arXiv subjects

Di Zhang

Publications and source records attributed to Di Zhang.

At least 181 records · Page 10Linked to original sources

arXiv2025

KineDex: Learning Tactile-Informed Visuomotor Policies via Kinesthetic Teaching for Dexterous Manipulation

Collecting demonstrations enriched with fine-grained tactile information is critical for dexterous manipulation, particularly in contact-rich tasks that require precise force control and physical interaction. While prior works primarily focus on teleoperation or video-based retargeting, they often suffer from kinematic mismatches and the absence of real-time tactile feedback, hindering the acquisition of high-fidelity tactile data. To mitigate this issue, we propose KineDex, a hand-over-hand kinesthetic teaching paradigm in which the operator's motion is directly transferred to the dexterous hand, enabling the collection of physically grounded demonstrations enriched with accurate tactile feedback. To resolve occlusions from human hand, we apply inpainting technique to preprocess the visual observations. Based on these demonstrations, we then train a visuomotor policy using tactile-augmented inputs and implement force control during deployment for precise contact-rich manipulation. We evaluate KineDex on a suite of challenging contact-rich manipulation tasks, including particularly difficult scenarios such as squeezing toothpaste onto a toothbrush, which require precise multi-finger coordination and stable force regulation. Across these tasks, KineDex achieves an average success rate of 74.4%, representing a 57.7% improvement over the variant without force control. Comparative experiments with teleoperation and user studies further validate the advantages of KineDex in data collection efficiency and operability. Specifically, KineDex collects data over twice as fast as teleoperation across two tasks of varying difficulty, while maintaining a near-100% success rate, compared to under 50% for teleoperation.

cs.RO↗

arXiv2025

A Survey of Interactive Generative Video

Interactive Generative Video (IGV) has emerged as a crucial technology in response to the growing demand for high-quality, interactive video content across various domains. In this paper, we define IGV as a technology that combines generative capabilities to produce diverse high-quality video content with interactive features that enable user engagement through control signals and responsive feedback. We survey the current landscape of IGV applications, focusing on three major domains: 1) gaming, where IGV enables infinite exploration in virtual worlds; 2) embodied AI, where IGV serves as a physics-aware environment synthesizer for training agents in multimodal interaction with dynamically evolving scenes; and 3) autonomous driving, where IGV provides closed-loop simulation capabilities for safety-critical testing and validation. To guide future development, we propose a comprehensive framework that decomposes an ideal IGV system into five essential modules: Generation, Control, Memory, Dynamics, and Intelligence. Furthermore, we systematically analyze the technical challenges and future directions in realizing each component for an ideal IGV system, such as achieving real-time generation, enabling open-domain control, maintaining long-term coherence, simulating accurate physics, and integrating causal reasoning. We believe that this systematic analysis will facilitate future research and development in the field of IGV, ultimately advancing the technology toward more sophisticated and practical applications.

cs.CV↗

arXiv2025

Koala-36M: A Large-scale Video Dataset Improving Consistency between Fine-grained Conditions and Video Content

With the continuous progress of visual generation technologies, the scale of video datasets has grown exponentially. The quality of these datasets plays a pivotal role in the performance of video generation models. We assert that temporal splitting, detailed captions, and video quality filtering are three crucial determinants of dataset quality. However, existing datasets exhibit various limitations in these areas. To address these challenges, we introduce Koala-36M, a large-scale, high-quality video dataset featuring accurate temporal splitting, detailed captions, and superior video quality. The essence of our approach lies in improving the consistency between fine-grained conditions and video content. Specifically, we employ a linear classifier on probability distributions to enhance the accuracy of transition detection, ensuring better temporal consistency. We then provide structured captions for the splitted videos, with an average length of 200 words, to improve text-video alignment. Additionally, we develop a Video Training Suitability Score (VTSS) that integrates multiple sub-metrics, allowing us to filter high-quality videos from the original corpus. Finally, we incorporate several metrics into the training process of the generation model, further refining the fine-grained conditions. Our experiments demonstrate the effectiveness of our data processing pipeline and the quality of the proposed Koala-36M dataset. Our dataset and code have been released at https://koala36m.github.io/.

