs associated with RL-based approaches have led to growing interest in alternative paradigms, such as Direct Preference Optimization (DPO). In this study, we investigate the effectiveness of DPO in facilitating self-improvement for LLMs through iterative preference-based learning. We demonstrate that a single round of DPO with coarse filtering significantly enhances mathematical reasoning performance, particularly for strong base model. Furthermore, we design an iterative enhancement framework for both the generator and the reward model (RM), enabling their mutual improvement through online interaction across multiple rounds of DPO. Finally, with simple verifiable rewards, our model DPO-VP achieves RL-level performance with significantly lower computational overhead. These findings highlight DPO as a scalable and cost-effective alternative to RL, offering a practical solution for enhancing LLM reasoning in resource-constrained situations.

cs.CL↗

arXiv2025

DAG-AFL:Directed Acyclic Graph-based Asynchronous Federated Learning

Due to the distributed nature of federated learning (FL), the vulnerability of the global model and the need for coordination among many client devices pose significant challenges. As a promising decentralized, scalable and secure solution, blockchain-based FL methods have attracted widespread attention in recent years. However, traditional consensus mechanisms designed for Proof of Work (PoW) similar to blockchain incur substantial resource consumption and compromise the efficiency of FL, particularly when participating devices are wireless and resource-limited. To address asynchronous client participation and data heterogeneity in FL, while limiting the additional resource overhead introduced by blockchain, we propose the Directed Acyclic Graph-based Asynchronous Federated Learning (DAG-AFL) framework. We develop a tip selection algorithm that considers temporal freshness, node reachability and model accuracy, with a DAG-based trusted verification strategy. Extensive experiments on 3 benchmarking datasets against eight state-of-the-art approaches demonstrate that DAG-AFL significantly improves training efficiency and model accuracy by 22.7% and 6.5% on average, respectively.

cs.LG↗

arXiv2025

Assess Space-Based Solar Power in European-Scale Power System Decarbonization

Meeting net-zero targets remains formidable as terrestrial renewables grapple with intermittency and regional variability. Here, we integrate space-based solar power (SBSP) -- a potential near-constant, orbital solar technology -- into a high-resolution, Europe-wide capacity-expansion and dispatch model to quantify its contribution under net-zero constraints. We examine two advanced SBSP designs: (1) a near-baseload, low Technology Readiness Level (TRL) concept (heliostat-based Representative Design RD1) and (2) a partially intermittent, higher-TRL concept (planar-based RD2), both drawing on NASA's 2050 cost and performance projections. Our results show that RD1 can reduce total system costs by 7--15%, displace up to 80% of intermittent wind and solar, and cut battery usage by over 70%, if it meets its forecast cost reductions -- though long-duration storage (e.g., hydrogen) remains essential for seasonal balancing. By contrast, RD2 is economically unattractive at its projected 2050 costs. Through extensive sensitivity analyses, we identify cost thresholds at which SBSP shifts from cost-prohibitive to complementary and ultimately to a dominant baseload technology. Specifically, RD1 becomes complementary at roughly 14x and dominant at 9x the 2050 solar PV capital cost, benefiting from its continuous power generation. Meanwhile, RD2 must achieve even lower cost levels (9x to be complementary and 6x to dominate) and would rely on short-duration storage to mitigate its partial intermittency. These findings provide quantified techno-economic benchmarks and reveal alternative net-zero pathways, offering critical guidance for policymakers and industry stakeholders seeking large-scale, centrally coordinated renewable solutions with non- or low-intermittency.

physics.soc-ph↗

arXiv2025

LangTime: A Language-Guided Unified Model for Time Series Forecasting with Proximal Policy Optimization

