CABLE - Extending the Reach of Memory Retrieval via Complementary Antecedent-Based Linking and Expansion

Sep 15, 2026

Uploaded by Maeve

As LLM agents operate across structured workflows and sessions, long-term history does not guarantee that later contexts can recover relevant evidence through a bounded memory interface. The study targets evidence-reachability in long-term conversational memory, where retrieval often depends on semantic similarity: it succeeds for topical recall yet misses earlier experiences, plans, or motivations that explain later events. The proposed CABLE plugin builds sparse retriever-complementary links and expands retrieved seeds to surface implicit supporting evidence. Evaluations on LoCoMo and MA-LongMemEval show higher mean LLM-judge scores, with the largest gains when evidence is distributed across memories or sessions.

LanguageEnglish

CategoryResearch & Report

FormatPDF

Pages20

Reading Time50 min

AccessFree to download

Introduction

Evidence-reachability problem in long-term conversational memory

Limits of semantic-overlap memory graph expansion

Design criterion: retriever-complementary associations

Proposed method: CABLE and its linking/expansion mechanism

Experiments and results

arXiv : 2608 . 179 1 1v 1 [cs .CL] 18 Aug 2026 CABLE: Extending the Reach of Memory Retrieval via Complementary Antecedent-Based Linking and Expansion Zheling Tan1, Jin Gao1, Dequan Wang1,2 ∗ 1Shanghai Jiao Tong University 2Shanghai Innovation Institute Abstract As LLM agents operate across structured workflows and sessions, preserving long-term history does not ensure that later contexts can recover relevant evidence through a bounded memory interface. We study this evidence-reachability problem in long-term conversational memory, where retrieval still relies heavily on semantic similarity. This works well for topical recall, but it often misses earlier experiences, plans, or motivations that are semantically distant from the later events they help explain. Existing memory graphs provide cross-memory structure, yet links driven mainly by semantic overlap can duplicate what the host retriever already recovers. We argue that link construction should instead prioritize a sparse set of retriever-complementary associations. We present CABLE (Complementary Antecedent-Based Linking and Expansion), a plug-in augmentation that constructs links designed to extend the host retriever’s direct semantic reach. For each new memory, CABLE generates antecedent-oriented queries, retrieves prior memories, subtracts candidates in the direct semantic neighborhood, and verifies the remainder before adding the accepted complementary associations into a sparse directed graph. At retrieval time, CABLE expands the host system’s retrieved seeds along these links to surface implicit supporting evidence. We evaluate CABLE with A-MEM on LoCoMo and MA-LongMemEval, and further integrate it into SimpleMem and Mem0g on LoCoMo, using Qwen3.5-27B, DeepSeek-chat, and GPT-4o-mini. CABLE yields higher mean LLM-judge scores in every evaluated system-level setting, with the largest gains in categories where useful evidence is distributed across memories or sessions, including open-domain, multi-session, and preference-oriented questions. These results support prioritizing sparse, reasoning-relevant associations that complement rather than duplicate the host retriever. The core CABLE implementation is available at [https://github.com/TanZheling/CABLE](https://github.com/TanZheling/CABLE). 1 Introduction LLM agent systems are increasingly organized as execution graphs of specialized components and explicit state transitions (Wu et al., 2024; Zhuge et al., 2024; Zhang et al., 2025a) . As these systems operate across component and session boundaries, long-term history can no longer be assumed to remain inside one prompt. This creates an information-access problem: execution graphs specify where computation proceeds, but not how later components can recover the relevant history within a bounded context. Long-term conversational memory provides a controlled setting for studying one concrete instance of this evidence-reachability problem: whether stored evidence can still be retrieved when it lies outside the direct semantic neighborhood of a later query. Existing systems typically store past interactions as compact memory entries and retrieve a small subset for each new query (Park et al., 2023; Packer et al., 2023; Hu et al., 2025b) . Semantic retrieval is effective when the query and the required memory are close in embedding space, even if their surface forms differ. It is less reliable when the query concerns a later ∗ Corresponding author: [dequanwang@sjtu.edu.cn](dequanwang@sjtu.edu.cn) Figure 1: Stored evidence can remain outside direct retrieval. Direct semantic retrieval returns Memory B but not Memory A, which contains the required answer. During construction, CABLE stores the antecedent link (A → B); at query time, expansion starts from the retrieved Memory B and follows its incoming link to recover Memory A. event but answering it requires an earlier memory about an experience, plan, motivation, or background event. In such cases, the r

FAQs

What evidence-reachability problem does the paper address?

It studies whether stored evidence can be retrieved when it lies outside the direct semantic neighborhood of a later query, despite the use of bounded memory interfaces.

How does CABLE construct memory links differently from existing approaches?

CABLE generates antecedent-oriented queries for new memories, retrieves prior memories, subtracts candidates within the direct semantic neighborhood, and adds only complementary remainder associations into a sparse directed graph.

What does CABLE do during retrieval time?

It expands the host system’s retrieved seeds along the incoming complementary links to surface implicit supporting evidence that would otherwise remain inaccessible.

Sep 15, 2026

Uploaded by Maeve