Speculate with Memory: Lossless Acceleration for LLM Agents

arXiv cs.LG Papers

Summary

This paper introduces memory-augmented speculative execution for LLM agents, using three online memory systems to improve prediction accuracy by 19-39% on action prediction and up to 2.5x on observation prediction, all while being lossless with zero added wall-clock cost.

arXiv:2607.12236v1 Announce Type: new Abstract: Speculative execution accelerates LLM agents by using a smaller, cheaper model to predict and pre-launch the next step while the environment is idle. However, existing speculators are stateless and discard all information between tasks, preventing prediction quality from improving with experience. We equip the speculator with three online memory systems that learn from past agent trajectories: a contrastive transition table tracking action-sequence statistics, an episodic memory retrieving contextually similar segments, and a confusion tracker suppressing recurring errors. We evaluate this approach on six benchmarks spanning three speculation types: action prediction, observation prediction, and chained prediction. Memory-augmented speculation yields a 19--39\% relative accuracy improvement on action prediction and up to a $2.5\times$ increase on observation prediction tasks with repetitive action spaces. These gains grow continuously as memory accumulates and generalize across speculator models of varying cost. All speculation is lossless because it runs during idle time at zero added wall-clock cost, and the actor's trajectory is identical to non-speculative execution.
Original Article
View Cached Full Text

Cached at: 07/15/26, 04:18 AM

# Speculate with Memory: Lossless Acceleration for LLM Agents
Source: [https://arxiv.org/abs/2607.12236](https://arxiv.org/abs/2607.12236)
[View PDF](https://arxiv.org/pdf/2607.12236)

> Abstract:Speculative execution accelerates LLM agents by using a smaller, cheaper model to predict and pre\-launch the next step while the environment is idle\. However, existing speculators are stateless and discard all information between tasks, preventing prediction quality from improving with experience\. We equip the speculator with three online memory systems that learn from past agent trajectories: a contrastive transition table tracking action\-sequence statistics, an episodic memory retrieving contextually similar segments, and a confusion tracker suppressing recurring errors\. We evaluate this approach on six benchmarks spanning three speculation types: action prediction, observation prediction, and chained prediction\. Memory\-augmented speculation yields a 19\-\-39\\% relative accuracy improvement on action prediction and up to a $2\.5\\times$ increase on observation prediction tasks with repetitive action spaces\. These gains grow continuously as memory accumulates and generalize across speculator models of varying cost\. All speculation is lossless because it runs during idle time at zero added wall\-clock cost, and the actor's trajectory is identical to non\-speculative execution\.

## Submission history

From: Yu Li \[[view email](https://arxiv.org/show-email/b7ad814a/2607.12236)\] **\[v1\]**Tue, 14 Jul 2026 00:36:31 UTC \(423 KB\)

Similar Articles

SimpleMem: Efficient Lifelong Memory for LLM Agents

Papers with Code Trending

Introduces SimpleMem, an efficient memory framework for LLM agents that uses semantic lossless compression to improve accuracy and reduce token consumption, achieving 26.4% F1 improvement and up to 30x reduction in inference-time token usage.

Selective Memory Retention for Long-Horizon LLM Agents

arXiv cs.AI

This paper presents TraceRetain, a lightweight framework for bounded external memory in frozen LLM agents, demonstrating that selective retention differentiates from cache heuristics primarily when memory streams contain noise, offering task-success and efficiency benefits.