LabLAB-002The Memory Field Guide

A field guide to memory — human & artificial

A language model is brilliant, and it forgets everything the moment you close the tab.

Its context window is working memory — a fixed buffer where thinking happens. When the conversation ends, it’s gone. In human terms that’s severe anterograde amnesia: intact intelligence, no ability to form a lasting memory. Solving it means borrowing from the one system that already did — the brain. This is a guide to how memory actually works, how AI is trying to copy it, and where the ground is still open.

Explore the mechanisms

The word for a memory

An engram is the physical trace a memory leaves in the brain — a term coined by Richard Semon in 1904Semon 1904, when it was pure theory. A century later, Susumu Tonegawa’s lab found the cells: they tagged the neurons active during an experience, then reactivated them with light and watched the memory returnTonegawa 2012–2015. Engrams are real, locatable, and switchable. This guide is organized around the mechanisms those traces are built from — one specimen at a time.

The spine

Eleven mechanisms

Each specimen reads the same four ways: how the brain does it, the model that formalizes it, how AI implements or fails it, and the mapof who has claimed it. Where the map runs thin, there’s open ground.

Showing 11 of 11 mechanisms

EncodingAI: occupied

Encoding & salience

Not everything that happens is worth keeping. Something has to decide.

In the brain

The brain can’t keep everything, and shouldn’t. Attention and arousal gate what the hippocampus bothers to consolidate, and a dopamine signal from the midbrain tags an event as worth keeping when it’s novel, surprising, or tied to a goal. The amygdala adds a second vote for anything with emotional stakesMcGaugh 2000. Most of a day never crosses that threshold — and that’s the point.

The model

Formalized, it’s a weighting problem: each experience earns an encoding strength from several factors at once — novelty, relevance to current goals, emotional charge, likely future use. Higher-salience items start with more strength and, in richer models, decay more slowly. Salience isn’t one number; it’s a small profile.

In AI

Nearly every AI memory system scores candidate memories at write time — an importance heuristic, or an LLM simply asked “how important is this?” engram scores four dimensions (novelty, relevance, emotional, predictive) as it encodes; Stanford’s Generative Agents scored importance on write; DeWix’s granted patent weights parts of speech including an emotional axis. The hard part was never scoring — it’s scoring well, and adapting the weights to one particular person.

The map

Multi-dimensional salience — partially anticipated (KR102809164B1 claim 2).

11 prior-art references

1 post-priority (field context, not prior art)
  • SCM~2026SCM — multi-dimensional importance + dual-phase consolidationpaperarXiv

Salience scoring is thoroughly occupied. The live edge is learned salience — weights that adapt to what a specific user actually reinforces, rather than hand-tuned constants.

TransformationAI: occupied

Consolidation

Memories aren't saved when they're made. They're rebuilt, offline, while you sleep.

Contested — How consolidation works long-term is actively debated — standard systems consolidation vs. multiple-trace theory.

In the brain

A memory isn’t fixed the moment it forms. Over hours to years it’s reorganized: the hippocampus captures an episode fast, then replays it — especially during slow-wave sleep, carried on sharp-wave ripples — to the neocortex, which slowly integrates it into long-term knowledgeDiekelmann 2010.

Whether the hippocampus ever fully hands off, or is always needed to replay a detailed memory, is still debatedNadel 1997.

The model

The Complementary Learning Systems account: a fast hippocampal learner and a slow neocortical one. Interleaving replay of old and new lets the slow system extract structure without catastrophically overwriting what it already knewMcClelland 1995 — the reason sleep matters for memory at all.

In AI

In practice: an offline “sleep” pass that merges redundant memories and extracts patterns. engram runs a consolidation cycle between sessions (merge duplicates, distill old episodics to gist); Microsoft filings cluster a vector store with DBSCAN/HDBSCAN and select by age and effectiveness. engram’s two-pass design — a cheap model finds merge candidates, a stronger one rewrites them — mirrors the fast/slow split directly.

The map

Two-pass consolidation — anticipated (WO2025147358A1, US20250021474A1).

13 prior-art references

3 post-priority (field context, not prior art)

Consolidation is heavily patented for the single-session case. The open ground is long-horizon consolidation: what should distill over months and years, not one night.

StorageAI: occupied

Decay & forgetting

Forgetting is not the failure of memory. It is one of its functions.

