Institution profile

Pengcheng Laboratory

Academic institutionasia · cn
Official website
Research library1,089linked papers
Opportunities0open roles
Selected work

Representative Papers

FedBCGD: Communication-Efficient Accelerated Block Coordinate Gradient Descent for Federated Learning

Oct 28, 2024ACM Multimedia

This work addresses the high communication overhead of large-scale models, such as Vision Transformers, in federated learning by proposing Federated Block Coordinate Gradient Descent (FedBCGD) and its accelerated variant, FedBCGD+. The method introduces, for the first time in federated learning, a block-wise parameter communication mechanism that uploads only a subset of parameter blocks per round, combined with stochastic variance reduction and client drift control strategies. Theoretical analysis shows that the communication complexity is reduced by a factor of 1/N compared to existing methods, where N denotes the number of blocks. Experimental results demonstrate that the proposed algorithms achieve faster convergence and higher communication efficiency than current state-of-the-art approaches.

37 citations2 influentialRead paper

AtomThink: A Slow Thinking Framework for Multimodal Mathematical Reasoning

Nov 18, 2024arXiv.org

Multimodal large language models (MLLMs) exhibit limited capability in solving complex mathematical reasoning problems due to insufficient granular, stepwise reasoning over multimodal inputs. Method: This paper introduces the “slow-thinking” paradigm, integrating long-chain, atomic-level reasoning into MLLMs via AtomThink—a novel atomic thinking framework comprising (i) an automatic Chain-of-Thought (CoT) annotation engine, (ii) atomic-step fine-tuning, and (iii) a policy-based search method guided by a four-category strategy reward model (PRM). The approach unifies vision–math joint fine-tuning, reinforcement learning–driven search, and interpretable CoT generation. Contribution/Results: We release AtomMATH, a large-scale multimodal mathematical dataset, and propose fine-grained atomic capability evaluation metrics. On MathVista and MathVerse benchmarks, our method achieves relative accuracy improvements of ~50% and ~120%, respectively, significantly enhancing MLLMs’ hierarchical, adaptive reasoning on complex mathematical problems.

18 citations3 influentialRead paper

TrimCaching: Parameter-sharing Edge Caching for AI Model Downloading

Apr 22, 2024arXiv.org

Existing edge caching mechanisms for AI model delivery in 5G/6G networks overlook parameter-block reuse—e.g., shared knowledge units across CNNs or LLMs—leading to low storage efficiency and limited cache hit rates under stringent latency constraints. Method: We propose a parameter-sharing-aware edge model caching framework that, for the first time, formulates parameter-block reuse as a submodular optimization problem. We design a polynomial-time algorithm with theoretical approximation guarantees and provide a general greedy solution. The framework jointly optimizes storage efficiency and service latency in multi-edge wireless networks. Results: Simulation results demonstrate that our approach significantly improves cache hit rates over conventional content-based caching, validating the effectiveness and practicality of parameter-level sharing for edge AI deployment.

12 citationsRead paper

Consistency of Local and Global Flatness for Federated Learning

Oct 27, 2025ACM Multimedia

This work addresses the challenge in federated learning where multiple local updates under data heterogeneity often drive the global model toward sharp minima, degrading generalization. Existing sharpness-aware methods struggle to align local and global flatness. To this end, we propose FedNSAM, an algorithm that leverages global Nesterov momentum to guide local updates, constructing an estimated direction of global perturbation and performing extrapolation to harmonize local and global flatness. We introduce a novel “flatness distance” metric to quantify the inconsistency between local and global landscapes and establish a tighter convergence bound than FedSAM in our theoretical analysis. Empirical results demonstrate that FedNSAM significantly enhances both generalization performance and training efficiency across CNN and Transformer architectures, particularly in highly heterogeneous settings.

11 citationsRead paper

CL-MVSNet: Unsupervised Multi-view Stereo with Dual-level Contrastive Learning

Oct 01, 2023IEEE International Conference on Computer Vision

To address incomplete and brittle depth estimation in unsupervised multi-view stereo (MVS) caused by low-texture regions and view-dependent effects (e.g., reflections), this paper proposes a two-level contrastive learning framework: image-level and scene-level contrastive branches are jointly optimized to enhance contextual awareness and feature representation robustness. Additionally, we introduce an L₀.₅ photometric consistency loss that selectively emphasizes high-confidence correspondences, mitigating the over-penalization of low-gradient regions inherent in conventional L₁/L₂ losses. The method is fully unsupervised—requiring no ground-truth depth annotations. Evaluated on DTU and Tanks & Temples benchmarks, it achieves state-of-the-art performance among unsupervised MVS approaches and surpasses leading supervised methods without fine-tuning. Our core contributions are the first-ever dual-granularity contrastive mechanism for MVS and an L₀.₅ norm-driven photometric constraint, jointly advancing robustness and accuracy in texture-deficient and view-dependent scenarios.

7 citations2 influentialRead paper
Recent publications

Latest Papers