cs.CV↗

arXiv2025

Critic-V: VLM Critics Help Catch VLM Errors in Multimodal Reasoning

Vision-language models (VLMs) have shown remarkable advancements in multimodal reasoning tasks. However, they still often generate inaccurate or irrelevant responses due to issues like hallucinated image understandings or unrefined reasoning paths. To address these challenges, we introduce Critic-V, a novel framework inspired by the Actor-Critic paradigm to boost the reasoning capability of VLMs. This framework decouples the reasoning process and critic process by integrating two independent components: the Reasoner, which generates reasoning paths based on visual and textual inputs, and the Critic, which provides constructive critique to refine these paths. In this approach, the Reasoner generates reasoning responses according to text prompts, which can evolve iteratively as a policy based on feedback from the Critic. This interaction process was theoretically driven by a reinforcement learning framework where the Critic offers natural language critiques instead of scalar rewards, enabling more nuanced feedback to boost the Reasoner's capability on complex reasoning tasks. The Critic model is trained using Direct Preference Optimization (DPO), leveraging a preference dataset of critiques ranked by Rule-based Reward~(RBR) to enhance its critic capabilities. Evaluation results show that the Critic-V framework significantly outperforms existing methods, including GPT-4V, on 5 out of 8 benchmarks, especially regarding reasoning accuracy and efficiency. Combining a dynamic text-based policy for the Reasoner and constructive feedback from the preference-optimized Critic enables a more reliable and context-sensitive multimodal reasoning process. Our approach provides a promising solution to enhance the reliability of VLMs, improving their performance in real-world reasoning-heavy multimodal applications such as autonomous driving and embodied intelligence.

cs.CV↗

arXiv2025

Joint Security-Latency Design for Short Packet-Based Low-Altitude Communications

In this article, a joint security and latency analysis of short packet-based low-altitude communications when the eavesdropper is close to the receiver is addressed. To reveal the impacts of the signal-to-noise ratio (SNR) and block-length on latency in communications, we propose a new metric named secure latency (SL) and derive the expressions for the effective secure probability (ESP) and the average SL. To minimize the average SL, different transmission designs are analyzed, in which the optimal solutions of SNR and block-length are provided. Numerical results validate our analysis and reveal the trade-off between reliability and security and the impacts of the block-length, SNR, and packet-generating rate on average SL, of which SNR and the block-length account for main factors. In addition, we find that the performance of SL can be enhanced by allocating less SNR.

eess.SP↗

arXiv2025

InstructEngine: Instruction-driven Text-to-Image Alignment

Reinforcement Learning from Human/AI Feedback (RLHF/RLAIF) has been extensively utilized for preference alignment of text-to-image models. Existing methods face certain limitations in terms of both data and algorithm. For training data, most approaches rely on manual annotated preference data, either by directly fine-tuning the generators or by training reward models to provide training signals. However, the high annotation cost makes them difficult to scale up, the reward model consumes extra computation and cannot guarantee accuracy. From an algorithmic perspective, most methods neglect the value of text and only take the image feedback as a comparative signal, which is inefficient and sparse. To alleviate these drawbacks, we propose the InstructEngine framework. Regarding annotation cost, we first construct a taxonomy for text-to-image generation, then develop an automated data construction pipeline based on it. Leveraging advanced large multimodal models and human-defined rules, we generate 25K text-image preference pairs. Finally, we introduce cross-validation alignment method, which refines data efficiency by organizing semantically analogous samples into mutually comparable pairs. Evaluations on DrawBench demonstrate that InstructEngine improves SD v1.5 and SDXL's performance by 10.53% and 5.30%, outperforming state-of-the-art baselines, with ablation study confirming the benefits of InstructEngine's all components. A win rate of over 50% in human reviews also proves that InstructEngine better aligns with human preferences.

cs.CV↗

arXiv2025

VLM as Policy: Common-Law Content Moderation Framework for Short Video Platform

Exponentially growing short video platforms (SVPs) face significant challenges in moderating content detrimental to users' mental health, particularly for minors. The dissemination of such content on SVPs can lead to catastrophic societal consequences. Although substantial efforts have been dedicated to moderating such content, existing methods suffer from critical limitations: (1) Manual review is prone to human bias and incurs high operational costs. (2) Automated methods, though efficient, lack nuanced content understanding, resulting in lower accuracy. (3) Industrial moderation regulations struggle to adapt to rapidly evolving trends due to long update cycles. In this paper, we annotate the first SVP content moderation benchmark with authentic user/reviewer feedback to fill the absence of benchmark in this field. Then we evaluate various methods on the benchmark to verify the existence of the aforementioned limitations. We further propose our common-law content moderation framework named KuaiMod to address these challenges. KuaiMod consists of three components: training data construction, offline adaptation, and online deployment & refinement. Leveraging large vision language model (VLM) and Chain-of-Thought (CoT) reasoning, KuaiMod adequately models video toxicity based on sparse user feedback and fosters dynamic moderation policy with rapid update speed and high accuracy. Offline experiments and large-scale online A/B test demonstrates the superiority of KuaiMod: KuaiMod achieves the best moderation performance on our benchmark. The deployment of KuaiMod reduces the user reporting rate by 20% and its application in video recommendation increases both Daily Active User (DAU) and APP Usage Time (AUT) on several Kuaishou scenarios. We have open-sourced our benchmark at https://kuaimod.github.io.

