Pengfei Zhou

Pengfei Zhou, Weiqing Min, Yang Zhang et al.

Food detection is becoming a fundamental task in food computing that supports various multimedia applications, including food recommendation and dietary monitoring. To deal with real-world scenarios, food detection needs to localize and recognize novel food objects that are not seen during training, demanding Zero-Shot Detection (ZSD). However, the complexity of semantic attributes and intra-class feature diversity poses challenges for ZSD methods in distinguishing fine-grained food classes. To tackle this, we propose the Semantic Separable Diffusion Synthesizer (SeeDS) framework for Zero-Shot Food Detection (ZSFD). SeeDS consists of two modules: a Semantic Separable Synthesizing Module (S$^3$M) and a Region Feature Denoising Diffusion Model (RFDDM). The S$^3$M learns the disentangled semantic representation for complex food attributes from ingredients and cuisines, and synthesizes discriminative food features via enhanced semantic information. The RFDDM utilizes a novel diffusion model to generate diversified region features and enhances ZSFD via fine-grained synthesized features. Extensive experiments show the state-of-the-art ZSFD performance of our proposed method on two food datasets, ZSFooD and UECFOOD-256. Moreover, SeeDS also maintains effectiveness on general ZSD datasets, PASCAL VOC and MS COCO. The code and dataset can be found at https://github.com/LanceZPF/SeeDS.

Weidong Tang, Jierui Li, Yueling Hou et al.

True general intelligence requires not only a model of the physical world but also a social world model: the capacity to infer how individual mental states interact and crystallize into group-level outcomes. Despite notable progress in individual-level Theory of Mind (ToM) reasoning, existing multimodal large language models fail at this broader task. Collective behavior emerges non-linearly from social tensions, conformity dynamics, and structural constraints, meaning it cannot be recovered by merely summing individual intentions. We present GroupToM-Bench, the first multimodal benchmark for group-level ToM, built around a causal chain spanning micro-level BDI states (belief, desire, intention), meso-level group tension and structural constraints, and macro-level outcome prediction and mechanistic attribution. To probe this full arc, we develop a seven-level cognitive audit framework. Experiments reveal a gap between current models and human baselines, highlighting a failure to process social structures and non-linear collective dynamics.

Tong Bai, Zhenglin Wan, Pengfei Zhou et al.

As LLM agents adopt large skill libraries, selecting the right subset becomes a structural problem rather than a similarity-matching one: skills depend on, conflict with, specialize, or duplicate one another, a structure invisible to both full enumeration and embedding similarity. We present SkillDAG, which models inter-skill relationships as a typed directed graph and exposes it to an LLM agent as an inference-time, agent-callable structural retrieval interface, queried and evolved during execution rather than baked into a fixed retrieval pipeline: each search returns vector matches, typed-edge neighbors, and conflict signals, and a propose-then-commit protocol lets the agent register execution-backed edges so the graph accumulates structure across episodes. On ALFWorld and SkillsBench with MiniMax-M2.7, SkillDAG reaches 67.1% success and 27.3% reward, exceeding the strongest reported Graph-of-Skills baseline by +12.8 and +8.6 points; the advantage ports to gpt-5.2-codex, and intrinsic SkillsBench Ret@K rises from 65.5 to 78.2 under matched queries. These gains trace to isolable mechanisms: candidate ranking that stays robust as the pool grows 10x where a fixed seeding-diffusion pipeline degrades, and set-monotone online edits that enlarge ground-truth recall without evicting prior hits.

Pengfei Zhou, Shengcong Chen, Di Chen et al.

Robotic manipulation requires models that generate executable actions while anticipating and evaluating their future consequences before physical execution. We present $τ_0$-World Model ($τ_0$-WM), a unified video-action world model that integrates policy learning, video prediction, and action evaluation within a single future-predictive framework. Built on a shared video diffusion backbone, $τ_0$-WM provides two complementary interfaces. First, a video action model jointly predicts future visual latents and continuous action chunks from multi-view observations, language instructions, and robot state. Second, an action-conditioned video simulator rolls out candidate action chunks into multi-view futures and predicts dense task-progress scores. The model is trained on approximately $27{,}300$ hours of real-robot teleoperation, UMI-style interaction, egocentric human videos, and rollout or failure trajectories using modality-specific supervision masks. At inference time, $τ_0$-WM uses test-time computation to sample action candidates, rank them with re-denoising consistency, and invoke simulator-based rectification for low-quality candidates. On challenging long-horizon and fine-grained robotic manipulation tasks, $τ_0$-WM shows superior performance over other relevant baselines.

Hua Yan, Heng Tan, Yi Ding et al.

Human Activity Recognition (HAR) using Inertial Measurement Unit (IMU) sensors is critical for applications in healthcare, safety, and industrial production. However, variations in activity patterns, device types, and sensor placements create distribution gaps across datasets, reducing the performance of HAR models. To address this, we propose LanHAR, a novel system that leverages Large Language Models (LLMs) to generate semantic interpretations of sensor readings and activity labels for cross-dataset HAR. This approach not only mitigates cross-dataset heterogeneity but also enhances the recognition of new activities. LanHAR employs an iterative re-generation method to produce high-quality semantic interpretations with LLMs and a two-stage training framework that bridges the semantic interpretations of sensor readings and activity labels. This ultimately leads to a lightweight sensor encoder suitable for mobile deployment, enabling any sensor reading to be mapped into the semantic interpretation space. Experiments on five public datasets demonstrate that our approach significantly outperforms state-of-the-art methods in both cross-dataset HAR and new activity recognition. The source code is publicly available at https://github.com/DASHLab/LanHAR.

Dongfang Guo, Yuting Wu, Yimin Dai et al.