Recent research has shown an increasing interest in utilizing pre-trained large language models (LLMs) for a variety of time series applications. However, there are three main challenges when using LLMs as foundational models for time series forecasting: (1) Cross-domain generalization. (2) Cross-modality alignment. (3) Error accumulation in autoregressive frameworks. To address these challenges, we proposed LangTime, a language-guided unified model for time series forecasting that incorporates cross-domain pre-training with reinforcement learning-based fine-tuning. Specifically, LangTime constructs Temporal Comprehension Prompts (TCPs), which include dataset-wise and channel-wise instructions, to facilitate domain adaptation and condense time series into a single token, enabling LLMs to understand better and align temporal data. To improve autoregressive forecasting, we introduce TimePPO, a reinforcement learning-based fine-tuning algorithm. TimePPO mitigates error accumulation by leveraging a multidimensional rewards function tailored for time series and a repeat-based value estimation strategy. Extensive experiments demonstrate that LangTime achieves state-of-the-art cross-domain forecasting performance, while TimePPO fine-tuning effectively enhances the stability and accuracy of autoregressive forecasting.

cs.LG↗

arXiv2025

OmniEval: A Benchmark for Evaluating Omni-modal Models with Visual, Auditory, and Textual Inputs

In this paper, we introduce OmniEval, a benchmark for evaluating omni-modality models like MiniCPM-O 2.6, which encompasses visual, auditory, and textual inputs. Compared with existing benchmarks, our OmniEval has several distinctive features: (i) Full-modal collaboration: We design evaluation tasks that highlight the strong coupling between audio and video, requiring models to effectively leverage the collaborative perception of all modalities; (ii) Diversity of videos: OmniEval includes 810 audio-visual synchronized videos, 285 Chinese videos and 525 English videos; (iii) Diversity and granularity of tasks: OmniEval contains 2617 question-answer pairs, comprising 1412 open-ended questions and 1205 multiple-choice questions. These questions are divided into 3 major task types and 12 sub-task types to achieve comprehensive evaluation. Among them, we introduce a more granular video localization task named Grounding. Then we conduct experiments on OmniEval with several omni-modality models. We hope that our OmniEval can provide a platform for evaluating the ability to construct and understand coherence from the context of all modalities. Codes and data could be found at https://omnieval-benchmark.github.io/.

cs.CV↗

arXiv2025

iDiT-HOI: Inpainting-based Hand Object Interaction Reenactment via Video Diffusion Transformer

Digital human video generation is gaining traction in fields like education and e-commerce, driven by advancements in head-body animation and lip-syncing technologies. However, realistic Hand-Object Interaction (HOI) - the complex dynamics between human hands and objects - continues to pose challenges. Generating natural and believable HOI reenactments is difficult due to issues such as occlusion between hands and objects, variations in object shapes and orientations, and the necessity for precise physical interactions, and importantly, the ability to generalize to unseen humans and objects. This paper presents a novel framework iDiT-HOI that enables in-the-wild HOI reenactment generation. Specifically, we propose a unified inpainting-based token process method, called Inp-TPU, with a two-stage video diffusion transformer (DiT) model. The first stage generates a key frame by inserting the designated object into the hand region, providing a reference for subsequent frames. The second stage ensures temporal coherence and fluidity in hand-object interactions. The key contribution of our method is to reuse the pretrained model's context perception capabilities without introducing additional parameters, enabling strong generalization to unseen objects and scenarios, and our proposed paradigm naturally supports long video generation. Comprehensive evaluations demonstrate that our approach outperforms existing methods, particularly in challenging real-world scenes, offering enhanced realism and more seamless hand-object interactions.

cs.GR↗

arXiv2025

SemEval-2025 Task 1: AdMIRe -- Advancing Multimodal Idiomaticity Representation

Idiomatic expressions present a unique challenge in NLP, as their meanings are often not directly inferable from their constituent words. Despite recent advancements in Large Language Models (LLMs), idiomaticity remains a significant obstacle to robust semantic representation. We present datasets and tasks for SemEval-2025 Task 1: AdMiRe (Advancing Multimodal Idiomaticity Representation), which challenges the community to assess and improve models' ability to interpret idiomatic expressions in multimodal contexts and in multiple languages. Participants competed in two subtasks: ranking images based on their alignment with idiomatic or literal meanings, and predicting the next image in a sequence. The most effective methods achieved human-level performance by leveraging pretrained LLMs and vision-language models in mixture-of-experts settings, with multiple queries used to smooth over the weaknesses in these models' representations of idiomaticity.