Contested — Whether forgetting is true decay or retrieval interference is a century-old, still-unsettled question.

0%25%50%75%100%day 07d14d21d30d
19% retained after 4 days

In the brain

Memories weaken with time unless they’re renewed. Ebbinghaus measured it on himself in 1885 — a steep initial drop that flattens into a long tailEbbinghaus 1885. Forgetting isn’t simply failure: it clears capacity and keeps one-off events from overwriting durable knowledge.

Whether the trace physically fades or is just crowded out by competitors is a century-old open question.

The model

The forgetting curve. Retention falls off as an exponential or, better-fitting, a power-law function of elapsed timeWickelgren 1974, modulated by how strongly it was encoded and flattened each time it’s retrieved — the mathematics behind spaced repetition.

In AI

The most-copied idea in AI memory. MemoryBank applies Ebbinghaus directly; a granted EMC/Dell patent decays relevance by linear, exponential, or step functions; DeWix claims R = e^(−a·x). engram computes strength on read from salience, access history, and elapsed time rather than storing a stale number. Most systems still decay linearly, which doesn’t match the brain’s power-law shape.

The map

Forgetting-curve decay — anticipated (KR102809164B1 claim 1; Alai Vault spec).

11 prior-art references

Decay curves are anticipated many times over. What almost no one models: the difference between true forgetting (the trace is gone) and retrieval failure(it’s intact, but the path to it is lost).

StorageAI: partial

Interference

New memories crowd out old ones. Old ones distort new ones. They compete.

In the brain

Memories compete. Proactive interference: older memories impair learning and recall of newer, similar ones. Retroactive: new learning degrades oldUnderwood 1957. The brain actively inhibits competing traces during encoding and retrieval — it doesn’t just file everything side by side and hope.

The model

Interference theory frames forgetting as competition among similar traces at retrieval — scaled by how similar they are and how many compete — rather than as passive decay. The more crowded the neighborhood, the harder any one memory is to find.

In AI

Suddenly urgent for agents. The Unable to Forgetbenchmark showed an LLM’s retrieval accuracy collapsing toward zero as conflicting values accumulate — proactive interference, measured. FadeMem applies contradiction-triggered graded attenuation (an LLM classifies the conflict); engram weakens an old memory’s salience the moment a new one supersedes it; Beijing Lingxin resolves conflicts with a BERT classifier.

The map

Proactive interference — eroded to a narrow claim (FadeMem §2.3, ~5 weeks pre-priority).

5 prior-art references

1 post-priority (field context, not prior art)

Interference is well diagnosed and barely solved. There are benchmarks and a few attenuation tricks, but reliably suppressing outdated information — without losing what’s still true — is open.

RetrievalAI: occupied

Retrieval & strengthening

Every time you remember something, you change it — and usually make it stronger.

In the brain

Retrieval is not a passive read. Recalling something strengthens it more than re-studying it would — the testing effectRoediger 2006. Recall is cue-driven and reconstructive: a partial cue reactivates the whole pattern, and neurons that fire together wire together, so use is itself a form of encoding.

The model

Retrieval-induced strengthening: each successful recall adds to a memory’s strength and resets its decay clock. Storage strength (how well-learned) and retrieval strength (how accessible right now) are distinct — you can know something deeply yet fail to find it, and finding it makes it easier to find again.

In AI

Retrieval is the one part every RAG system does. The subtler move — letting a read changethe memory — is rarer. engram boosts strength when a memory is recalled; Generative Agents fold recency, importance, and relevance into a score computed at read time. A real gap: if the agent leans on a session briefing instead of querying, no strengthening fires — “use it or lose it” only triggers on explicit lookups.

The map

Dynamic strength on read — §103 exposed (Generative Agents retrieval-time scoring).

14 prior-art references

1 post-priority (field context, not prior art)

Retrieval-time scoring is §103-occupied. Passive strengthening — a memory gaining weight because it was used in conversation, not explicitly queried — is thin.

TransformationAI: occupied

Reconsolidation

Recall makes a memory briefly editable. Then it's re-stored, sometimes rewritten.

Contested — The boundary conditions for reconsolidation — when a recalled trace becomes labile — are still debated.

In the brain

Retrieving a settled memory can return it to a fragile, editable state — it has to be re-stored, and can be altered before it re-stabilizesNader 2000. This is why memory is reconstructive, not reproductive: every recall is a chance to rewrite.