Camera-based computer vision is essential to autonomous vehicle's perception. This paper presents an attack that uses light-emitting diodes and exploits the camera's rolling shutter effect to create adversarial stripes in the captured images to mislead traffic sign recognition. The attack is stealthy because the stripes on the traffic sign are invisible to human. For the attack to be threatening, the recognition results need to be stable over consecutive image frames. To achieve this, we design and implement GhostStripe, an attack system that controls the timing of the modulated light emission to adapt to camera operations and victim vehicle movements. Evaluated on real testbeds, GhostStripe can stably spoof the traffic sign recognition results for up to 94\% of frames to a wrong class when the victim vehicle passes the road section. In reality, such attack effect may fool victim vehicles into life-threatening incidents. We discuss the countermeasures at the levels of camera sensor, perception model, and autonomous driving system.

Pengfei Zhou, Weiqing Min, Jiajun Song et al.

Food computing brings various perspectives to computer vision like vision-based food analysis for nutrition and health. As a fundamental task in food computing, food detection needs Zero-Shot Detection (ZSD) on novel unseen food objects to support real-world scenarios, such as intelligent kitchens and smart restaurants. Therefore, we first benchmark the task of Zero-Shot Food Detection (ZSFD) by introducing FOWA dataset with rich attribute annotations. Unlike ZSD, fine-grained problems in ZSFD like inter-class similarity make synthesized features inseparable. The complexity of food semantic attributes further makes it more difficult for current ZSD methods to distinguish various food categories. To address these problems, we propose a novel framework ZSFDet to tackle fine-grained problems by exploiting the interaction between complex attributes. Specifically, we model the correlation between food categories and attributes in ZSFDet by multi-source graphs to provide prior knowledge for distinguishing fine-grained features. Within ZSFDet, Knowledge-Enhanced Feature Synthesizer (KEFS) learns knowledge representation from multiple sources (e.g., ingredients correlation from knowledge graph) via the multi-source graph fusion. Conditioned on the fusion of semantic knowledge representation, the region feature diffusion model in KEFS can generate fine-grained features for training the effective zero-shot detector. Extensive evaluations demonstrate the superior performance of our method ZSFDet on FOWA and the widely-used food dataset UECFOOD-256, with significant improvements by 1.8% and 3.7% ZSD mAP compared with the strong baseline RRFS. Further experiments on PASCAL VOC and MS COCO prove that enhancement of the semantic knowledge can also improve the performance on general ZSD. Code and dataset are available at https://github.com/LanceZPF/KEFS.

Pengfei Zhou, Fangxiang Feng, Xiaojie Wang

Image harmonization, which involves adjusting the foreground of a composite image to attain a unified visual consistency with the background, can be conceptualized as an image-to-image translation task. Diffusion models have recently promoted the rapid development of image-to-image translation tasks . However, training diffusion models from scratch is computationally intensive. Fine-tuning pre-trained latent diffusion models entails dealing with the reconstruction error induced by the image compression autoencoder, making it unsuitable for image generation tasks that involve pixel-level evaluation metrics. To deal with these issues, in this paper, we first adapt a pre-trained latent diffusion model to the image harmonization task to generate the harmonious but potentially blurry initial images. Then we implement two strategies: utilizing higher-resolution images during inference and incorporating an additional refinement stage, to further enhance the clarity of the initially harmonized images. Extensive experiments on iHarmony4 datasets demonstrate the superiority of our proposed method. The code and model will be made publicly available at https://github.com/nicecv/DiffHarmony .

Pengfei Zhou, Fanrui Zhang, Xiaopeng Peng et al.

Multimodal reasoning, which integrates language and visual cues into problem solving and decision making, is a fundamental aspect of human intelligence and a crucial step toward artificial general intelligence. However, the evaluation of multimodal reasoning capabilities in Multimodal Large Language Models (MLLMs) remains inadequate. Most existing reasoning benchmarks are constrained by limited data size, narrow domain coverage, and unstructured knowledge distribution. To close these gaps, we introduce MDK12-Bench, a multi-disciplinary benchmark assessing the reasoning capabilities of MLLMs via real-world K-12 examinations. Spanning six disciplines (math, physics, chemistry, biology, geography, and information science), our benchmark comprises 140K reasoning instances across diverse difficulty levels from primary school to 12th grade. It features 6,827 instance-level knowledge point annotations based on a well-organized knowledge structure, detailed answer explanations, difficulty labels and cross-year partitions, providing a robust platform for comprehensive evaluation. Additionally, we present a novel dynamic evaluation framework to mitigate data contamination issues by bootstrapping question forms, question types, and image styles during evaluation. Extensive experiment on MDK12-Bench reveals the significant limitation of current MLLMs in multimodal reasoning. The findings on our benchmark provide insights into the development of the next-generation models. Our data and codes are available at https://github.com/LanceZPF/MDK12.

Ziqiao Wang, Wangbo Zhao, Yuhao Zhou et al.

Diffusion Transformers (DiTs) deliver state-of-the-art image quality, yet their training remains notoriously slow. A recent remedy -- representation alignment (REPA) that matches DiT hidden features to those of a non-generative teacher (e.g. DINO) -- dramatically accelerates the early epochs but plateaus or even degrades performance later. We trace this failure to a capacity mismatch: once the generative student begins modelling the joint data distribution, the teacher's lower-dimensional embeddings and attention patterns become a straitjacket rather than a guide. We then introduce HASTE (Holistic Alignment with Stage-wise Termination for Efficient training), a two-phase schedule that keeps the help and drops the hindrance. Phase I applies a holistic alignment loss that simultaneously distills attention maps (relational priors) and feature projections (semantic anchors) from the teacher into mid-level layers of the DiT, yielding rapid convergence. Phase II then performs one-shot termination that deactivates the alignment loss, once a simple trigger such as a fixed iteration is hit, freeing the DiT to focus on denoising and exploit its generative capacity. HASTE speeds up training of diverse DiTs without architecture changes. On ImageNet 256X256, it reaches the vanilla SiT-XL/2 baseline FID in 50 epochs and matches REPA's best FID in 500 epochs, amounting to a 28X reduction in optimization steps. HASTE also improves text-to-image DiTs on MS-COCO, demonstrating to be a simple yet principled recipe for efficient diffusion training across various tasks. Our code is available at https://github.com/NUS-HPC-AI-Lab/HASTE .

