惯性聚合 高效追踪和阅读你感兴趣的博客、新闻、科技资讯
阅读原文 在惯性聚合中打开

推荐订阅源

奇客Solidot–传递最新科技情报
奇客Solidot–传递最新科技情报
爱范儿
爱范儿
博客园 - 三生石上(FineUI控件)
Vercel News
Vercel News
M
MIT News - Artificial intelligence
L
LangChain Blog
大猫的无限游戏
大猫的无限游戏
钛媒体:引领未来商业与生活新知
钛媒体:引领未来商业与生活新知
Microsoft Azure Blog
Microsoft Azure Blog
J
Java Code Geeks
Recent Announcements
Recent Announcements
Stack Overflow Blog
Stack Overflow Blog
人人都是产品经理
人人都是产品经理
IT之家
IT之家
F
Fortinet All Blogs
博客园 - 聂微东
U
Unit 42
Martin Fowler
Martin Fowler
腾讯CDC
博客园_首页
OSCHINA 社区最新新闻
OSCHINA 社区最新新闻
量子位
阮一峰的网络日志
阮一峰的网络日志
博客园 - Franky

cs.DC updates on arXiv.org

DUAL-BLADE: Dual-Path NVMe-Direct KV-Cache Offloading for Edge LLM Inference Progressive Semantic Communication for Efficient Edge-Cloud Vision-Language Models Efficient, VRAM-Constrained xLM Inference on Clients Folding Tensor and Sequence Parallelism for Memory-Efficient Transformer Training & Inference DORA: A Scalable Asynchronous Reinforcement Learning System for Language Model Training AMMA: A Multi-Chiplet Memory-Centric Architecture for Low-Latency 1M Context Attention Serving RaMP: Runtime-Aware Megakernel Polymorphism for Mixture-of-Experts Spark Policy Toolkit: Semantic Contracts and Scalable Execution for Policy Learning in Spark Internet of Everything in the 6G Era: Paradigms, Enablers, Potentials and Future Directions PolyKV: A Shared Asymmetrically-Compressed KV Cache Pool for Multi-Agent LLM Inference A Survey on Split Learning for LLM Fine-Tuning: Models, Systems, and Privacy Optimizations ITAS: A Multi-Agent Architecture for LLM-Based Intelligent Tutoring Latency and Cost of Multi-Agent Intelligent Tutoring at Scale TACO: Efficient Communication Compression of Intermediate Tensors for Scalable Tensor-Parallel LLM Training FreeScale: Distributed Training for Sequence Recommendation Models with Minimal Scaling Cost CommFuse: Hiding Tail Latency via Communication Decomposition and Fusion for Distributed LLM Training A Taxonomy and Resolution Strategy for Client-Level Disagreements in Federated Learning Usable Agent Discovery for Decentralized AI Systems Cloud to Edge: Benchmarking LLM Inference On Hardware-Accelerated Single-Board Computers Data-Free Contribution Estimation in Federated Learning using Gradient von Neumann Entropy Shard the Gradient, Scale the Model: Serverless Federated Aggregation via Gradient Partitioning Promoting Simple Agents: Ensemble Methods for Event-Log Prediction GraphLeap: Decoupling Graph Construction and Convolution for Vision GNN Acceleration on FPGA AGNT2: Autonomous Agent Economies on Interaction-Optimized Layer 2 Infrastructure FedSIR: Spectral Client Identification and Relabeling for Federated Learning with Noisy Labels Stream-CQSA: Avoiding Out-of-Memory in Attention Computation via Flexible Workload Scheduling A Delta-Aware Orchestration Framework for Scalable Multi-Agent Edge Computing Federated Learning over Blockchain-Enabled Cloud Infrastructure Optimal Routing for Federated Learning over Dynamic Satellite Networks: Tractable or Not? Sherpa.ai Privacy-Preserving Multi-Party Entity Alignment without Intersection Disclosure for Noisy Identifiers
The Blocklace: A Byzantine-repelling and Universal Confli...
Paulo Sérgio Almeida, Ehud Shapiro · 2024-02-13 · via cs.DC updates on arXiv.org

Conflict-free Replicated Data Types (CRDTs) are designed for replica convergence without global coordination or consensus. Recent work has achieved the same in a Byzantine environment, through DAG-like structures based on cryptographic hashes of content. The blocklace is a partially-ordered generalization of the blockchain in which each block has any finite number of signed hash pointers to preceding blocks. We show that the blocklace datatype, with the sole operation of adding a single block, is a CRDT: it is both a pure operation-based CRDT, with self-tagging; and a delta-state CRDT, under a slight generalization of the delta framework. Allowing arbitrary values as payload, the blocklace can also be seen as a universal Byzantine fault-tolerant implementation for arbitrary CRDTs, under the operation-based approach. Current approaches only care about CRDT convergence, being equivocation-tolerant (they do not detect or prevent equivocations), allowing a Byzantine node to cause an arbitrary amount of harm by polluting the CRDT state with an unbounded number of equivocations. We show that the blocklace can be used not only in an equivocation-tolerant way, but also so as to detect and eventually exclude Byzantine nodes, including equivocators, even under the presence of undetectable colluders. The blocklace CRDT protocol ensures that a Byzantine node may harm only a finite prefix of the computation.