cs.SI↗

arXiv2025

PH2ST:ST-Prompt Guided Histological Hypergraph Learning for Spatial Gene Expression Prediction

Spatial Transcriptomics (ST) reveals the spatial distribution of gene expression in tissues, offering critical insights into biological processes and disease mechanisms. However, the high cost, limited coverage, and technical complexity of current ST technologies restrict their widespread use in clinical and research settings, making obtaining high-resolution transcriptomic profiles across large tissue areas challenging. Predicting ST from H\&E-stained histology images has emerged as a promising alternative to address these limitations but remains challenging due to the heterogeneous relationship between histomorphology and gene expression, which is affected by substantial variability across patients and tissue sections. In response, we propose PH2ST, a prompt-guided hypergraph learning framework, which leverages limited ST signals to guide multi-scale histological representation learning for accurate and robust spatial gene expression prediction. Extensive evaluations on two public ST datasets and multiple prompt sampling strategies simulating real-world scenarios demonstrate that PH2ST not only outperforms existing state-of-the-art methods, but also shows strong potential for practical applications such as imputing missing spots, ST super-resolution, and local-to-global prediction, highlighting its value for scalable and cost-effective spatial gene expression mapping in biomedical contexts.

cs.CV↗

arXiv2025

SlimPipe: Memory-Thrifty and Efficient Pipeline Parallelism for Long-Context LLM Training

Pipeline Parallelism (PP) serves as a crucial technique for training Large Language Models (LLMs), owing to its capability to alleviate memory pressure from model states with relatively low communication overhead. However, in long-context scenarios, existing pipeline parallelism methods fail to address the substantial activation memory pressure, primarily due to the peak memory consumption resulting from the accumulation of activations across multiple microbatches. Moreover, these approaches inevitably introduce considerable pipeline bubbles, further hindering efficiency. To tackle these challenges, we propose SlimPipe, a novel approach to fine-grained pipeline parallelism that employs uniform sequence slicing coupled with one-forward-one-backward (1F1B) schedule. It reduces the accumulated activations from several microbatches to just one, which is split into several slices. Although the slices are evenly partitioned, the computation cost is not equal across slices due to causal attention. We develop a sophisticated workload redistribution technique to address this load imbalance. SlimPipe achieves (1) near-zero memory overhead and (2) minimal pipeline bubbles simultaneously. The effectiveness of SlimPipe has been proven by thorough testing with diverse model architectures, context window sizes, and SlimPipe-specific configurations. For example, on the Llama 70B model, compared to state-of-the-art methods, SlimPipe significantly boosts the Model FLOPs Utilization (MFU) to up to $1.57\times$ for a context length of 512K. More notably, for a context length of 2048K, it maintains over 45% utilization on 256 NVIDIA Hopper 80GB GPUs, while other approaches either suffer significant performance drops or fail entirely due to memory constraints.

cs.LG↗

arXiv2025

E-TRIALS: Empowering Data-Driven Decisions to Enhance Computer-Based Learning Platforms

Computer-based learning platforms (CBLPs) have become a common medium in schools, transforming how students learn and interact with educational content. However, researchers still lack adequate tools to address the diverse set of challenges that students face in these environments. In this paper, we introduce \textbf{Ed-Tech Research Infrastructure to Advance Learning Sciences (E-TRIALS)}, a free tool developed by ASSISTments to help researchers conduct randomized controlled trials in the realm of learning sciences. We describe its features, the types of experiments it supports, and how it can address critical research questions. We showcase E-TRIALS' capabilities through two real-world interventions. Finally, we evaluate the efficacy of interventions using three average treatment effect (ATE) estimators. Student's t-test, regression, and Leave-One-Out Potential outcomes (LOOP). The results demonstrate that the unbiased LOOP estimator can achieve greater precision by adjusting for baseline covariates compared to the Student's t test. Our work demonstrates the potential of E-TRIALS to advance research and contribute to the development of more effective, inclusive, and adaptive CBLP. The code used for this work is available at https://osf.io/xp6ch/.