Jiaxin Ai, Pengfei Zhou, Zhaopan Xu et al.

As multi-modal large language models (MLLMs) frequently exhibit errors when solving scientific problems, evaluating the validity of their reasoning processes is critical for ensuring reliability and uncovering fine-grained model weaknesses. Since human evaluation is laborious and costly, prompting MLLMs as automated process judges has become a common practice. However, the reliability of these model-based judges remains uncertain. To address this, we introduce ProJudgeBench, the first comprehensive benchmark specifically designed for evaluating abilities of MLLM-based process judges. ProJudgeBench comprises 2,400 test cases and 50,118 step-level labels, spanning four scientific disciplines with diverse difficulty levels and multi-modal content. In ProJudgeBench, each step is meticulously annotated by human experts for correctness, error type, and explanation, enabling a systematic evaluation of judges' capabilities to detect, classify and diagnose errors. Evaluation on ProJudgeBench reveals a significant performance gap between open-source and proprietary models. To bridge this gap, we further propose ProJudge-173k, a large-scale instruction-tuning dataset, and a Dynamic Dual-Phase fine-tuning strategy that encourages models to explicitly reason through problem-solving before assessing solutions. Both contributions significantly enhance the process evaluation capabilities of open-source models. All the resources will be released to foster future research of reliable multi-modal process evaluation.

Junli Liang, Pengfei Zhou, Wangqiu Zhou et al.

Traditional Retrieval-Augmented Generation (RAG) effectively supports single-hop question answering with large language models but faces significant limitations in multi-hop question answering tasks, which require combining evidence from multiple documents. Existing chunk-based retrieval often provides irrelevant and logically incoherent context, leading to incomplete evidence chains and incorrect reasoning during answer generation. To address these challenges, we propose SentGraph, a sentence-level graph-based RAG framework that explicitly models fine-grained logical relationships between sentences for multi-hop question answering. Specifically, we construct a hierarchical sentence graph offline by first adapting Rhetorical Structure Theory to distinguish nucleus and satellite sentences, and then organizing them into topic-level subgraphs with cross-document entity bridges. During online retrieval, SentGraph performs graph-guided evidence selection and path expansion to retrieve fine-grained sentence-level evidence. Extensive experiments on four multi-hop question answering benchmarks demonstrate the effectiveness of SentGraph, validating the importance of explicitly modeling sentence-level logical dependencies for multi-hop reasoning.

Pengfei Zhou, Yunlong Liu, Junli Liang et al.

Time series forecasting with exogenous variables is a critical emerging paradigm that presents unique challenges in modeling dependencies between variables. Traditional models often struggle to differentiate between endogenous and exogenous variables, leading to inefficiencies and overfitting. In this paper, we introduce CrossLinear, a novel Linear-based forecasting model that addresses these challenges by incorporating a plug-and-play cross-correlation embedding module. This lightweight module captures the dependencies between variables with minimal computational cost and seamlessly integrates into existing neural networks. Specifically, it captures time-invariant and direct variable dependencies while disregarding time-varying or indirect dependencies, thereby mitigating the risk of overfitting in dependency modeling and contributing to consistent performance improvements. Furthermore, CrossLinear employs patch-wise processing and a global linear head to effectively capture both short-term and long-term temporal dependencies, further improving its forecasting precision. Extensive experiments on 12 real-world datasets demonstrate that CrossLinear achieves superior performance in both short-term and long-term forecasting tasks. The ablation study underscores the effectiveness of the cross-correlation embedding module. Additionally, the generalizability of this module makes it a valuable plug-in for various forecasting tasks across different domains. Codes are available at https://github.com/mumiao2000/CrossLinear.

Zheqi He, Xinya Wu, Pengfei Zhou et al.

Multi-modal large language models(MLLMs) have achieved remarkable progress and demonstrated powerful knowledge comprehension and reasoning abilities. However, the mastery of domain-specific knowledge, which is essential for evaluating the intelligence of MLLMs, continues to be a challenge. Current multi-modal benchmarks for domain-specific knowledge concentrate on multiple-choice questions and are predominantly available in English, which imposes limitations on the comprehensiveness of the evaluation. To this end, we introduce CMMU, a novel benchmark for multi-modal and multi-type question understanding and reasoning in Chinese. CMMU consists of 3,603 questions in 7 subjects, covering knowledge from primary to high school. The questions can be categorized into 3 types: multiple-choice, multiple-response, and fill-in-the-blank, bringing greater challenges to MLLMs. In addition, we propose an evaluation strategy called Positional Error Variance for assessing multiple-choice questions. The strategy aims to perform a quantitative analysis of position bias. We evaluate seven open-source MLLMs along with GPT4-V, Gemini-Pro, and Qwen-VL-Plus. The results demonstrate that CMMU poses a significant challenge to the recent MLLMs. The data and code are available at https://github.com/FlagOpen/CMMU.

Pengfei Zhou, Liliang Chen, Shengcong Chen et al.

