{"id":"memformers-gradient-memory","name":"Memformers (gradient memory)","url":null,"page":"https://mrpeppersdev.github.io/agent-infrastructure-landscape/systems/memformers-gradient-memory","family":"memory-research","section":"Recent method papers — theorized, no distinct product","sections":["Recent method papers — theorized, no distinct product"],"tier":4,"last_verified_at":"2026-10-28","lifecycle":{"status":"active","severity":0,"note":null},"taxonomy":{"storage":[{"value":"parametric","primary":true,"reason":null}],"retrieval":[{"value":"attention","primary":true,"reason":null}],"persistence":[{"value":"session","primary":true,"reason":null}],"update":[{"value":"parametric-edit","primary":true,"reason":null}],"unit":[{"value":"weight","primary":true,"reason":null}],"governance":[{"value":"opaque","primary":true,"reason":null}],"conflict":[{"value":"n/a","primary":true,"reason":"not applicable (past-gradient registers; architectural)"}]},"cells":{"type":{"value":"Past-gradient memory registers","citation":"https://arxiv.org/abs/2508.10824","status":"real-data","tier":"T2","source_class":"paper"},"desc":{"value":"Treats past optimisation gradients as memory registers for procedural computation. Distinct from the 2010.06891 \"Memformer\" above. (Survey citation — distinct paper identifier not recovered.)","citation":"https://arxiv.org/abs/2508.10824","status":"real-data","tier":"T2","source_class":"paper"},"claims":{"value":"Memory-augmented transformers can implement linear first-order optimization (gradient descent, conjugate gradient, momentum); MoE adapts learned methods to OOD scenarios. Headline: competitive with classical CGD/Nesterov AGM/momentum on synthetic linear regression (no specific headline metric); baselines: Conjugate Gradient Descent, Nesterov Accelerated Gradient, Momentum Gradient Descent; primary dataset: synthetic random linear regression (d=5, n=20, Gaussian inputs).","citation":"https://arxiv.org/abs/2410.07263","status":"real-data","tier":"T2","source_class":"paper"},"modalities":{"value":"text","citation":"https://arxiv.org/abs/2010.06891","status":"real-data","tier":"T2","source_class":"paper"},"created":{"value":"2025-08","citation":"https://arxiv.org/abs/2508.10824","status":"real-data","tier":"T2","last_verified_at":"2026-05-12","source_class":"paper"},"latest-release":{"value":"not applicable — not OSS","citation":null,"status":"not-applicable","tier":"T3","last_verified_at":"2026-05-06"},"license":{"value":"not applicable — not OSS","citation":null,"status":"not-applicable","tier":"T3"},"gh":{"value":"not applicable — no GitHub repo","citation":null,"status":"not-applicable","tier":"T3","last_verified_at":"2026-05-06"},"mindshare":{"value":"0 cites · ?/mo","citation":"https://api.semanticscholar.org/graph/v1/paper/arXiv:2410.07263","status":"real-data","tier":"T2","last_verified_at":"2026-05-12","source_class":"paper"},"citations":{"value":"7 cites 7/yr","citation":"https://www.semanticscholar.org/paper/d5df3249a372fcd7994fa43fa48a94d322b43161","status":"real-data","tier":"T2","last_verified_at":"2026-05-12","source_class":"paper"},"funding":{"value":"not applicable — not commercial","citation":null,"status":"not-applicable","tier":"T3","last_verified_at":"2026-05-06"},"customers":{"value":"not applicable — not commercial","citation":null,"status":"not-applicable","tier":"T3"},"pricing":{"value":"not applicable — not commercial","citation":null,"status":"not-applicable","tier":"T3"},"compliance":{"value":"not applicable — not commercial","citation":null,"status":"not-applicable","tier":"T3"},"data-handling":{"value":"not applicable — not commercial","citation":null,"status":"not-applicable","tier":"T3"},"deployment":{"value":"not applicable — not a deployable product","citation":null,"status":"not-applicable","tier":"T3"},"hq":{"value":"not applicable — not a company","citation":null,"status":"not-applicable","tier":"T3"},"founders":{"value":"no data — no URL in landscape HTML; searched gradient memory Memformers; no source found","citation":"https://arxiv.org/abs/2010.06891","status":"real-data","tier":"T2","source_class":"paper"},"volume":{"value":"no scale claim — research","citation":"https://arxiv.org/abs/2010.06891","status":"real-data","tier":"T2","source_class":"paper"},"perf":{"value":"not applicable — qualitative-only paper (theory paper showing memory-augmented Transformers can implement entire class of linear first-order methods; empirical results competitive with CGD/Nesterov AGM/momentum but no headline metric)","citation":"https://arxiv.org/abs/2410.07263","status":"not-applicable","tier":"T2","source_class":"paper"},"repro":{"value":"No reproductions found — published 2024-10-08 (19 mo ago); no public code","citation":"https://arxiv.org/abs/2410.07263","status":"real-data","tier":"T2","source_class":"paper"},"code-release":{"value":"depth-floor-reached: