Tag
This paper proposes Polar, a multimodal memory-augmented framework for personalizing embodied MLLM agents over long-term user interactions, using a knowledge graph and episodic memory to ground user-intended instances from accumulated context.
Memory-R2 introduces LoGo-GRPO, a training framework that combines local and global group-relative optimization to provide fairer credit assignment for long-horizon memory-augmented LLM agents, improving accuracy and inference latency across backbones.
This paper proposes a memory-augmented reinforcement learning framework for CAD generation agents that integrates geometric kernel toolchains, dual-track memory, and dynamic utility retrieval to handle complex CAD models with long operation sequences and geometric constraints, achieving improved success rate and geometric consistency.
This paper proposes ARS, a memory-augmented agentic recommender system that treats recommendation as a partially observable problem with a hierarchical belief-state memory structure. It achieves state-of-the-art performance on four benchmarks with significant improvements over baselines.
SafeHarbor is a novel framework for LLM agent safety that uses hierarchical memory and self-evolution to balance safety and utility, achieving state-of-the-art performance on benign and malicious tasks.