Specifying robotic manipulation tasks in a manner that is both expressive and precise remains a central challenge. While visual goals provide a compact and unambiguous task specification, existing goal-conditioned policies often struggle with long-horizon manipulation due to their reliance on single-step action prediction without explicit modeling of task progress. We propose Act2Goal, a general goal-conditioned manipulation policy that integrates a goal-conditioned visual world model with multi-scale temporal control. Given a current observation and a target visual goal, the world model generates a plausible sequence of intermediate visual states that captures long-horizon structure. To translate this visual plan into robust execution, we introduce Multi-Scale Temporal Hashing (MSTH), which decomposes the imagined trajectory into dense proximal frames for fine-grained closed-loop control and sparse distal frames that anchor global task consistency. The policy couples these representations with motor control through end-to-end cross-attention, enabling coherent long-horizon behavior while remaining reactive to local disturbances. Act2Goal achieves strong zero-shot generalization to novel objects, spatial layouts, and environments. We further enable reward-free online adaptation through hindsight goal relabeling with LoRA-based finetuning, allowing rapid autonomous improvement without external supervision. Real-robot experiments demonstrate that Act2Goal improves success rates from 30% to 90% on challenging out-of-distribution tasks within minutes of autonomous interaction, validating that goal-conditioned world models with multi-scale temporal control provide structured guidance necessary for robust long-horizon manipulation. Project page: https://act2goal.github.io/

Fanqing Meng, Lingxiao Du, Zijian Wu et al.

Language-model agents are increasingly used as persistent coworkers that assist users across multiple working days. During such workflows, the surrounding environment may change independently of the agent: new emails arrive, calendar entries shift, knowledge-base records are updated, and evidence appears across images, scanned PDFs, audio, video, and spreadsheets. Existing benchmarks do not adequately evaluate this setting because they typically run within a single static episode and remain largely text-centric. We introduce \bench{}, a benchmark for coworker agents built around multi-turn multi-day tasks, a stateful sandboxed service environment whose state evolves between turns, and rule-based verification. The current release contains 100 tasks across 13 professional scenarios, executed against five stateful sandboxed services (filesystem, email, calendar, knowledge base, spreadsheet) and scored by 1537 deterministic Python checkers over post-execution service state; no LLM-as-judge is invoked during scoring. We benchmark seven frontier agent systems. The strongest model reaches 75.8 weighted score, but the best strict Task Success is only 20.0\%, indicating that partial progress is common while complete end-to-end workflow completion remains rare. Turn-level analysis shows that performance drops after the first exogenous environment update, highlighting adaptation to changing state as a key open challenge. We release the benchmark, evaluation harness, and construction pipeline to support reproducible coworker-agent evaluation.

Siyuan Huang, Liliang Chen, Pengfei Zhou et al.

We introduce EnerVerse, a generative robotics foundation model that constructs and interprets embodied spaces. EnerVerse employs a chunk-wise autoregressive video diffusion framework to predict future embodied spaces from instructions, enhanced by a sparse context memory for long-term reasoning. To model the 3D robotics world, we adopt a multi-view video representation, providing rich perspectives to address challenges like motion ambiguity and 3D grounding. Additionally, EnerVerse-D, a data engine pipeline combining generative modeling with 4D Gaussian Splatting, forms a self-reinforcing data loop to reduce the sim-to-real gap. Leveraging these innovations, EnerVerse translates 4D world representations into physical actions via a policy head (EnerVerse-A), achieving state-of-the-art performance in both simulation and real-world tasks. For efficiency, EnerVerse-A reuses features from the first denoising step and predicts action chunks, achieving about 280 ms per 8-step action chunk on a single RTX 4090. Further video demos, dataset samples could be found in our project page.

Xiaoda Yang, Majun Zhang, Changhao Pan et al.

Unified audio-visual generation is rapidly gaining industrial and creative relevance, enabling applications in virtual production and interactive media. However, when moving from general audio-video synthesis to music-dance co-generation, the task becomes substantially harder: musical rhythm, phrasing, and accents must drive choreographic motion at fine temporal resolution, and such rhythmic coupling is not captured by unimodal metrics or generic audiovisual consistency scores used in current evaluation practice. We introduce TMD-Bench, a benchmark for text-driven music-dance co-generation that assesses systems across unimodal generation quality, instruction adherence, and cross-modal rhythmic alignment. The benchmark integrates computable physical metrics with perceptual multimodal judgments, and is supported by a curated rhythm-aligned music-dance dataset and a fine-grained Music Captioner for structured music semantics. TMD-Bench further reveals that (i) modern commercial audio-visual models, such as Veo 3 and Sora 2, produce high-quality music and video, while rhythmic coupling remains less consistently optimized and leaves room for improvement, and (ii) our unified baseline RhyJAM trained on rhythm-aligned data achieves competitive beat-level synchronization while maintaining competitive unimodal fidelity. This presents prospects for building next-generation music-dance models that explicitly optimize rhythmic and kinetic coherence.

Xiyue Bai, Ronghao Yu, Jia Xiu et al.

Generating or editing images directly from Neural signals has immense potential at the intersection of neuroscience, vision, and Brain-computer interaction. In this paper, We present Uni-Neur2Img, a unified framework for neural signal-driven image generation and editing. The framework introduces a parameter-efficient LoRA-based neural signal injection module that independently processes each conditioning signal as a pluggable component, facilitating flexible multi-modal conditioning without altering base model parameters. Additionally, we employ a causal attention mechanism accommodate the long-sequence modeling demands of conditional generation tasks. Existing neural-driven generation research predominantly focuses on textual modalities as conditions or intermediate representations, resulting in limited exploration of visual modalities as direct conditioning signals. To bridge this research gap, we introduce the EEG-Style dataset. We conduct comprehensive evaluations across public benchmarks and self-collected neural signal datasets: (1) EEG-driven image generation on the public CVPR40 dataset; (2) neural signal-guided image editing on the public Loongx dataset for semantic-aware local modifications; and (3) EEG-driven style transfer on our self-collected EEG-Style dataset. Extensive experimental results demonstrate significant improvements in generation fidelity, editing consistency, and style transfer quality while maintaining low computational overhead and strong scalability to additional modalities. Thus, Uni-Neur2Img offers a unified, efficient, and extensible solution for bridging neural signals and visual content generation.

Pengfei Zhou, Xiaopeng Peng, Jiajun Song et al.