arXiv 2410.07263 preprint; no official code repo discovered via WebSearch + arXiv abstract scan + GitHub API","citation":"https://arxiv.org/abs/2410.07263","status":"real-data","tier":"T2","source_class":"paper"},"api-surface":{"value":"not applicable — research paper","citation":null,"status":"not-applicable","tier":"T3"},"latency":{"value":"not applicable — research paper","citation":null,"status":"not-applicable","tier":"T3"},"throughput":{"value":"not applicable — research paper","citation":null,"status":"not-applicable","tier":"T3"},"backend-storage":{"value":"not applicable — research paper","citation":null,"status":"not-applicable","tier":"T3"},"multi-tenancy":{"value":"not applicable — research paper","citation":null,"status":"not-applicable","tier":"T3"},"encryption":{"value":"not applicable — research paper","citation":null,"status":"not-applicable","tier":"T3"},"sso-rbac":{"value":"not applicable — research paper","citation":null,"status":"not-applicable","tier":"T3"},"embedding-model":{"value":"not applicable — research paper","citation":null,"status":"not-applicable","tier":"T3"},"consistency":{"value":"not applicable","citation":null,"status":"not-applicable","tier":"T3"},"versioning":{"value":"snapshots","citation":"https://arxiv.org/abs/2010.06891","status":"real-data","tier":"T2","source_class":"paper"},"tombstoning":{"value":"not specified - paper does not address this dimension","citation":"https://arxiv.org/abs/2010.06891","status":"depth-floor-reached","tier":"T2","source_class":"paper"},"schema-evolution":{"value":"schema locked","citation":"https://arxiv.org/abs/2010.06891","status":"real-data","tier":"T2","source_class":"paper"},"namespace":{"value":"not applicable","citation":null,"status":"not-applicable","tier":"T3"},"contradiction":{"value":"not applicable","citation":null,"status":"not-applicable","tier":"T3"},"forgetting":{"value":"TTL-based","citation":"https://arxiv.org/abs/2010.06891","status":"real-data","tier":"T2","source_class":"paper"},"mcp-support":{"value":"not applicable — research paper, no deployed product","citation":null,"status":"not-applicable","tier":"T3"},"a2a-support":{"value":"not applicable — research paper, no deployed product","citation":null,"status":"not-applicable","tier":"T3"},"otel":{"value":"not applicable — research paper, no deployed product","citation":null,"status":"not-applicable","tier":"T3"},"webhooks":{"value":"not applicable — research paper, no deployed product","citation":null,"status":"not-applicable","tier":"T3"},"import-export":{"value":"not applicable — research paper, no deployed product","citation":null,"status":"not-applicable","tier":"T3"},"integration-count":{"value":"not applicable — wrong section","citation":null,"status":"not-applicable","tier":"T3"},"orchestration":{"value":"not applicable — wrong section","citation":null,"status":"not-applicable","tier":"T3"},"programmatic-control":{"value":"not applicable — wrong section","citation":null,"status":"not-applicable","tier":"T3"},"vendor-benchmarks":{"value":"not applicable — wrong section","citation":null,"status":"not-applicable","tier":"T3","last_verified_at":"2026-05-06"},"pricing-specifics":{"value":"not applicable — wrong section","citation":null,"status":"not-applicable","tier":"T3"},"agent-abstraction":{"value":"not applicable — wrong section","citation":null,"status":"not-applicable","tier":"T3"},"memory-primitives":{"value":"not applicable — wrong section","citation":null,"status":"not-applicable","tier":"T3"},"llm-lock":{"value":"not applicable — wrong section","citation":null,"status":"not-applicable","tier":"T3"},"runtimes":{"value":"not applicable — wrong section","citation":null,"status":"not-applicable","tier":"T3"},"session-handling":{"value":"not applicable — wrong section","citation":null,"status":"not-applicable","tier":"T3"},"validated-verticals":{"value":"not applicable - research paper","citation":null,"status":"not-applicable","tier":"T3"},"time-to-running":{"value":"not applicable - research paper","citation":null,"status":"not-applicable","tier":"T3"},"anti-fit":{"value":"not applicable - research paper","citation":null,"status":"not-applicable","tier":"T3"},"optimised-for":{"value":"not applicable - research paper","citation":null,"status":"not-applicable","tier":"T3"},"adjacent-infrastructure":{"value":"not applicable - research paper","citation":null,"status":"not-applicable","tier":"T3"},"pros":{"value":"Treats past optimisation gradients as memory registers for procedural computation.","citation":"https://arxiv.org/abs/2508.10824","status":"real-data","tier":"T2","source_class":"paper"},"cons":{"value":"Survey citation only; paper identifier not directly recovered; reproducibility limited.","citation":"https://arxiv.org/abs/2508.10824","status":"real-data","tier":"T2","source_class":"paper"},"links":{"value":"via survey 