arXiv 公告日 · 周一
回补：候选来自 arXiv /list/<分类>/recent 列表页，摘要来自 arXiv API。当天 cs.OS 没有新论文。
LiHA: Worse is Better for Lightweight Hash-based Source Authentication
Negin Dehghanchaleshtori, Homit Dalia 等 6 人
互联网至今缺少数据面上的 IP 源地址认证，已有方案依赖路由可信度、拓扑假设、每流状态或大范围部署。LiHA 用最长前缀匹配把密钥绑定到源前缀，每个包只附一个 MAC，数据面不需要握手、每流状态或控制面参与，发送方和验证方都只做一次前缀查找和一次定长哈希。作者分别实现了用户态库、eBPF 程序和 FPGA 三个版本，并明确说明它主要防范离路径伪造，不能消除所有拒绝服务威胁。
参考价值同一机制的用户态库、eBPF 和 FPGA 实现放在一起比较：用户态库开销很高，eBPF 版用软件哈希也有不错的单线程性能，这组对比可作为 eBPF 执行路径性能讨论的参考。
这个主题本期没有论文。
LLM Fine-Tuning Workloads: Characterization, Challenges, and Opportunities for Serverless AI
cs.DC2610.40983Migrating a Legacy Scientific Application with an AI Coding Agent: NICAM from Fortran to Kokkos
cs.DC2610.41313MemX: Memory-Aware Automatic Parallelism for Cross-Vendor Heterogeneous GPU Training
cs.DC2610.41525Analysis of Triggered Packet Streams: Stochastic Bounds on System Time under General Triggering Delays
math.PR2610.40865Falcons: Pushing the Speed-Ratio Limits of Lossy Compression for Parallel Scientific Applications
cs.DC2610.41209AuditFL: A Contract-Backed Clearing Layer for Open Federated Learning
cs.DC2610.41278Supervised and LLM-Augmented Cardinality Estimation for Similarity Queries on Knowledge Graphs
cs.DB2610.41669Understanding the Energy Cost of Quantized Agents on Shared GPUs
cs.PF2610.41672A Deterministic Commitment Rule over Unstructured DAGs for Optimistic Replication
cs.DC2610.41837CLIC: Closed-Loop Intent Consistency for Securing LLM-Driven Control in Open RAN
cs.CR2610.40744Merging and Evolution of Data Bases: A Many-Logic Approach
cs.LO2610.40812How Global Is Anycast?: Evaluating Performance Gaps
cs.NI2610.40833HippocamFLEX: FPGA Prototyping of Hippocampus-Inspired Architecture for Context-Aware Sequence Prediction
cs.AR2610.40932Adaptive Dynamic Security for Operational Technology Against AI-Powered Attacks
cs.CR2610.41007An Incentive-Aligned Trust-Centric Coordination Layer for Heterogeneous Blockchain Interoperability
cs.CR2610.41032ZenKO: Guided and Iterative Quantum-Classical Method for Molecular Electronic Structure using Sample-Based Quantum Diagonalization
cs.ET2610.41146ARK: Avoiding Routing Collisions for KV Cache Transfer in Disaggregated LLM Inference
cs.NI2610.41160Iris on Indirection Theory: Lightweight Higher-order Ghost State
cs.PL2610.41202Resource Locks and Fast Unlock for Shared Balances in Certificate-Based Payment Systems
cs.DC2610.41207Fidelity Synopses: Pre-Execution Plan Selection for Approximate Vector Search
cs.DB2610.41407Beamfocusing Beyond the Near Field: Frequency Diverse Arrays for OFDM-ISAC Systems
cs.NI2610.41415Vehicle clustering and resource scheduling scheme based on collaborative perception of wireless resources
cs.NI2610.41423RecIS: Unifying Sparse and Dense Computation for Efficient Recommendation Training
cs.DC2610.41457Model Sharing Bayesian Optimization for Mechanical Wi-Fi Antenna Control in Multi-Station Environments
cs.NI2610.41469HiState: Hierarchical Prefix Cache Management for Hybrid LLM Serving in Agentic Workloads
cs.DC2610.41508DeepWeave: Weaving Neural and Symbolic Evidence for ML-Based Network Intrusion Detection System
cs.CR2610.41541ORION: An Outcome-Driven LLM Advisor for RL-Based Multi-Domain 6G Slice Placement
cs.LG2610.41563Provisioning Networks for AI Supercomputers: A Trace-Driven Study of Performance Sensitivity at Unprecedented Scale
cs.NI2610.41700Ordered Chaos: Towards Adaptive Multipath Load Balancing with Precise Fast Loss Detection
cs.NI2610.41707CypherGuard: Database-Grounded Correction for Reliable Text-to-Cypher
cs.DB2610.41755Clinical-Protocol-Aware Cyber Resilience for Connected Clinical Technologies Under AI-Enabled Adversarial Conditions
cs.AI2610.41784Where KV-Cache Error Lives in Looped Language Models: A Diagnostic Study of Ouro and Huginn
cs.DC2610.418256G-Enabled Cooperative Connected and Automated Mobility: Are We There Yet?
cs.NI2610.41836Layered Reproducibility for High-Performance Computing Applications: The Feel++ and Ktirio Urban Buildings Case Study
cs.DC2610.41861Decentralized Dynamic Stable Matching
cs.DC2610.41912A Multimodal Type 2 Diabetes Registry on DHIS2: A Design and Feasibility Study with the AI-READI Dataset
cs.CY2610.41918Matching Workloads to Blackwell: Power-Resolved Benchmarking of NVIDIA B200 and B300 across HPC and AI Workloads
cs.DC2610.41919RAMA: Retrieval-Augmented Multi-Agent Routing with Episodic Memory for LEO Satellite Networks
cs.NI2610.42068