Multimodal Large Language Models (MLLMs) have made significant strides in visual understanding and generation tasks. However, generating interleaved image-text content remains a challenge, which requires integrated multimodal understanding and generation abilities. While the progress in unified models offers new solutions, existing benchmarks are insufficient for evaluating these methods due to limitations in data size and diversity. To bridge this gap, we introduce OpenING, a comprehensive benchmark comprising 5,400 high-quality human-annotated instances across 56 real-world tasks. OpenING covers diverse daily scenarios such as travel guide, design, and brainstorming, offering a robust platform for challenging interleaved generation methods. In addition, we present IntJudge, a judge model for evaluating open-ended multimodal generation methods. Trained with a novel data pipeline, our IntJudge achieves an agreement rate of 82.42% with human judgments, outperforming GPT-based evaluators by 11.34%. Extensive experiments on OpenING reveal that current interleaved generation methods still have substantial room for improvement. Key findings on interleaved image-text generation are further presented to guide the development of next-generation models.

Yue Liao, Pengfei Zhou, Siyuan Huang et al.

We introduce Genie Envisioner (GE), a unified world foundation platform for robotic manipulation that integrates policy learning, evaluation, and simulation within a single video-generative framework. At its core, GE-Base is a large-scale, instruction-conditioned video diffusion model that captures the spatial, temporal, and semantic dynamics of real-world robotic interactions in a structured latent space. Built upon this foundation, GE-Act maps latent representations to executable action trajectories through a lightweight, flow-matching decoder, enabling precise and generalizable policy inference across diverse embodiments with minimal supervision. To support scalable evaluation and training, GE-Sim serves as an action-conditioned neural simulator, producing high-fidelity rollouts for closed-loop policy development. The platform is further equipped with EWMBench, a standardized benchmark suite measuring visual fidelity, physical consistency, and instruction-action alignment. Together, these components establish Genie Envisioner as a scalable and practical foundation for instruction-driven, general-purpose embodied intelligence. All code, models, and benchmarks will be released publicly.

Yuxin Jiang, Shengcong Chen, Siyuan Huang et al.

Robotic imitation learning has advanced from solving static tasks to addressing dynamic interaction scenarios, but testing and evaluation remain costly and challenging due to the need for real-time interaction with dynamic environments. We propose EnerVerse-AC (EVAC), an action-conditional world model that generates future visual observations based on an agent's predicted actions, enabling realistic and controllable robotic inference. Building on prior architectures, EVAC introduces a multi-level action-conditioning mechanism and ray map encoding for dynamic multi-view image generation while expanding training data with diverse failure trajectories to improve generalization. As both a data engine and evaluator, EVAC augments human-collected trajectories into diverse datasets and generates realistic, action-conditioned video observations for policy testing, eliminating the need for physical robots or complex simulations. This approach significantly reduces costs while maintaining high fidelity in robotic manipulation evaluation. Extensive experiments validate the effectiveness of our method. Code, checkpoints, and datasets can be found at <https://annaj2178.github.io/EnerverseAC.github.io>.

Zekai Li, Xinhao Zhong, Samir Khaki et al.

In recent years, dataset distillation has provided a reliable solution for data compression, where models trained on the resulting smaller synthetic datasets achieve performance comparable to those trained on the original datasets. To further improve the performance of synthetic datasets, various training pipelines and optimization objectives have been proposed, greatly advancing the field of dataset distillation. Recent decoupled dataset distillation methods introduce soft labels and stronger data augmentation during the post-evaluation phase and scale dataset distillation up to larger datasets (e.g., ImageNet-1K). However, this raises a question: Is accuracy still a reliable metric to fairly evaluate dataset distillation methods? Our empirical findings suggest that the performance improvements of these methods often stem from additional techniques rather than the inherent quality of the images themselves, with even randomly sampled images achieving superior results. Such misaligned evaluation settings severely hinder the development of DD. Therefore, we propose DD-Ranking, a unified evaluation framework, along with new general evaluation metrics to uncover the true performance improvements achieved by different methods. By refocusing on the actual information enhancement of distilled datasets, DD-Ranking provides a more comprehensive and fair evaluation standard for future research advancements.

Zhaopan Xu, Pengfei Zhou, Jiaxin Ai et al.

Reasoning is an essential capacity for large language models (LLMs) to address complex tasks, where the identification of process errors is vital for improving this ability. Recently, process-level reward models (PRMs) were proposed to provide step-wise rewards that facilitate reinforcement learning and data production during training and guide LLMs toward correct steps during inference, thereby improving reasoning accuracy. However, existing benchmarks of PRMs are text-based and focus on error detection, neglecting other scenarios like reasoning search. To address this gap, we introduce MPBench, a comprehensive, multi-task, multimodal benchmark designed to systematically assess the effectiveness of PRMs in diverse scenarios. MPBench employs three evaluation paradigms, each targeting a specific role of PRMs in the reasoning process: (1) Step Correctness, which assesses the correctness of each intermediate reasoning step; (2) Answer Aggregation, which aggregates multiple solutions and selects the best one; and (3) Reasoning Process Search, which guides the search for optimal reasoning steps during inference. Through these paradigms, MPBench makes comprehensive evaluations and provides insights into the development of multimodal PRMs.

Pengfei Zhou, Xiangyue Zhang, Xukun Shen et al.

Masked generative models have become a strong paradigm for text-to-motion synthesis, but they still treat motion frames too uniformly during masking, attention, and decoding. This is a poor match for motion, where local dynamic complexity varies sharply over time. We show that current masked motion generators degrade disproportionately on dynamically complex motions, and that frame-wise generation error is strongly correlated with motion dynamics. Motivated by this mismatch, we introduce the Motion Spectral Descriptor (MSD), a simple and parameter-free measure of local dynamic complexity computed from the short-time spectrum of motion velocity. Unlike learned difficulty predictors, MSD is deterministic, interpretable, and derived directly from the motion signal itself. We use MSD to make masked motion generation complexity-aware. In particular, MSD guides content-focused masking during training, provides a spectral similarity prior for self-attention, and can additionally modulate token-level sampling during iterative decoding. Built on top of masked motion generators, our method, DynMask, improves motion generation most clearly on dynamically complex motions while also yielding stronger overall FID on HumanML3D and KIT-ML. These results suggest that respecting local motion complexity is a useful design principle for masked motion generation. Project page: https://xiangyue-zhang.github.io/DynMask

Pengfei Zhou, Xiaopeng Peng, Fanrui Zhang et al.