Exactly which memories become labile, and when, is still being worked out.

The model

A recalled trace briefly becomes writable; an updated version overwrites the old while keeping the same identity. It’s how “Mike has one son” becomes “Mike has two sons” without deleting and recreating the memory — the fact evolves in place.

In AI

Mem0’s UPDATE operation — edit a memory’s content in place, keep its id, log the prior version — is exactly this, and it’s the closest §102 prior art there is. engram’s reinforce-with-new-content does the same: it rewrites the text while strengthening the memory, logging the old wording first as a safety net.

The map

In-place update — §102 anticipated (Mem0 UPDATE + history log).

1 prior-art reference

In-place update is anticipated. What’s unexplored is reconsolidation as deliberate distortion— letting recall reshape a memory toward the agent’s current understanding, the way human memory quietly does.

TransformationAI: occupied

Episodic ↔ semantic

"I met her on Tuesday" slowly becomes "I know her" — the event fades, the meaning stays.

Contested — Whether gist is compressed episodic detail or an independent multi-trace abstraction is unresolved.

In the brain

Two systems, first distinguished by Tulving: episodicmemory holds specific events (“I met her Tuesday”); semanticmemory holds general knowledge (“I know her”)Tulving 1972. Over time, repeated episodes shed their verbatim detail and leave gist — Fuzzy-Trace Theory holds the two as parallel traces, the verbatim one fading fasterBrainerd 2002.

Whether semantic memory is just compressed episodes or a separate abstraction is unsettled.

The model

Verbatim-to-gist degradation: detailed traces decay quickly while the extracted meaning persists. You keep “we argued about consciousness” long after the exact words are gone. Consolidation is the process that does the distilling.

In AI

engram tags each memory episodic or semantic and, during consolidation, compresses aging episodics into semantic gist with a small model; merged and generalized memories are born semantic. Microsoft’s layered working/episodic/collective memory and Alai Vault’s episodic/semantic tiers cover the split. One honest caveat: real semantic memory is gist across many episodes, not one old episode squeezed down.

The map

Episodic→semantic degradation — §103 exposed (Alai Vault, US20250371317A1).

7 prior-art references

1 post-priority (field context, not prior art)

The tiered split is §103-occupied. Distilling gist across many episodes — rather than compressing a single old one — is far less trodden.

EncodingAI: partial

Emotional modulation

The moments that mattered are the ones you keep. Feeling is the encoder's thumb on the scale.

In the brain

Emotion is the encoder’s thumb on the scale. Arousing events are remembered better and longer: the amygdala modulates hippocampal consolidation, and stress hormones flag an event as important as it’s laid downMcGaugh 2000. It’s why a flashbulb memory survives while a thousand ordinary Tuesdays vanish — and emotional memories don’t just start stronger, they decay more slowly.

The model

Two effects, not one: an affective weight on initial encoding strength, and a reduced decay rate. Emotion changes both the height and the slope of the forgetting curve — a fact most implementations quietly drop.

In AI

Usually collapsed into a single salience axis. DeWix weights adjectives “reflecting the emotional context”; Dynamic Affective Memory Management keeps an affective dimension in a numeric salience profile; engram scores an emotional dimension at encoding. But almost nothing lets emotion change the decay ratethe way the amygdala does — it’s treated as starting weight only.

The map

Emotional salience axis — partially anticipated (KR claim 2; Dynamic Affective Memory).

3 prior-art references

Emotion-as-initial-weight is partially claimed. Emotion-as-slower-forgetting — a distinct decay rate for what mattered — is largely open.

RetrievalAI: partial

Prospective memory

Remembering to do something later — the memory that fires itself at the right moment.

In the brain

Remembering to do something later— take the medicine at eight, pass on the message. Unlike memory for the past, a prospective memory has to fire itself at the right moment, cued by a time or an event, while you’re busy with something elseEinstein 1990. It leans on prefrontal monitoring and a spontaneous retrieval when the cue finally appears.

The model

An intention stored together with a trigger — time-based (“at 8pm”) or event-based (“when I see Katie”) — that has to be noticed and executed without an explicit prompt to remember it. The hard part is the noticing.