stat.AP↗

arXiv2025

Threshold Effects on the Massless Neutrino in the Canonical Seesaw Mechanism

In this work, we revisit the one-loop renormalization group equations (RGEs) among non-degenerate seesaw scales, i.e., threshold effects in the canonical seesaw mechanism, which have been obtained for more than two decades. Different from the previous work only focusing on the Weinberg operator, we derive the complete one-loop RGEs of all three dimension-five operators in the Standard Model effective field theory with right-handed neutrinos ($ν$SMEFT) and apply them to threshold effects in the canonical seesaw mechanism. We find some contributions from the Weinberg operator to its Wilson coefficient, the neutrino Yukawa coupling matrix, and the Higgs quartic coupling absent in the previous calculations. Based on the updated one-loop RGEs, we derive the RGE of the effective neutrino mass matrix's determinant without any approximation. Then, for the first time, we provide a strict proof that the one-loop RG running effects among non-degenerate seesaw scales can not generate a non-zero mass for the initial massless neutrino in the minimal type-I seesaw mechanism or in the canonical one with a rank-degenerate neutrino Yukawa coupling matrix. One has to include two- or higher-loop corrections to achieve a non-zero mass for the massless neutrino.

hep-ph↗

arXiv2025

SketchVideo: Sketch-based Video Generation and Editing

Video generation and editing conditioned on text prompts or images have undergone significant advancements. However, challenges remain in accurately controlling global layout and geometry details solely by texts, and supporting motion control and local modification through images. In this paper, we aim to achieve sketch-based spatial and motion control for video generation and support fine-grained editing of real or synthetic videos. Based on the DiT video generation model, we propose a memory-efficient control structure with sketch control blocks that predict residual features of skipped DiT blocks. Sketches are drawn on one or two keyframes (at arbitrary time points) for easy interaction. To propagate such temporally sparse sketch conditions across all frames, we propose an inter-frame attention mechanism to analyze the relationship between the keyframes and each video frame. For sketch-based video editing, we design an additional video insertion module that maintains consistency between the newly edited content and the original video's spatial feature and dynamic motion. During inference, we use latent fusion for the accurate preservation of unedited regions. Extensive experiments demonstrate that our SketchVideo achieves superior performance in controllable video generation and editing.

cs.GR↗

arXiv2025

Neutrino Theory in the Precision Era

This document summarises discussions on future directions in theoretical neutrino physics, which are the outcome of a neutrino theory workshop held at CERN in February 2025. The starting point is the realisation that neutrino physics offers unique opportunities to address some of the most fundamental questions in physics. This motivates a vigorous experimental programme which the theory community fully supports. \textbf{A strong effort in theoretical neutrino physics is paramount to optimally take advantage of upcoming neutrino experiments and to explore the synergies with other areas of particle, astroparticle, and nuclear physics, as well as cosmology.} Progress on the theory side has the potential to significantly boost the physics reach of experiments, as well as go well beyond their original scope. Strong collaboration between theory and experiment is essential in the precision era. To foster such collaboration, \textbf{we propose to establish a CERN Neutrino Physics Centre.} Taking inspiration from the highly successful LHC Physics Center at Fermilab, the CERN Neutrino Physics Centre would be the European hub of the neutrino community, covering experimental and theoretical activities.

hep-ph↗

arXiv2025

SARGes: Semantically Aligned Reliable Gesture Generation via Intent Chain

Co-speech gesture generation enhances human-computer interaction realism through speech-synchronized gesture synthesis. However, generating semantically meaningful gestures remains a challenging problem. We propose SARGes, a novel framework that leverages large language models (LLMs) to parse speech content and generate reliable semantic gesture labels, which subsequently guide the synthesis of meaningful co-speech gestures.First, we constructed a comprehensive co-speech gesture ethogram and developed an LLM-based intent chain reasoning mechanism that systematically parses and decomposes gesture semantics into structured inference steps following ethogram criteria, effectively guiding LLMs to generate context-aware gesture labels. Subsequently, we constructed an intent chain-annotated text-to-gesture label dataset and trained a lightweight gesture label generation model, which then guides the generation of credible and semantically coherent co-speech gestures. Experimental results demonstrate that SARGes achieves highly semantically-aligned gesture labeling (50.2% accuracy) with efficient single-pass inference (0.4 seconds). The proposed method provides an interpretable intent reasoning pathway for semantic gesture synthesis.