Multimodal large language models (MLLMs), which integrate language and visual cues for problem-solving, are crucial for advancing artificial general intelligence (AGI). However, current benchmarks for measuring the intelligence of MLLMs suffer from limited scale, narrow coverage, and unstructured knowledge, offering only static and undifferentiated evaluations. To bridge this gap, we introduce MDK12-Bench, a large-scale multidisciplinary benchmark built from real-world K-12 exams spanning six disciplines with 141K instances and 6,225 knowledge points organized in a six-layer taxonomy. Covering five question formats with difficulty and year annotations, it enables comprehensive evaluation to capture the extent to which MLLMs perform over four dimensions: 1) difficulty levels, 2) temporal (cross-year) shifts, 3) contextual shifts, and 4) knowledge-driven reasoning. We propose a novel dynamic evaluation framework that introduces unfamiliar visual, textual, and question form shifts to challenge model generalization while improving benchmark objectivity and longevity by mitigating data contamination. We further evaluate knowledge-point reference-augmented generation (KP-RAG) to examine the role of knowledge in problem-solving. Key findings reveal limitations in current MLLMs in multiple aspects and provide guidance for enhancing model robustness, interpretability, and AI-assisted education.

Zhaopan Xu, Pengfei Zhou, Weidong Tang et al.

Multimodal large language models (MLLMs) have achieved remarkable success in vision-language tasks, but their reliance on vast, internet-sourced data raises significant privacy and security concerns. Machine unlearning (MU) has emerged as a critical technique to address these issues, enabling the selective removal of targeted information from pre-trained models without costly retraining. However, the evaluation of MU for MLLMs remains inadequate. Existing benchmarks often lack a comprehensive scope, focusing narrowly on entities while overlooking the unlearning of broader visual concepts and the inherent semantic coupling between them. To bridge this gap, we introduce, PEBench, a novel benchmark designed to facilitate a thorough assessment of MU in MLLMs. PEBench features a fictitious dataset of personal entities and corresponding event scenes to evaluate unlearning across these distinct yet entangled concepts. We leverage this benchmark to evaluate five MU methods, revealing their unique strengths and weaknesses. Our findings show that unlearning one concept can unintentionally degrade performance on related concepts within the same image, a challenge we term cross-concept interference. Furthermore, we demonstrate the difficulty of unlearning person and event concepts simultaneously and propose an effective method to mitigate these conflicting objectives. The source code and benchmark are publicly available at https://pebench.github.io.

Wangbo Zhao, Yizeng Han, Zhiwei Tang et al.

Diffusion Transformers (DiTs) excel at visual generation yet remain hampered by slow sampling. Existing training-free accelerators - step reduction, feature caching, and sparse attention - enhance inference speed but typically rely on a uniform heuristic or a manually designed adaptive strategy for all images, leaving quality on the table. Alternatively, dynamic neural networks offer per-image adaptive acceleration, but their high fine-tuning costs limit broader applicability. To address these limitations, we introduce RAPID3: Tri-Level Reinforced Acceleration Policies for Diffusion Transformers, a framework that delivers image-wise acceleration with zero updates to the base generator. Specifically, three lightweight policy heads - Step-Skip, Cache-Reuse, and Sparse-Attention - observe the current denoising state and independently decide their corresponding speed-up at each timestep. All policy parameters are trained online via Group Relative Policy Optimization (GRPO) while the generator remains frozen. Meanwhile, an adversarially learned discriminator augments the reward signal, discouraging reward hacking by boosting returns only when generated samples stay close to the original model's distribution. Across state-of-the-art DiT backbones, including Stable Diffusion 3 and FLUX, RAPID3 achieves nearly 3x faster sampling with competitive generation quality.

Pengfei Zhou, Jie Xia, Xiaopeng Peng et al.

Traditional image editing typically relies on manual prompting, making it labor-intensive and inaccessible to individuals with limited motor control or language abilities. Leveraging recent advances in brain-computer interfaces (BCIs) and generative models, we propose LoongX, a hands-free image editing approach driven by multimodal neurophysiological signals. LoongX utilizes state-of-the-art diffusion models trained on a comprehensive dataset of 23,928 image editing pairs, each paired with synchronized electroencephalography (EEG), functional near-infrared spectroscopy (fNIRS), photoplethysmography (PPG), and head motion signals that capture user intent. To effectively address the heterogeneity of these signals, LoongX integrates two key modules. The cross-scale state space (CS3) module encodes informative modality-specific features. The dynamic gated fusion (DGF) module further aggregates these features into a unified latent space, which is then aligned with edit semantics via fine-tuning on a diffusion transformer (DiT). Additionally, we pre-train the encoders using contrastive learning to align cognitive states with semantic intentions from embedded natural language. Extensive experiments demonstrate that LoongX achieves performance comparable to text-driven methods (CLIP-I: 0.6605 vs. 0.6558; DINO: 0.4812 vs. 0.4636) and outperforms them when neural signals are combined with speech (CLIP-T: 0.2588 vs. 0.2549). These results highlight the promise of neural-driven generative models in enabling accessible, intuitive image editing and open new directions for cognitive-driven creative technologies. Datasets and code will be released to support future work and foster progress in this emerging area.