In AI

Thin. Assistant “reminders” (Meta’s proactive-reminder patents) and cognitive architectures like ACT-R cover narrow, scheduled forms. But a general agent that spontaneously acts on a stored future intention when a matching context appears— not when it’s asked — is barely built. engram lists it as a TODO, not a feature.

The map

Prospective memory — anticipated in narrow forms (ACT-R/Soar; Meta US11567788B1).

3 prior-art references

Open ground. Genuine prospective memory for agents — self-triggering intentions that fire on context, not on a prompt — is largely unclaimed.

RetrievalAI: occupied

Temporal context & association

Recalling one thing pulls its neighbors with it — everything you learned in the same hour.

In the brain

Memories formed close together in time get linked. Recalling one tends to drag its neighbors along — the contiguity effectHoward 2002. A slowly drifting internal “context” is bound to each memory as it forms; reinstating that context reactivates everything you encoded in the same stretch. It’s why a smell can return an entire afternoon.

The model

The Temporal Context Model: a drifting context vector is associated with each item, and retrieval reinstates that context, which in turn cues temporally-nearby itemsHoward 2002. Association falls off smoothly with temporal distance.

In AI

engram returns “temporal siblings” — memories formed in the same session — alongside direct recall matches, and boosts the current project’s memories at session start (context-dependent retrieval by working directory). Zep/Graphiti build an explicit temporal knowledge graph; PAM retrieves through temporal co-occurrence.

The map

Temporal association — §103 obvious (Temporal Context Model; PAM).

5 prior-art references

§103-occupied through the Temporal Context Model and graph systems. Rich context reinstatement — reconstructing a whole prior context, not just same-session tagging — is under-built.

TransformationAI: open ground

Schema & assimilation

New facts don't land on blank ground. They're folded into what you already believe.

In the brain

New information never lands on blank ground. It’s assimilated into existing schemas— mental frameworks that shape what gets encoded and how it’s later reconstructed. Bartlett’s readers, retelling an unfamiliar folk tale, unconsciously reshaped it toward their own cultural expectationsBartlett 1932. Schema-consistent facts consolidate faster; there’s already a slot for them.

The model

Assimilation versus accommodation: fit new data into an existing frame, or revise the frame to fit the data. Schemas speed the encoding of consistent information and bias the reconstruction of anything ambiguous — memory as an active model, not a passive log.

In AI

The least-built mechanism here. Knowledge-graph memories (Microsoft’s brain-modeling KG; Cerego’s related-concept generation) gesture at structure, and entity tracking would be a first step — but AI memory systems overwhelmingly store isolated memory atoms, not memories assimilated into an evolving model of the world. engram’s own review flags this as its biggest missing piece.

The map

Schema assimilation — largely open on the AI side; classic in cognitive science.

2 prior-art references

Open ground. Assimilating memories into an evolving schema — rather than accumulating independent atoms — is close to untouched on the AI side.

The map

The landscape

Every patent, paper, and open-source system found across five rounds of adversarial prior-art search. Sorted by priority date, the shape is the finding: the field is occupied, not crowded — a convergence in 2023–2026, with Microsoft, the ex-Cerego team behind Alai Vault, DeWix, Snap, Meta, Samsung, and a fast-moving research frontier all arriving at once.

