Top Headlines

Feeds

Microsoft Proposes SageServe to Cut GPU Costs for LLM Inference

Cached

Researchers Highlight Gaps in CPU Isolation Against Microarchitectural Leaks

Cached

DroidSpeak Boosts Multi‑LLM Inference Efficiency by Up to Threefold

Cached

HarvestContainers Enables Up to 75% CPU Core Sharing for Latency‑Sensitive Kubernetes Pods

Cached

Researchers Test Brain Foundation Models for Continuous Cognitive‑Load Monitoring

Cached

Concord Tool Learns Network Configuration Contracts to Detect Misconfigurations

Cached

EgoBrain dataset launches multimodal human‑action research platform

Cached

New Methods Reduce Low‑Probability Token Dominance in RL Training of LLMs

Cached

MetaMuse Framework Boosts Cloud Algorithm Performance

Cached

VidGuard‑R1 Sets New Benchmark in AI‑Generated Video Detection

Cached

Adversarial RL Framework Boosts LLM Unit Test Generation

Cached

Generative UI Workshop Set for CHI 2026 Explores AI‑Driven Interface Design

Cached

Engaging Disability Communities in AI Image Representation

Cached

VeriStruct Extends AI‑Assisted Formal Verification to Rust Data‑Structure Modules

Cached

Niyama Boosts LLM Inference Capacity by 32% with Fine‑Grained QoS

Cached

MSCCL++ Proposes New GPU Communication Abstractions for AI Inference

Cached

SUTRADHARA Identifies Latency Bottlenecks in Tool‑Based LLM Agents

Published Cached

OrbitalBrain Demonstrates Up to 12.4× Faster In‑Space ML Training

Published Cached

Researchers Propose SRv6 for Fully Controllable AI Backend Traffic

Published Cached

Predictive Inverse Dynamics Models Show Sample‑Efficiency Gains Over Behavior Cloning

Published Cached

UniRG Framework Sets New State‑of‑the‑Art in Chest X‑Ray Report Generation

Published Cached

Benchmarking Affordance Generalization with BusyBox

Published Cached

New Algorithms Produce Tight Composable Coresets for Constrained Determinant Maximization

Published Cached

SALAD‑VAE Achieves State‑of‑the‑Art Semantic Audio Compression

Published Cached

Sublinear‑Time Approximation for Metric Steiner Forest and MIS Size Estimation

Published Cached

Terabyte‑Scale Analytics Achieve 60× Speedup on GPU Clusters

Published Cached

New Retrieval Method Boosts Long‑Document QA Performance

Published Cached

Real-Time Generative Speech Restoration Achieves 20 ms Latency

Published Cached

Sci‑Phi: First Audio LLM for Full Spatial‑Scene Description

Published Cached

Predictive Models for Kidney Offer Acceptance Face Data and Decision Gaps

Published Cached