Memory Tax: 에이전트 메모리의 순증가 효과 분석
Memory Tax: Dissecting the Net Value of Agent Memory
- 매체
- Hugging Face
- 게시
- 중요도
- 3/5
- 화제 지수
- 16
한국어 요약
장시간 에이전트에서 메모리 시스템의 실제 가치를 평가한 논문이다. 많은 전용 메모리 시스템이 단순 RAG보다 성능이 떨어지는 경우가 많으며, 메모리 이득과 손실을 정량화한 '메모리 세금' 개념을 제시했다.
같은 소식, 다른 매체 · 8곳
- Retrieval as Traversal: Value-Indexed Memory for Stateful Long-Context Agents
- HABIT-Bench: Evaluating Habit Induction from Longitudinal Weak Evidence in Agent Memory
- Preserving Utility in Continual LLM Unlearning via Orthogonal Projections
- Glance: The Generator Is Its Own Reranker
- Wizard: Manipulate Any World in Real Time via Agentic Control
- RFW-A: Agentic Procedural Generation of 400K Articulated Assets over 500 Categories
- WorldProbe: When Video World Models Fall Apart under Long-Horizon Spatial Exploration
- Distilling Large-Scale Machine Learning Interatomic Potentials to Predict Long-Timescale Electrolyte Transport Properties
원문 발췌 · English
Memory is increasingly used in long-horizon agents, yet the net value of building memory over direct retrieval from raw history remains largely unclear. Through a large-scale evaluation across multiple benchmarks and model backbones, we find many dedicated memory systems underperform naive RAG in many cases, indicating a negative net value. Taking RAG as a reference, we dissect the net value of memory into the gains and losses over direct retrieval, termed memory benefit and memory tax. These gains and losses arise because memory process changes how the raw information is organized and accessed. We view such changes as the double-edged deviations from the raw history. They are originally designed and introduced to benefit the task, for example by reorganizing facts or distilling preferences, but some may also remove or obscure evidence, giving rise to memory tax. Since memory benefit reflects the expected gains of its design, we then focus more on the overlooked memory tax, study how it arises from the system and evaluation side, and generalize the concept to more application scenarios of memory. We hope the notion of memory tax can provide a new lens and shift more attention from standalone performance to the net value of memory, informing more reliable memory designs and evaluations. We will open-source the code and results to the community. Project Page: https://hexuan21.github.io/MemoryTax/