2023202420252026pre-’23Cerego — 1999 — Learning-effectiveness / memory-retention modelingHoward — ~2002 — Temporal Context ModelAnderson — ~2013 — ACT-R / Soar / ACT-RΦ — cognitive architecturesDeepMind — ~2015 — Prioritized experience replay; salience-weighted transitions (RL)Cerego (Smith Lewis) — 2015 — Personalized learning / learner-adaptive enginesCerego (Smith Lewis, Harlow) — 2017 — Personalized learning, automated generationCerego (Smith Lewis, Harlow) — 2017 — Predictive memory-retention / forgetting model with thresholds + schedulingCerego — 2017 — Auto-generating related conceptsMicrosoft — ~2018 — Knowledge graph that "models the human brain memory system"ServiceNow — 2018 — Focused conversation context management; overlay context from stored episodesMeta — ~2019 — Memory-graph traversal; "reactive/proactive memory" for assistantsMeta — 2019 — Proactive reminders for assistant systemsEMC / Dell — ~2020 — "Relevance decay" — linear / exponential / step functionsHuman AI Labs (Personal.ai) — 2020 — "Memory retention system" — automatic capture of searchable human memoryCerego (Harlow) — 2020 — Personalized learning systemMicrosoft — ~2023 — Foundational RAG (feature-vector similarity → augmented prompt)Modulus AI — ~2023 — Preserves salient entities/facts under high-compression summarizationAlai Vault LLC (Smith Lewis, Harlow) — 2023 — Biologically-inspired memory: power decay of salience, episodic/semantic tiers, agent memory toolShinn — 2023 — Reflexion: Language Agents with Verbal Reinforcement LearningPark — ~2023 — Generative Agents: Interactive Simulacra of Human BehaviorMicrosoft — 2023 — Dynamic LLM prompt construction from summary + scoring modelZhong — ~2023 — MemoryBank: Enhancing LLMs with Long-Term MemoryMicrosoft — 2023 — Prompt construction by dynamically selecting context by semantic relevanceSnap — 2023 — Chatbot long-term memory: summarize → store → semantic-search → augment promptMicrosoft — 2023 — Extract memories via generative model → DBSCAN/HDBSCAN cluster into memory banksPacker — 2023 — MemGPT: Towards LLMs as Operating SystemsNavinfo Europe — ~2024 — Catastrophic forgetting + episodic/semantic (neural-net training)Escendant Ltd — 2024 — Hierarchical topic manager; session boundaries before context window fillsMicrosoft — 2024 — Layered working/episodic/collective memory; embedding-compressed historyMicrosoft — ~2024 — Semantic memory vector-DB consolidation; DBSCAN clustering; age/effectiveness selectionBeijing Lingxin Intelligent Tech — 2024 — LLM memory retrieval; BERT-based "Memory Conflict Resolution"Samsung — ~2025 — KV-cache compression; frequency + recency importance scoringRasmussen — 2025 — Zep / Graphiti — temporal knowledge-graph memoryXu — 2025 — A-MEM: Agentic Memory for LLM AgentsDeWix Co., Ltd. — 2025 — Memory-curve context maintenanceTaranjeet — 2025 — Mem0 — scalable long-term memory for AI agentsWang — 2025 — PI-LLM ("Unable to Forget") — proactive interference in LLMsDynamic Affective Memory Management — 2025 — Dynamic Affective Memory ManagementWei — 2026 — FadeMem: Biologically-Inspired Forgetting for Efficient Agent MemoryPAM — ~2026 — PAM — retrieval through temporal co-occurrenceGentleman-Programming — ~2026 — Gentleman-Programming/engramforamorent — 2026 — foramorent/engram-ai-memory
Each dot is a prior-art reference. The cluster is the point — the field is occupied, not crowded.

46 of 46 references

CoverageCite
CeregoLearning-effectiveness / memory-retention modelingUS6652283B1patent1999Granted 2003

Foundational predictive modeling of learning effectiveness and memory retention — the root of the Cerego lineage.

Howard, KahanaTemporal Context Modeltcmpaper~2002Howard & Kahana, 2002

The temporal-contiguity effect and the drifting-context account underlying session-based association.

Anderson, Laird, DancyACT-R / Soar / ACT-RΦ — cognitive architecturesact-r-soarpaper~2013Cognitive architectures, 1990s–2013

Decades of cognitive-architecture work: prospective memory, affect-modulated declarative memory, base-level activation decay.

DeepMindPrioritized experience replay; salience-weighted transitions (RL)US9679258B2patent~2015Granted 2017 (~2015 priority)

Prioritized experience replay weights transitions by salience — reinforcement learning, not LLM memory, but foundational salience-weighting art.

Cerego (Smith Lewis)Personalized learning / learner-adaptive enginesUS20170075881A1patent2015Application

Personalized learning with learner-adaptive scheduling engines.

Cerego (Smith Lewis, Harlow)Personalized learning, automated generationUS11721230B2patent2017Granted

Personalized learning with automated content generation. Family also includes US11081018B2.

Cerego (Smith Lewis, Harlow)Predictive memory-retention / forgetting model with thresholds + schedulingUS11776417B2patent2017Granted 2023

The most on-point Cerego patent: predictive retention/forgetting model with retention thresholds and review scheduling.

CeregoAuto-generating related conceptsUS11487804B2patent2017Granted

Automatically generating related concepts from learning material. Family also includes US11238085B2.