cs.CL↗

arXiv2025

A Stochastic Conjugate Subgradient Algorithm for Two-stage Stochastic Programming

Stochastic Optimization is a cornerstone of operations research, providing a framework to solve optimization problems under uncertainty. Despite the development of numerous algorithms to tackle these problems, several persistent challenges remain, including: (i) selecting an appropriate sample size, (ii) determining an effective search direction, and (iii) choosing a proper step size. This paper introduces a comprehensive framework, the Stochastic Conjugate Subgradient (SCS) framework, designed to systematically address these challenges. Specifically, The SCS framework offers structured approaches to determining the sample size, the search direction, and the step size. By integrating various stochastic algorithms within the SCS framework, we have developed a novel stochastic algorithm for two-stage stochastic programming. The convergence and convergence rates of the algorithm have been rigorously established, with preliminary computational results support the theoretical analysis.

math.OC↗

arXiv2025

FullDiT: Multi-Task Video Generative Foundation Model with Full Attention

Current video generative foundation models primarily focus on text-to-video tasks, providing limited control for fine-grained video content creation. Although adapter-based approaches (e.g., ControlNet) enable additional controls with minimal fine-tuning, they encounter challenges when integrating multiple conditions, including: branch conflicts between independently trained adapters, parameter redundancy leading to increased computational cost, and suboptimal performance compared to full fine-tuning. To address these challenges, we introduce FullDiT, a unified foundation model for video generation that seamlessly integrates multiple conditions via unified full-attention mechanisms. By fusing multi-task conditions into a unified sequence representation and leveraging the long-context learning ability of full self-attention to capture condition dynamics, FullDiT reduces parameter overhead, avoids conditions conflict, and shows scalability and emergent ability. We further introduce FullBench for multi-task video generation evaluation. Experiments demonstrate that FullDiT achieves state-of-the-art results, highlighting the efficacy of full-attention in complex multi-task video generation.

cs.CV↗

arXiv2025

DiffMoE: Dynamic Token Selection for Scalable Diffusion Transformers

Diffusion models have demonstrated remarkable success in various image generation tasks, but their performance is often limited by the uniform processing of inputs across varying conditions and noise levels. To address this limitation, we propose a novel approach that leverages the inherent heterogeneity of the diffusion process. Our method, DiffMoE, introduces a batch-level global token pool that enables experts to access global token distributions during training, promoting specialized expert behavior. To unleash the full potential of the diffusion process, DiffMoE incorporates a capacity predictor that dynamically allocates computational resources based on noise levels and sample complexity. Through comprehensive evaluation, DiffMoE achieves state-of-the-art performance among diffusion models on ImageNet benchmark, substantially outperforming both dense architectures with 3x activated parameters and existing MoE approaches while maintaining 1x activated parameters. The effectiveness of our approach extends beyond class-conditional generation to more challenging tasks such as text-to-image generation, demonstrating its broad applicability across different diffusion model applications. Project Page: https://shiml20.github.io/DiffMoE/

cs.CV↗

arXiv2025

Solving Token Gradient Conflict in Mixture-of-Experts for Large Vision-Language Model

The Mixture-of-Experts (MoE) has gained increasing attention in studying Large Vision-Language Models (LVLMs). It uses a sparse model to replace the dense model, achieving comparable performance while activating fewer parameters during inference, thus significantly reducing the inference cost. Existing MoE methods in LVLM encourage different experts to specialize in different tokens, and they usually employ a router to predict the routing of each token. However, the router is not optimized concerning distinct parameter optimization directions generated from tokens within an expert. This may lead to severe interference between tokens within an expert. To address this problem, we propose to use the token-level gradient analysis to Solving Token Gradient Conflict (STGC) in this paper. Specifically, we first use token-level gradients to identify conflicting tokens in experts. After that, we add a regularization loss tailored to encourage conflicting tokens routing from their current experts to other experts, for reducing interference between tokens within an expert. Our method can serve as a plug-in for diverse LVLM methods, and extensive experimental results demonstrate its effectiveness. The code will be publicly available at https://github.com/longrongyang/STGC.

cs.CV↗