cs.CL↗

arXiv2025

Symmetry-protected electronic metastability in an optically driven cuprate ladder

Optically excited quantum materials exhibit nonequilibrium states with remarkable emergent properties, but these phenomena are usually transient, decaying on picosecond timescales and limiting practical applications. Advancing the design and control of nonequilibrium phases requires the development of targeted strategies to achieve long-lived, metastable phases. Here, we report the discovery of symmetry-protected electronic metastability in the model cuprate ladder Sr$_{14}$Cu$_{24}$O$_{41}$. Using femtosecond resonant x-ray scattering and spectroscopy, we show that this metastability is driven by a transfer of holes from chain-like charge reservoirs into the ladders. This ultrafast charge redistribution arises from the optical dressing and activation of a hopping pathway that is forbidden by symmetry at equilibrium. Relaxation back to the ground state is hence suppressed after the pump coherence dissipates. Our findings highlight how dressing materials with electromagnetic fields can dynamically activate terms in the electronic Hamiltonian, and provide a rational design strategy for nonequilibrium phases of matter.

cond-mat.str-el↗

arXiv2025

Pangu Embedded: An Efficient Dual-system LLM Reasoner with Metacognition

This work presents Pangu Embedded, an efficient Large Language Model (LLM) reasoner developed on Ascend Neural Processing Units (NPUs), featuring flexible fast and slow thinking capabilities. Pangu Embedded addresses the significant computational costs and inference latency challenges prevalent in existing reasoning-optimized LLMs. We propose a two-stage training framework for its construction. In Stage 1, the model is finetuned via an iterative distillation process, incorporating inter-iteration model merging to effectively aggregate complementary knowledge. This is followed by reinforcement learning on Ascend clusters, optimized by a latency-tolerant scheduler that combines stale synchronous parallelism with prioritized data queues. The RL process is guided by a Multi-source Adaptive Reward System (MARS), which generates dynamic, task-specific reward signals using deterministic metrics and lightweight LLM evaluators for mathematics, coding, and general problem-solving tasks. Stage 2 introduces a dual-system framework, endowing Pangu Embedded with a "fast" mode for routine queries and a deeper "slow" mode for complex inference. This framework offers both manual mode switching for user control and an automatic, complexity-aware mode selection mechanism that dynamically allocates computational resources to balance latency and reasoning depth. Experimental results on benchmarks including AIME 2024, GPQA, and LiveCodeBench demonstrate that Pangu Embedded with 7B parameters, outperforms similar-size models like Qwen3-8B and GLM4-9B. It delivers rapid responses and state-of-the-art reasoning quality within a single, unified model architecture, highlighting a promising direction for developing powerful yet practically deployable LLM reasoners.

cs.CL↗

arXiv2025

Building-Guided Pseudo-Label Learning for Cross-Modal Building Damage Mapping

Accurate building damage assessment using bi-temporal multi-modal remote sensing images is essential for effective disaster response and recovery planning. This study proposes a novel Building-Guided Pseudo-Label Learning Framework to address the challenges of mapping building damage from pre-disaster optical and post-disaster SAR images. First, we train a series of building extraction models using pre-disaster optical images and building labels. To enhance building segmentation, we employ multi-model fusion and test-time augmentation strategies to generate pseudo-probabilities, followed by a low-uncertainty pseudo-label training method for further refinement. Next, a change detection model is trained on bi-temporal cross-modal images and damaged building labels. To improve damage classification accuracy, we introduce a building-guided low-uncertainty pseudo-label refinement strategy, which leverages building priors from the previous step to guide pseudo-label generation for damaged buildings, reducing uncertainty and enhancing reliability. Experimental results on the 2025 IEEE GRSS Data Fusion Contest dataset demonstrate the effectiveness of our approach, which achieved the highest mIoU score (54.28%) and secured first place in the competition.

