REVIEW 6 cited by
On Memory Construction and Retrieval for Personalized Conversational Agents
Not yet reviewed by Pith; the record is open.
This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.
SPECIMEN: schema-true, not a live event
T0 review · schema-true
One-sentence machine reading of the paper's core claim.
pith:XXXXXXXX · record.json · timestamp
read the original abstract
To deliver coherent and personalized experiences in long-term conversations, existing approaches typically perform retrieval augmented response generation by constructing memory banks from conversation history at either the turn-level, session-level, or through summarization techniques.In this paper, we present two key findings: (1) The granularity of memory unit matters: turn-level, session-level, and summarization-based methods each exhibit limitations in both memory retrieval accuracy and the semantic quality of the retrieved content. (2) Prompt compression methods, such as LLMLingua-2, can effectively serve as a denoising mechanism, enhancing memory retrieval accuracy across different granularities. Building on these insights, we propose SeCom, a method that constructs the memory bank at segment level by introducing a conversation segmentation model that partitions long-term conversations into topically coherent segments, while applying compression based denoising on memory units to enhance memory retrieval. Experimental results show that SeCom exhibits a significant performance advantage over baselines on long-term conversation benchmarks LOCOMO and Long-MT-Bench+. Additionally, the proposed conversation segmentation method demonstrates superior performance on dialogue segmentation datasets such as DialSeg711, TIAGE, and SuperDialSeg.
Forward citations
Cited by 6 Pith papers
-
LeanMem: Simple and Efficient Long-Term Memory for LLM Agents
By routing dialogue segments into profile, event, and record memory, updating only events, and planning retrieval per query, LeanMem reports accuracy gains up to 15.1 points over memory baselines at lower cost.
-
A-TMA: Decoupling State-Aware Memory Failures in Long-Term Agent Memory
Explicit bank, retrieval, and QA state roles reduce ghost-memory failures on conflict-heavy LTP and improve some temporal scores on LoCoMo, with host-dependent gains.
-
Persona2Web: Benchmarking Personalized Web Agents for Contextual Reasoning with User History
Persona2Web is a new open-web benchmark where agents must infer a user's preferences from synthetic browsing history to solve intentionally ambiguous queries; current best agents score 13% success.
-
Locomo-Plus: Beyond-Factual Cognitive Memory Evaluation Framework for LLM Agents
Current LLMs and memory systems score far lower when a later question is semantically unrelated to an earlier implicit constraint, and LoCoMo-Plus evaluates this using constraint-consistency judging instead of string ...
-
MemTool: Optimizing Short-Term Memory Management for Dynamic Tool Calling in LLM Agent Multi-Turn Conversations
MemTool is a short-term memory framework with three modes (autonomous, workflow, hybrid) that lets LLM agents add and remove tools across multi-turn conversations, evaluated over 100 turns on 13+ models.
-
RGMem: Renormalization Group-inspired Memory Evolution for Language Agents
RGMem, a hierarchical memory framework with thresholded updates inspired by renormalization group ideas, reports state-of-the-art scores on the LOCOMO long-term conversational memory benchmark.
Discussion (0). Continue with ORCID to comment.