MicrosoftKnowledge graph that "models the human brain memory system"US20180114111A1patent~2018Application (~2018)

Knowledge-graph memory explicitly framed as modeling the human brain memory system.

ServiceNowFocused conversation context management; overlay context from stored episodesUS20190294675A1patent2018Application

Manages focused conversation context, overlaying context retrieved from stored episodes.

MetaMemory-graph traversal; "reactive/proactive memory" for assistantsUS11657094B2patent~2019Granted (~2019 priority)

Traverses a memory graph; distinguishes reactive and proactive memory for assistant systems.

MetaProactive reminders for assistant systemsUS11567788B1patent2019Granted

Generates proactive reminders — a prospective-memory analog for assistants.

EMC / Dell"Relevance decay" — linear / exponential / step functionsUS10885464B1patent~2020Granted 2021 (~2020 priority)

Applies relevance decay via linear, exponential, or step functions.

Human AI Labs (Personal.ai)"Memory retention system" — automatic capture of searchable human memoryUS20210182020A1patent2020Application

Automatic capture of a searchable store of human memory — the user-facing half of a dual-system memory design.

Cerego (Harlow)Personalized learning systemUS11545042B2patent2020Granted 2023

Later Cerego personalized-learning patent extending the retention-modeling lineage.

MicrosoftFoundational RAG (feature-vector similarity → augmented prompt)US20240346256A1patent~2023Application (~2023)

Foundational retrieval-augmented generation: feature-vector similarity retrieval feeding an augmented prompt.

Modulus AIPreserves salient entities/facts under high-compression summarizationUS12008332B1patent~2023Granted 2024 (~2023 priority)

Preserves salient entities and facts when summarizing under high compression — pre-compression preservation.

Alai Vault LLC (Smith Lewis, Harlow)Biologically-inspired memory: power decay of salience, episodic/semantic tiers, agent memory toolUS20240289863A1patent2023Applications pending

Strongest single §103 obviousness reference. Specification teaches parameterized age-dependent decay of salience, power decay of episodic memories, episodic/semantic tiers, vector-DB long-term memory, a conversational memory tool — though the 20 claims cover brand-marketing agent personalization only. Inventors are ex-Cerego. Family: WO2024178435A1, CA3284082A1.

Shinn, et al.Reflexion: Language Agents with Verbal Reinforcement Learningreflexionpaper2023arXiv:2303.11366

First-person self-reflective episodic buffer — anticipates first-person memory encoding.

Park, et al.Generative Agents: Interactive Simulacra of Human Behaviorgenerative-agentspaper~2023Stanford, Apr 2023

Retrieval-time scoring by recency + importance + relevance — §103 art for dynamic strength and context-dependent retrieval.

MicrosoftDynamic LLM prompt construction from summary + scoring modelUS12462095B2patent2023Granted 2025-11-04

Constructs LLM prompts dynamically from a running summary plus a memory-scoring model.

Zhong, et al.MemoryBank: Enhancing LLMs with Long-Term Memorymemorybankpaper~2023Zhong et al., 2023

Applies the Ebbinghaus forgetting curve to LLM memory.

MicrosoftPrompt construction by dynamically selecting context by semantic relevanceUS20240394477A1patent2023Application

Selects context for prompt construction by semantic relevance.

SnapChatbot long-term memory: summarize → store → semantic-search → augment promptUS20240414108A1patent2023Application

End-to-end chatbot long-term memory: summarize, store, semantic-search, augment the prompt.

MicrosoftExtract memories via generative model → DBSCAN/HDBSCAN cluster into memory banksUS20250021474A1patent2023Application

Extract-then-cluster consolidation (DBSCAN/HDBSCAN) into memory banks — anticipates two-pass consolidation.

Packer, et al.MemGPT: Towards LLMs as Operating Systemsmemgptpaper2023Packer et al., 2023

OS-inspired tiered memory (main context vs. archival) with paging between tiers.

Navinfo EuropeCatastrophic forgetting + episodic/semantic (neural-net training)US20240119280A1patent~2024Application (~2024)

Addresses catastrophic forgetting with an episodic/semantic split — for neural-network training, not a retrievable memory store.

Escendant LtdHierarchical topic manager; session boundaries before context window fillsUS12093846B1patent2024Granted 2024-09-17

Hierarchical topic management that draws session boundaries before the context window fills — granted pre-compression preservation art.