Pengfei Zhou, Xukun Shen, Yong Hu

Recent advances in Score Distillation Sampling (SDS) have improved 3D human generation from textual descriptions. However, existing methods still face challenges in accurately aligning 3D models with long and complex textual inputs. To address this challenge, we propose a novel framework that introduces contrastive preferences, where human-level preference models, guided by both positive and negative prompts, assist SDS for improved alignment. Specifically, we design a preference optimization module that integrates multiple models to comprehensively capture the full range of textual features. Furthermore, we introduce a negation preference module to mitigate over-optimization of irrelevant details by leveraging static-dynamic negation prompts, effectively preventing ``reward hacking". Extensive experiments demonstrate that our method achieves state-of-the-art results, significantly enhancing texture realism and visual alignment with textual descriptions, particularly for long and complex inputs.

Yukai Song, Pengfei Zhou, César Escobar-Viera et al.

Suicide rates have risen worldwide in recent years, underscoring the urgent need for proactive prevention strategies. Social media provides valuable signals, as many at-risk individuals - who often avoid formal help due to stigma - choose instead to share their distress online. Yet detecting implicit suicidal ideation, conveyed indirectly through metaphor, sarcasm, or subtle emotional cues, remains highly challenging. Lightweight models like BERT handle explicit signals but fail on subtle implicit ones, while large language models (LLMs) capture nuance at prohibitive computational cost. To address this gap, we propose a two-stage voting architecture that balances efficiency and robustness. In Stage 1, a lightweight BERT classifier rapidly resolves high-confidence explicit cases. In Stage 2, ambiguous inputs are escalated to either (i) a multi-perspective LLM voting framework to maximize recall on implicit ideation, or (ii) a feature-based ML ensemble guided by psychologically grounded indicators extracted via prompt-engineered LLMs for efficiency and interpretability. To the best of our knowledge, this is among the first works to operationalize LLM-extracted psychological features as structured vectors for suicide risk detection. On two complementary datasets - explicit-dominant Reddit and implicit-only DeepSuiMind - our framework outperforms single-model baselines, achieving 98.0% F1 on explicit cases, 99.7% on implicit ones, and reducing the cross-domain gap below 2%, while significantly lowering LLM cost.

Suorong Yang, Jie Zong, Lihang Wang et al.

Data augmentation has been widely employed to improve the generalization of deep neural networks. Most existing methods apply fixed or random transformations. However, we find that sample difficulty evolves along with the model's generalization capabilities in dynamic training environments. As a result, applying uniform or stochastic augmentations, without accounting for such dynamics, can lead to a mismatch between augmented data and the model's evolving training needs, ultimately degrading training effectiveness. To address this, we introduce SADA, a Sample-Aware Dynamic Augmentation that performs on-the-fly adjustment of augmentation strengths based on each sample's evolving influence on model optimization. Specifically, we estimate each sample's influence by projecting its gradient onto the accumulated model update direction and computing the temporal variance within a local training window. Samples with low variance, indicating stable and consistent influence, are augmented more strongly to emphasize diversity, while unstable samples receive milder transformations to preserve semantic fidelity and stabilize learning. Our method is lightweight, which does not require auxiliary models or policy tuning. It can be seamlessly integrated into existing training pipelines as a plug-and-play module. Experiments across various benchmark datasets and model architectures show consistent improvements of SADA, including +7.3\% on fine-grained tasks and +4.3\% on long-tailed datasets, highlighting the method's effectiveness and practicality.

Hui Ji, Wei Gao, Pengfei Zhou

The 12-lead electrocardiogram (ECG) is the gold standard for cardiovascular monitoring, offering superior diagnostic granularity and specificity compared to photoplethysmography (PPG). However, existing 12-lead ECG systems rely on cumbersome multi-electrode setups, limiting sustained monitoring in ambulatory settings, while current PPG-based methods fail to reconstruct multi-lead ECG due to the absence of inter-lead constraints and insufficient modeling of spatial-temporal dependencies across leads. To bridge this gap, we introduce P2Es, an innovative demographic-aware diffusion framework designed to generate clinically valid 12-lead ECG from PPG signals via three key innovations. Specifically, in the forward process, we introduce frequency-domain blurring followed by temporal noise interference to simulate real-world signal distortions. In the reverse process, we design a temporal multi-scale generation module followed by frequency deblurring. In particular, we leverage KNN-based clustering combined with contrastive learning to assign affinity matrices for the reverse process, enabling demographic-specific ECG translation. Extensive experimental results show that P2Es outperforms baseline models in 12-lead ECG reconstruction.

Pengfei Zhou, Weiqing Min, Chaoran Fu et al.

Food is foundational to human life, serving not only as a source of nourishment but also as a cornerstone of cultural identity and social interaction. As the complexity of global dietary needs and preferences grows, food intelligence is needed to enable food perception and reasoning for various tasks, ranging from recipe generation and dietary recommendation to diet-disease correlation discovery and understanding. Towards this goal, for powerful capabilities across various domains and tasks in Large Language Models (LLMs), we introduce Food-oriented LLM FoodSky to comprehend food data through perception and reasoning. Considering the complexity and typicality of Chinese cuisine, we first construct one comprehensive Chinese food corpus FoodEarth from various authoritative sources, which can be leveraged by FoodSky to achieve deep understanding of food-related data. We then propose Topic-based Selective State Space Model (TS3M) and the Hierarchical Topic Retrieval Augmented Generation (HTRAG) mechanism to enhance FoodSky in capturing fine-grained food semantics and generating context-aware food-relevant text, respectively. Our extensive evaluations demonstrate that FoodSky significantly outperforms general-purpose LLMs in both chef and dietetic examinations, with an accuracy of 67.2% and 66.4% on the Chinese National Chef Exam and the National Dietetic Exam, respectively. FoodSky not only promises to enhance culinary creativity and promote healthier eating patterns, but also sets a new standard for domain-specific LLMs that address complex real-world issues in the food domain. An online demonstration of FoodSky is available at http://222.92.101.211:8200.

Hanyu Zeng, Pengfei Zhou, Xin Lou et al.