cs.CV↗

arXiv2025

Pangu Ultra MoE: How to Train Your Big MoE on Ascend NPUs

Sparse large language models (LLMs) with Mixture of Experts (MoE) and close to a trillion parameters are dominating the realm of most capable language models. However, the massive model scale poses significant challenges for the underlying software and hardware systems. In this paper, we aim to uncover a recipe to harness such scale on Ascend NPUs. The key goals are better usage of the computing resources under the dynamic sparse model structures and materializing the expected performance gain on the actual hardware. To select model configurations suitable for Ascend NPUs without repeatedly running the expensive experiments, we leverage simulation to compare the trade-off of various model hyperparameters. This study led to Pangu Ultra MoE, a sparse LLM with 718 billion parameters, and we conducted experiments on the model to verify the simulation results. On the system side, we dig into Expert Parallelism to optimize the communication between NPU devices to reduce the synchronization overhead. We also optimize the memory efficiency within the devices to further reduce the parameter and activation management overhead. In the end, we achieve an MFU of 30.0% when training Pangu Ultra MoE, with performance comparable to that of DeepSeek R1, on 6K Ascend NPUs, and demonstrate that the Ascend system is capable of harnessing all the training stages of the state-of-the-art language models. Extensive experiments indicate that our recipe can lead to efficient training of large-scale sparse language models with MoE. We also study the behaviors of such models for future reference.

cs.CL↗

arXiv2025

Pangu Ultra: Pushing the Limits of Dense Large Language Models on Ascend NPUs

We present Pangu Ultra, a Large Language Model (LLM) with 135 billion parameters and dense Transformer modules trained on Ascend Neural Processing Units (NPUs). Although the field of LLM has been witnessing unprecedented advances in pushing the scale and capability of LLM in recent years, training such a large-scale model still involves significant optimization and system challenges. To stabilize the training process, we propose depth-scaled sandwich normalization, which effectively eliminates loss spikes during the training process of deep models. We pre-train our model on 13.2 trillion diverse and high-quality tokens and further enhance its reasoning capabilities during post-training. To perform such large-scale training efficiently, we utilize 8,192 Ascend NPUs with a series of system optimizations. Evaluations on multiple diverse benchmarks indicate that Pangu Ultra significantly advances the state-of-the-art capabilities of dense LLMs such as Llama 405B and Mistral Large 2, and even achieves competitive results with DeepSeek-R1, whose sparse model structure contains much more parameters. Our exploration demonstrates that Ascend NPUs are capable of efficiently and effectively training dense models with more than 100 billion parameters. Our model and system will be available for our commercial customers.

cs.CL↗

arXiv2025

Performance of plastic scintillator modules for top veto tracker at Taishan Antineutrino Observatory

The Taishan Antineutrino Observatory (TAO) experiment incorporates a top veto tracker (TVT) system comprising 160 modules, each composed of plastic scintillator (PS) strips, embedded wavelength shifting fibers (WLS-fibers), and silicon photomultipliers (SiPMs). This article highlights the performance of all produced modules following the production and readout/trigger design, providing insights for scintillation detectors with WLS-fibers. Three kinds of trigger modes and its efficiency have been defined to comprehensively evaluate the performance of this unique design, which has been verified for the batch production, along with comprehensive measurement strategies and quality inspection methods. In "module" mode, the detection(tagging) efficiency of the PS exceeds 99.67\% at a 30 photoelectron threshold, and even in "AND" mode, it surpasses 99.60\% at a 15 photoelectron threshold. The muon tagging efficiency meets TAO's requirements. The production and performance of the PS module set a benchmark for other experiments, with optimized optical fiber arrangements that enhance light yield and muon detection efficiency.

physics.ins-det↗