MicrosoftLayered working/episodic/collective memory; embedding-compressed historyUS20250371317A1patent2024Application

Layered working / episodic / collective memory with embedding-compressed history.

MicrosoftSemantic memory vector-DB consolidation; DBSCAN clustering; age/effectiveness selectionWO2025147358A1patent~2024PCT published 2025-07-10

Consolidates a semantic-memory vector DB via DBSCAN clustering with age/effectiveness-based selection — anticipates two-pass consolidation.

Beijing Lingxin Intelligent TechLLM memory retrieval; BERT-based "Memory Conflict Resolution"CN119903125Bpatent2024Granted 2025-08-22

Encoding-time memory-conflict resolution: a BERT model detects conflict and resolves by elimination or integration — overlaps part of the interference mechanism (a §103 reference, not anticipation).

SamsungKV-cache compression; frequency + recency importance scoringUS20250348732A1patent~2025Application (~2025)

KV-cache compression driven by tensor-level frequency + recency importance scoring.

Rasmussen, et al.Zep / Graphiti — temporal knowledge-graph memoryzep-graphitioss2025arXiv:2501.13956 (OSS)

Bounded edge deduplication for complexity reduction in a temporal knowledge graph.

Xu, et al.A-MEM: Agentic Memory for LLM Agentsa-mempaper2025arXiv:2502.12110

A top-k similarity neighbor set drives memory evolution (Zettelkasten-style linking).

DeWix Co., Ltd.Memory-curve context maintenanceKR102809164B1patent2025Granted 2025-05-21

The most directly relevant reference — claims (not just spec) recite the machinery: exponential decay R=e^(-a·x), top-20% summarization, timestamped storage, delete-low-retention (claim 1); importance from parts of speech with adjectives weighted for emotional context (claim 2); ML-derived decay rate (claim 4); semantic-similarity clustering and pruning (claim 5). Claim analysis via machine translation.

Taranjeet, et al.Mem0 — scalable long-term memory for AI agentsmem0oss2025arXiv:2504.19413 (OSS)

ADD/UPDATE/DELETE/NOOP operations with a bounded top-s dedup context. The UPDATE-with-history operation is exactly the in-place-update ('reconsolidation') mechanism — §102 anticipation for it.

Wang, SunPI-LLM ("Unable to Forget") — proactive interference in LLMspi-llmpaper2025arXiv:2506.08184

A benchmark diagnosing proactive interference as an LLM problem — establishes the term/problem pre-priority, but discloses no mechanism to solve it.

Dynamic Affective Memory Managementdammpaper2025arXiv:2510.27418

A multi-dimensional numerical salience profile updated by a non-LLM arithmetic rule — §103 art for a salience vector with an affective axis.

Wei, Peng, Dong, Xie, WangFadeMem: Biologically-Inspired Forgetting for Efficient Agent Memoryfademempaper2026arXiv:2601.18642 (Alibaba; Peking University)

The closest research reference. Does contradiction-triggered, encoding-time, graded arithmetic attenuation of memory strength in an LLM agent — most of what a proactive-interference mechanism does. But conflict is LLM-classified (GPT-4o-mini), it attenuates a scalar not a multi-dimensional vector, strength is a stored field batch-rewritten (not computed on read), and it uses one model tier. Predates engram's priority by ~5 weeks.

PAM — retrieval through temporal co-occurrencepampaper~2026arXiv:2602.11322

Retrieval through temporal co-occurrence — §103 art for temporal association.

Gentleman-ProgrammingGentleman-Programming/engramgentleman-engramoss~2026Open source

Conventional CRUD memory store (SQLite + FTS).

foramorentforamorent/engram-ai-memoryforamorent-engramoss2026Open source

Basic Ebbinghaus decay for AI memory.

SleepGate — proactive interference + sleep consolidationsleepgatepaper2026arXiv:2603.14517

Field context only — post-dates engram's priority, so not prior art. Combines proactive interference with sleep-style consolidation.

SCM — multi-dimensional importance + dual-phase consolidationscmpaper~2026arXiv:2604.20943

Field context only — post-priority. Multi-dimensional importance scoring with dual-phase consolidation.

HeLa-Mem — Hebbian episodic→semantic distillationhela-mempaper~2026arXiv:2604.16839

Field context only — post-priority. Hebbian distillation of episodic memories into semantic ones.