Modern power grids are undergoing significant changes driven by information and communication technologies (ICTs), and evolving into smart grids with higher efficiency and lower operation cost. Using ICTs, however, comes with an inevitable side effect that makes the power system more vulnerable to cyber attacks. In this paper, we propose a self-supervised learning-based framework to detect and identify various types of cyber attacks. Different from existing approaches, the proposed framework does not rely on large amounts of well-curated labeled data but makes use of the massive unlabeled data in the wild which are easily accessible. Specifically, the proposed framework adopts the BERT model from the natural language processing domain and learns generalizable and effective representations from the unlabeled sensing data, which capture the distinctive patterns of different attacks. Using the learned representations, together with a very small amount of labeled data, we can train a task-specific classifier to detect various types of cyber attacks. Meanwhile, real-world training datasets are usually imbalanced, i.e., there are only a limited number of data samples containing attacks. In order to cope with such data imbalance, we propose a new loss function, separate mean error (SME), which pays equal attention to the large and small categories to better train the model. Experiment results in a 5-area power grid system with 37 buses demonstrate the superior performance of our framework over existing approaches, especially when a very limited portion of labeled data are available, e.g., as low as 0.002\%. We believe such a framework can be easily adopted to detect a variety of cyber attacks in other power grid scenarios.

Kaining Ying, Zhenhua Wang, Cong Bai et al.

Most instance segmentation models are not end-to-end trainable due to either the incorporation of proposal estimation (RPN) as a pre-processing or non-maximum suppression (NMS) as a post-processing. Here we propose a novel end-to-end instance segmentation method termed ISDA. It reshapes the task into predicting a set of object masks, which are generated via traditional convolution operation with learned position-aware kernels and features of objects. Such kernels and features are learned by leveraging a deformable attention network with multi-scale representation. Thanks to the introduced set-prediction mechanism, the proposed method is NMS-free. Empirically, ISDA outperforms Mask R-CNN (the strong baseline) by 2.6 points on MS-COCO, and achieves leading performance compared with recent models. Code will be available soon.

Sujie Li, Feng Pan, Pengfei Zhou et al.

Restricted Boltzmann machines (RBM) and deep Boltzmann machines (DBM) are important models in machine learning, and recently found numerous applications in quantum many-body physics. We show that there are fundamental connections between them and tensor networks. In particular, we demonstrate that any RBM and DBM can be exactly represented as a two-dimensional tensor network. This representation gives an understanding of the expressive power of RBM and DBM using entanglement structures of the tensor networks, also provides an efficient tensor network contraction algorithm for the computing partition function of RBM and DBM. Using numerical experiments, we demonstrate that the proposed algorithm is much more accurate than the state-of-the-art machine learning methods in estimating the partition function of restricted Boltzmann machines and deep Boltzmann machines, and have potential applications in training deep Boltzmann machines for general machine learning tasks.

Jie Zhang, Pengfei Zhou, Hongyan Wu

Graph-structured data and their related algorithms have attracted significant attention in many fields, such as influenza prediction in public health. However, the variable influenza seasonality, occasional pandemics, and domain knowledge pose great challenges to construct an appropriate graph, which could impair the strength of the current popular graph-based algorithms to perform data analysis. In this study, we develop a novel method, Dynamic Virtual Graph Significance Networks (DVGSN), which can supervisedly and dynamically learn from similar "infection situations" in historical timepoints. Representation learning on the dynamic virtual graph can tackle the varied seasonality and pandemics, and therefore improve the performance. The extensive experiments on real-world influenza data demonstrate that DVGSN significantly outperforms the current state-of-the-art methods. To the best of our knowledge, this is the first attempt to supervisedly learn a dynamic virtual graph for time-series prediction tasks. Moreover, the proposed method needs less domain knowledge to build a graph in advance and has rich interpretability, which makes the method more acceptable in the fields of public health, life sciences, and so on.

Pengfei Zhou, Tianyi Li, Pan Zhang

We perform theoretical and algorithmic studies for the problem of clustering and semi-supervised classification on graphs with both pairwise relational information and single-point feature information, upon a joint stochastic block model for generating synthetic graphs with both edges and node features. Asymptotically exact analysis based on the Bayesian inference of the underlying model are conducted, using the cavity method in statistical physics. Theoretically, we identify a phase transition of the generative model, which puts fundamental limits on the ability of all possible algorithms in the clustering task of the underlying model. Algorithmically, we propose a belief propagation algorithm that is asymptotically optimal on the generative model, and can be further extended to a belief propagation graph convolution neural network (BPGCN) for semi-supervised classification on graphs. For the first time, well-controlled benchmark datasets with asymptotially exact properties and optimal solutions could be produced for the evaluation of graph convolution neural networks, and for the theoretical understanding of their strengths and weaknesses. In particular, on these synthetic benchmark networks we observe that existing graph convolution neural networks are subject to an sparsity issue and an ovefitting issue in practice, both of which are successfully overcome by our BPGCN. Moreover, when combined with classic neural network methods, BPGCN yields extraordinary classification performances on some real-world datasets that have never been achieved before.

Feng Pan, Pengfei Zhou, Hai-Jun Zhou et al.

We propose a method for solving statistical mechanics problems defined on sparse graphs. It extracts a small Feedback Vertex Set (FVS) from the sparse graph, converting the sparse system to a much smaller system with many-body and dense interactions with an effective energy on every configuration of the FVS, then learns a variational distribution parameterized using neural networks to approximate the original Boltzmann distribution. The method is able to estimate free energy, compute observables, and generate unbiased samples via direct sampling without auto-correlation. Extensive experiments show that our approach is more accurate than existing approaches for sparse spin glasses. On random graphs and real-world networks, our approach significantly outperforms the standard methods for sparse systems such as the belief-propagation algorithm; on structured sparse systems such as two-dimensional lattices our approach is significantly faster and more accurate than recently proposed variational autoregressive networks using convolution neural networks.