Cognitive Extensions for Dual-Process Language Agents: Memory and Self-Reflection in Interactive Environments
Researchers extend the SwiftSage dual-process agent with adaptive memory and self-reflection modules, improving scores in interactive environments.
The work adds an Adaptive Memory Module (AMM) for salience-gated episodic storage and trigger-driven retrieval, and a Self-Reflection Module (SRM) for bounded execution-time validation and corrective intervention, to the SwiftSage agent. Controlled ablations on ScienceWorld across four configurations show the full system achieves the best mean final score (64.62), success rate (43.17%), and successful-step efficiency (19.33 steps). SRM is the strongest standalone contributor, suggesting execution-time control is the dominant bottleneck while episodic memory helps once the runtime loop is stable.
- AMM provides salience-gated episodic memory with trigger-driven retrieval
- SRM performs bounded execution-time validation and corrective intervention
- Full system reaches 64.62 mean final score and 43.17% success on ScienceWorld
- Self-reflection beats episodic memory as a standalone improvement
Full article141 words · extracted from arxiv.org · click to collapse
Language agents remain brittle in interactive environments, where success requires long-horizon state tracking, valid action execution, and recovery from failed steps. We extend SwiftSage, a dual-process agent that combines a fast action proposer with a slower planner, using two modular cognitive extensions: an Adaptive Memory Module (AMM) for salience-gated episodic storage and trigger-driven retrieval, and a Self-Reflection Module (SRM) for bounded execution-time validation and corrective intervention. Both modules are implemented as feature-flagged extensions over the same execution substrate, enabling controlled ablations on ScienceWorld. Across four configurations---baseline, baseline+AMM, baseline+SRM, and the full system---the full system achieves the best mean final score (64.62), success rate (43.17%), and successful-step efficiency (19.33 steps), while SRM is the strongest standalone contributor. The results suggest that execution-time control is the dominant bottleneck in this setting, while episodic memory becomes most useful once the runtime loop is stabilized.
Text extracted automatically; images, tables and formatting may be missing. Original: https://arxiv.org/abs/2609.19128