Agent control cuts memory use for assistant queries by 92 percent
APDMem: Agent-Controlled Progressive Disclosure for Query-Adaptive Long-Term Memory
Computation and LanguageArtificial Intelligence
Summary
Answering questions using long conversations is hard because the right details may be buried deep in many messages. The authors created a system called APDMem where conversation memories are organized into layers, from simple summaries to full messages. When a question comes in, the system looks at summaries first and only digs deeper if it needs more details. This method helps find answers accurately without reading everything, saving time and computing resources.
What this means in practice
- •For personal assistant developers: Reduce computational costs while maintaining accuracy in long conversation query answering by adaptively retrieving layered memory details.
- •For customer support automation teams: Improve retrieval precision for multi-turn support conversations by accessing progressively detailed memory layers as query complexity increases.
Authors
Chin-Lun Fu, Anagha Kulkarni, Hong Ni, Behrouz Madahian
Abstract
Personalized LLM assistants must recover sparse evidence from long conversation histories across queries of varying complexity. We introduce APDMem (Agent-controlled Progressive Disclosure Memory), a hierarchical long-term memory architecture that applies progressive disclosure to memory retrieval. Rather than relying on a flat memory store or fixed retrieval granularity, APDMem represents conversation history as four progressively detailed layers: thematic summaries, personalized key facts, turn-level evidence notes, and raw messages. At inference time, a controller applies progressive disclosure to the memory hierarchy: it first reads high-level summaries and drills into finer evidence only when needed. This creates an adaptive cost-fidelity trade-off: simple queries can terminate early, while complex temporal, multi-hop, or exact-evidence queries trigger deeper inspection. A note synthesizer converts retrieved evidence into a query-focused structure that consolidates facts, orders events, and flags contradictions before final answer generation. Experiments on LongMemEval show that APDMem achieves strong performance for long-context memory reasoning while accessing only 8% of the total conversations.