Community GPU(Graphics Processing Unit) platforms are emerging as a cost-effective and democratized alternative to centralized GPU clusters for AI(Artificial Intelligence) workloads, aggregating idle consumer GPUs from globally distributed and heterogeneous environments. However, their extreme hardware/software diversity, volatile availability, and variable network conditions render traditional schedulers ineffective, leading to suboptimal task completion. In this work, we present REACH (Reinforcement Learning for Efficient Allocation in Community and Heterogeneous Networks), a Transformer-based reinforcement learning framework that redefines task scheduling as a sequence scoring problem to balance performance, reliability, cost, and network efficiency. By modeling both global GPU states and task requirements, REACH learns to adaptively co-locate computation with data, prioritize critical jobs, and mitigate the impact of unreliable resources. Extensive simulation results show that REACH improves task completion rates by up to 17%, more than doubles the success rate for high-priority tasks, and reduces bandwidth penalties by over 80% compared to state-of-the-art baselines. Stress tests further demonstrate its robustness to GPU churn and network congestion, while scalability experiments confirm its effectiveness in large-scale, high-contention scenarios.

Training Mixture-of-Experts (MoE) models introduces sparse and highly imbalanced all-to-all communication that dominates iteration time. Conventional load-balancing methods fail to exploit the deterministic topology of Rail architectures, leaving multi-NIC bandwidth underutilized. We present RailS, a distributed load-balancing framework that minimizes all-to-all completion time in MoE training. RailS leverages the Rail topology’s symmetry to prove that uniform sending ensures uniform receiving, transforming global coordination into local scheduling. Each node independently executes a Longest Processing Time First (LPT) spraying scheduler to proactively balance traffic using local information. RailS activates N parallel rails for fine-grained, topology-aware multipath transmission. Across synthetic and real-world MoE workloads, RailS improves bus bandwidth by 20%–78% and reduces completion time by 17%–78%. For Mixtral workloads, it shortens iteration time by 18%–40% and achieves near-optimal load balance, fully exploiting architectural parallelism in distributed training.

Variational autoencoders (VAEs) have been deployed in many privacy-sensitive domains, and their vulnerability to membership inference attacks (MIAs) poses giant privacy risks. While some existing privacy protection methods like differential privacy often compromise generative models’ utility, we present GuidedLatent, a novel mechanism that enhances membership privacy and preserves their generative performance. GuidedLatent allows the model to adjust latent representations dynamically based on distribution similarities, coupled with a two-phase training strategy that gradually incorporates privacy constraints. We also establish bounds on the privacy-utility trade-off theoretically and prove our mechanism reduces the performance of membership inference attacks compared to other baseline approaches. Extensive experiments demonstrate that our method maintains high-quality generation capabilities while minimizing degradation in quality metrics. Our method performs effectively across various VAE variants and architectures, providing a practical solution for privacy-preserving generative models.1

Graph-level clustering is a fundamental task of data mining, aiming at dividing unlabeled graphs into distinct groups. However, existing deep methods that are limited by pooling have difficulty extracting diverse and complex graph structure features, while traditional graph kernel methods rely on exhaustive substructure search, unable to adaptively handle multi-relational data. This limitation hampers producing robust and representative graph-level embeddings. To address this issue, we propose a novel Multi-Relation Graph-Kernel Strengthen Network for Graph-Level Clustering (MGSN), which integrates Multi-Relation Modeling (MRM) with graph kernel to fully employ their respective advantages. Specifically, MGSN constructs multi-relation graphs to capture diverse semantic relationships between nodes and graphs, which employ graph kernel methods to extract graph affinity, enriching the representation space. Moreover, a Relation-aware Embedding Strengthening Strategy (RESS) is designed, which adaptively aligns multi-relation information across views while strengthening graph-level features through a progressive fusion process. Extensive experiments on multiple benchmark datasets demonstrate the superiority of MGSN over state-of-the-art methods. The results highlight its ability to leverage multi-relation structures and graph kernel features, establishing a new paradigm for robust graph-level clustering.

In recent years, Vision Transformers (ViT) have achieved significant success in various complex visual tasks, but they also come with substantial computational costs and memory overheads. To address this issue, lightweight Vision Transformers have become an important research direction. Current research on lightweight ViTs mainly focuses on combining CNNs and Transformers, leveraging the advantages of CNNs in local feature extraction while utilizing Transformers’ ability to model global context. However, existing lightweight models often suffer from an imbalance in processing low-frequency global information and high-frequency local information. While sparse attention mechanisms effectively capture global context and reduce computational load, they typically adopt relatively simple strategies for handling high-frequency local information, failing to fully exploit the details of local features. To address this issue, we introduce a new lightweight Vision Transformer model, A Lightweight Vision Transformer for Balancing Global and Local Information (GLFormer). GLFormer combines dynamic weight adjustment with context-aware mechanisms to effectively aggregate high-frequency local information, optimizing the balance between global and local information. Additionally, we introduce a Depth Perception Feed-Forward Network (DPFFN), which further enhances feature fusion and detail refinement, thus enhancing the model’s performance and its capacity to generalize. Based on GLFormer and DPFFN, we design a novel visual backbone network—GLNet. Extensive experimental results show that GLNet consistently demonstrates excellent performance across various tasks, while maintaining a relatively low computational cost.

Large language models (LLMs) have achieved significant success in complex tasks across various domains, but these achievements come with high computational costs and long inference delays. Pruning, as an effective optimization technique, simplifies model structures by removing redundant components, thereby improving model generalization and operational efficiency. Although existing pruning retraining-free algorithms perform excellently in pruning time, these algorithms often focus on local optimal solutions in encoder-based language models, lacking comprehensive exploration of global optimal solutions, which may affect the overall model performance. To address this issue, we propose a novel retraining-free structured pruning algorithm, named RL-Pruner. The algorithm consists of two main stages: the Mask Rearrangement Based on Asynchronous Advantage Actor-Critic (MA3C) stage and the BiConjugate Gradient Solver for Mask Tuning (BGMT) stage. It aims to explore the intra-layer interactions of mask variables and efficiently find the global optimal solution without requiring retraining. We evaluate this method using BERTBASEand DistilBERT models on the GLUE and SQuAD benchmark tests. Experimental results show that RL-Pruner significantly improves accuracy on the SQuAD1.1benchmark. Under a 60% FLOPs constraint, compared with existing pruning retraining-free algorithms, the F1 score increases by 4.25%.