{"paper":{"title":"MemoryBank: Enhancing Large Language Models with Long-Term Memory","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"MemoryBank equips large language models with a long-term memory system modeled on human forgetting.","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"He Ye, Lianghong Guo, Qiqi Gao, Wanjun Zhong, Yanlin Wang","submitted_at":"2023-05-17T14:40:29Z","abstract_excerpt":"Revolutionary advancements in Large Language Models have drastically reshaped our interactions with artificial intelligence systems. Despite this, a notable hindrance remains-the deficiency of a long-term memory mechanism within these models. This shortfall becomes increasingly evident in situations demanding sustained interaction, such as personal companion systems and psychological counseling. Therefore, we propose MemoryBank, a novel memory mechanism tailored for LLMs. MemoryBank enables the models to summon relevant memories, continually evolve through continuous memory updates, comprehend"},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"MemoryBank enables the models to summon relevant memories, continually evolve through continuous memory updates, comprehend, and adapt to a user personality by synthesizing information from past interactions.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"The assumption that incorporating the Ebbinghaus Forgetting Curve will allow the AI to selectively preserve memory in a way that improves long-term interaction without introducing errors or forgetting critical information.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"MemoryBank equips LLMs with long-term memory using Ebbinghaus-inspired updates, allowing recall and personality adaptation in chatbots like SiliconFriend.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"MemoryBank equips large language models with a long-term memory system modeled on human forgetting.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"12c98d464c9fb365a6c4953ab7b754dbda37e9113f49b3164c643afbd0311ebc"},"source":{"id":"2305.10250","kind":"arxiv","version":3},"verdict":{"id":"997d2576-29e8-4f73-8423-942b83f751dd","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-16T07:13:57.247705Z","strongest_claim":"MemoryBank enables the models to summon relevant memories, continually evolve through continuous memory updates, comprehend, and adapt to a user personality by synthesizing information from past interactions.","one_line_summary":"MemoryBank equips LLMs with long-term memory using Ebbinghaus-inspired updates, allowing recall and personality adaptation in chatbots like SiliconFriend.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"The assumption that incorporating the Ebbinghaus Forgetting Curve will allow the AI to selectively preserve memory in a way that improves long-term interaction without introducing errors or forgetting critical information.","pith_extraction_headline":"MemoryBank equips large language models with a long-term memory system modeled on human forgetting."},"references":{"count":12,"sample":[{"doi":"","year":1901,"title":"Language models are few-shot learners","work_id":"677093e0-2019-45af-8c52-d9b33dec7e3d","ref_index":1,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":null,"title":"PaLM: Scaling Language Modeling with Pathways","work_id":"a94f3ef7-2c49-4445-93fe-6ec16aafd966","ref_index":2,"cited_arxiv_id":"2204.02311","is_internal_anchor":true},{"doi":"","year":null,"title":"Scaling Instruction-Finetuned Language Models","work_id":"8405abb1-7558-4fdf-af24-f4c52fa77a06","ref_index":3,"cited_arxiv_id":"2210.11416","is_internal_anchor":true},{"doi":"","year":null,"title":"Neural Turing Machines","work_id":"0a5ce53c-9670-42b9-8be5-386de7eed50c","ref_index":4,"cited_arxiv_id":"1410.5401","is_internal_anchor":true},{"doi":"","year":2004,"title":"Dense Passage Retrieval for Open-Domain Question Answering","work_id":"3d6f2008-b001-4542-ba3f-192f6880c74b","ref_index":5,"cited_arxiv_id":"2004.04906","is_internal_anchor":true}],"resolved_work":12,"snapshot_sha256":"900bfb09e40aeaa7a3b4558d4ad7f9eb29128b6563f1d6d13a313170758b509f","internal_anchors":8},"formal_canon":{"evidence_count":3,"snapshot_sha256":"524f18bee1e18f1d3ca3c63f8758116a206a03c9d09a03f450c3cfba73954e66"},"author_claims":{"count":0,"strong_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"builder_version":"pith-number-builder-2026-05-17-v1"}