YourMemoryyourmemoryoss~2026Open source

Field context only — post-priority OSS memory system.

The uncharted region

Where the ground is still open

If the obvious mechanisms are occupied, where is the real work? Read as a map, the search points fairly clearly.

Evaluation

Almost everyone is building memory systems; almost no one agrees how to measure them. Most results lean on synthetic benchmarks that don’t capture real interaction — the gap behind the RECALL conversation-grounded benchmark.

Interference, solved not just diagnosed

LLMs demonstrably can’t suppress outdated information. FadeMem and engram both attempt mechanisms — but “the model retrieves a value it should have forgotten” is far better characterized than it is solved.

Long-horizon memory

Nearly every system operates on days or weeks. What a memory should do over months and years — what genuinely consolidates, what becomes irrecoverable versus merely hard to find — is close to unexplored.

Forgetting vs. retrieval failure

Human memory distinguishes “the trace is gone” from “the trace exists but the path is lost.” Most AI systems only delete. Graded, partially reversible forgetting has very little prior art and real conceptual room.

How this was built

Method & caveats

The landscape was compiled from five escalating rounds of adversarial prior-art search conducted during the engram project (May 2026) — broad academic and OSS sweeps, then patents across the US, Europe, Korea, and China, then focused deep dives on the closest references. The goal was deliberately hostile: not to confirm an idea was new, but to find everything that would prove it wasn’t.

  • This is a research record and a field reference — not legal advice, and not a freedom-to-operate or patentability opinion.
  • Patent numbers were verified against live Google Patents pages during the searches; the Korean patent’s claim analysis used machine translation.
  • Contested points in the neuroscience are marked ; they reflect active scientific debate, not settled fact.
  • engram is open source (github.com/mlapeter/claude-engram); the RECALL benchmark is at github.com/mlapeter/recall-bench.

References

Sources

  1. R. Semon (1904). Die Mneme (introducing the term "engram"). Wilhelm Engelmann, Leipzig.
  2. H. Ebbinghaus (1885). Über das Gedächtnis (Memory: A Contribution to Experimental Psychology). Duncker & Humblot.
  3. W. A. Wickelgren (1974). Single-trace fragility theory of memory dynamics. Memory & Cognition, 2(4).
  4. S. Tonegawa, X. Liu, S. Ramirez, et al. (2012–2015). Optogenetic identification and manipulation of memory engram cells. Nature; Science.
  5. E. Tulving (1972). Episodic and semantic memory. Organization of Memory (Academic Press).
  6. C. J. Brainerd & V. F. Reyna (2002). Fuzzy-Trace Theory and false memory. Current Directions in Psychological Science, 11(5).
  7. K. Nader, G. Schafe & J. LeDoux (2000). Fear memories require protein synthesis in the amygdala for reconsolidation after retrieval. Nature, 406.
  8. J. L. McGaugh (2000). Memory — a century of consolidation (amygdala modulation of emotional memory). Science, 287.
  9. M. W. Howard & M. J. Kahana (2002). A distributed representation of temporal context. Journal of Mathematical Psychology, 46.
  10. J. L. McClelland, B. L. McNaughton & R. C. O'Reilly (1995). Why there are complementary learning systems in the hippocampus and neocortex. Psychological Review, 102(3).
  11. L. Nadel & M. Moscovitch (1997). Memory consolidation, retrograde amnesia and the hippocampal complex (multiple-trace theory). Current Opinion in Neurobiology, 7(2).
  12. S. Diekelmann & J. Born (2010). The memory function of sleep (sharp-wave ripples, replay). Nature Reviews Neuroscience, 11.
  13. H. L. Roediger & J. D. Karpicke (2006). Test-enhanced learning: the testing effect and retrieval-induced strengthening. Psychological Science, 17(3).
  14. F. C. Bartlett (1932). Remembering: A Study in Experimental and Social Psychology (schema theory). Cambridge University Press.
  15. G. O. Einstein & M. A. McDaniel (1990). Normal aging and prospective memory. Journal of Experimental Psychology: LMC, 16(4).
  16. B. J. Underwood (1957). Interference and forgetting (proactive interference). Psychological Review, 64(1).

42 prior-art references · 11 mechanisms · a defensive publication, not a patent. Take the map and go further.