A Token/KV-Cache Communication Media Selection and Resource Allocation Strategy for Multi-Agent Collaboration
eess.SP, cs.AI, cs.IT, math.IT
Submitted: 2026-05-25
Updated: 2026-05-25
Journal ref: SCIENCE CHINA Information Sciences, 2026
DOI: 10.1007/s11432-026-5057-2
License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Terminology
Sources
- When Intelligence Overloads Infrastructure: A Forecast Model for AI-Driven Bottlenecks
- Enabling Agents to Communicate Entirely in Latent Space
- Thought Communication in Multiagent Collaboration
- Cache-to-Cache: Direct Semantic Communication Between Large Language Models
- Latent Collaboration in Multi-Agent Systems
- Q-KVComm: Efficient Multi-Agent Communication Via Adaptive KV Cache Compression
- Agent Primitives: Reusable Latent Building Blocks for Multi-Agent Systems
- Efficient LLM Inference over Heterogeneous Edge Networks with Speculative Decoding
- Distributed On-Device LLM Inference With Over-the-Air Computation
- Beyond Self-Talk: A Communication-Centric Survey of LLM-Based Multi-Agent Systems
- A Survey of Large Language Models
- DualPath: Breaking the Storage Bandwidth Bottleneck in Agentic LLM Inference
- DeepSeek-V3 Technical Report
- LLaMA: Open and Efficient Foundation Language Models
Related papers
- What If They Took the Shot? A Hierarchical Bayesian Framework for Counterfactual Expected Goals
- Runtime Assurance Under Measurement Attack: Necessary and Sufficient Observability Conditions for Learned Control in Radio Access Networks
- Physics-Constrained Deep Learning Model for Contactless Blood Pressure Monitoring from Triaxial Bodyseismography
- SpellerSSL: Self-Supervised Learning with P300 Aggregation for Speller BCIs
- Uncertainty Quantification in Machine Learning for Biosignal Applications -- A Review
- Continuous Orthogonal Mode Decomposition: Haptic Signal Prediction in Tactile Internet