Receiver-conditioned communication boosts agent coordination efficiency
Receiver-Conditioned Latent Communication gives 94% CacheBack
Artificial Intelligence
Summary
Coordinating multiple AI agents to solve tasks requires passing information between them, often creating large data exchanges that strain computer memory and slow processing. The authors found that agents only need to send the parts of their information relevant to the receiving agent's task. They designed CacheBack, a method that filters and shrinks these data messages based on what the receiver needs, speeding up task completion and improving accuracy. Their tests showed big improvements in how much data is sent, task success rate, and processing speed.
What this means in practice
- •For distributed ai developers: Reduce memory and processing demands in multi-agent AI systems by tailoring communication to receiver needs, improving speed and accuracy of coordination.
- •For cloud service architects: Optimize infrastructure for collaborative AI workloads by limiting exchanged data size, lowering resource use and cost without model retraining.
Authors
Maximillian Rossi, Prajwal Raghunath, Haoqing Xuan, Yusen Zhang, Eugene Wu
Abstract
Multi-agent systems distribute large contexts across agents that communicate to solve a task. Text messages are compact but require decoding and may omit evidence the receiving agent needs. Recent latent communication instead transfers KV caches. This avoids text generation and can improve accuracy and latency. However, a full KV cache grows linearly with both the context an individual agent processes, and the number of agents that coordinate together. This raises memory and context costs, often far exceeding available GPU resources and context window sizes. Our key observation is that agents need only send what the receiving agent requires for its local task -- which we call receiver-conditioned communication. The receiver agent passes the sender a small description of its information needs, which serves to filter and compress the sender agent's KV cache. CacheBack is a simple, robust, training-free instance of receiver conditioning based on the sender's attention weights. On FanOutQA, CacheBack with Qwen 3 removes 75% of the state the agent would otherwise receive, improving accuracy by 14.7 percentage points and reducing median task-completion latency by 3.2x relative to text communication. We show comparable improvements across model families that span dense Transformers, Mamba-attention hybrids, and sliding-window attention.