2508.10824","citation":"https://arxiv.org/abs/2508.10824","status":"real-data","tier":"T2","source_class":"paper"},"obs-langsmith":{"value":"","citation":null,"status":"no-data","tier":"T3","last_verified_at":"2026-05-14"},"obs-opentelemetry":{"value":"","citation":null,"status":"no-data","tier":"T3","last_verified_at":"2026-05-14"},"obs-datadog":{"value":"","citation":null,"status":"no-data","tier":"T3","last_verified_at":"2026-05-14"},"obs-helicone":{"value":"","citation":null,"status":"no-data","tier":"T3","last_verified_at":"2026-05-14"},"obs-weave":{"value":"","citation":null,"status":"no-data","tier":"T3","last_verified_at":"2026-05-14"},"obs-langfuse":{"value":"","citation":null,"status":"no-data","tier":"T3","last_verified_at":"2026-05-14"},"obs-arize":{"value":"","citation":null,"status":"no-data","tier":"T3","last_verified_at":"2026-05-14"},"obs-custom":{"value":"","citation":null,"status":"no-data","tier":"T3","last_verified_at":"2026-05-14"},"cost-token-budget":{"value":"","citation":null,"status":"no-data","tier":"T3","last_verified_at":"2026-05-14"},"cost-prompt-caching":{"value":"","citation":null,"status":"no-data","tier":"T3","last_verified_at":"2026-05-14"},"cost-semantic-caching":{"value":"","citation":null,"status":"no-data","tier":"T3","last_verified_at":"2026-05-14"},"cost-batching":{"value":"","citation":null,"status":"no-data","tier":"T3","last_verified_at":"2026-05-14"},"cost-model-routing":{"value":"","citation":null,"status":"no-data","tier":"T3","last_verified_at":"2026-05-14"},"cost-streaming-only":{"value":"","citation":null,"status":"no-data","tier":"T3","last_verified_at":"2026-05-14"},"cost-observability-cost-attribution":{"value":"","citation":null,"status":"no-data","tier":"T3","last_verified_at":"2026-05-14"},"eval-langsmith-evals":{"value":"","citation":null,"status":"no-data","tier":"T3","last_verified_at":"2026-05-14"},"eval-braintrust":{"value":"","citation":null,"status":"no-data","tier":"T3","last_verified_at":"2026-05-14"},"eval-weights-and-biases-agent":{"value":"","citation":null,"status":"no-data","tier":"T3","last_verified_at":"2026-05-14"},"eval-helicone-evals":{"value":"","citation":null,"status":"no-data","tier":"T3","last_verified_at":"2026-05-14"},"eval-custom-test-harness":{"value":"","citation":null,"status":"no-data","tier":"T3","last_verified_at":"2026-05-14"},"eval-human-loop":{"value":"","citation":null,"status":"no-data","tier":"T3","last_verified_at":"2026-05-14"},"eval-production-traffic-replay":{"value":"","citation":null,"status":"no-data","tier":"T3","last_verified_at":"2026-05-14"},"commit-trajectory":{"value":"not applicable — no GitHub repo","citation":null,"status":"not-applicable","tier":"T3","last_verified_at":"2026-05-14"},"citation-trajectory":{"value":"[]","citation":"https://www.semanticscholar.org/paper/d5df3249a372fcd7994fa43fa48a94d322b43161","status":"real-data","tier":"T2","last_verified_at":"2026-05-14","source_class":"paper"},"download-trajectory":{"value":"not applicable — not published as a package","citation":null,"status":"not-applicable","tier":"T3","last_verified_at":"2026-05-14"},"cost-input-usd-per-mtok":{"value":"not applicable — not commercial","citation":null,"status":"not-applicable","tier":"T3"},"cost-output-usd-per-mtok":{"value":"not applicable — not commercial","citation":null,"status":"not-applicable","tier":"T3"},"cost-tier":{"value":"not applicable — not commercial","citation":null,"status":"not-applicable","tier":"T3"},"cost-pricing-model":{"value":"not applicable — not commercial","citation":null,"status":"not-applicable","tier":"T3"},"cost-last-verified":{"value":"not applicable — not commercial","citation":null,"status":"not-applicable","tier":"T3"},"capability-composite-score":{"value":"not applicable — research paper, not a runnable capable system","citation":null,"status":"not-applicable","tier":"T3"},"capability-band":{"value":"not applicable — research paper, not a runnable capable system","citation":null,"status":"not-applicable","tier":"T3"},"capability-benchmark-sources":{"value":"not applicable — research paper, not a runnable capable system","citation":null,"status":"not-applicable","tier":"T3"},"capability-last-verified":{"value":"not applicable — research paper, not a runnable capable system","citation":null,"status":"not-applicable","tier":"T3"},"use-case-tags":{"value":"not applicable — research method paper, no deployable system","citation":null,"status":"not-applicable","tier":"T3"},"use-case-anti-tags":{"value":"not applicable — research method paper, no deployable system","citation":null,"status":"not-applicable","tier":"T3"},"adoption-accelerator":{"value":"","citation":null,"status":"no-data","tier":"T3"},"architecture-deployment-locus":{"value":"","citation":null,"status":"no-data","tier":"T3"},"licensing-bucket":{"value":"not applicable — research paper","citation":null,"status":"not-applicable","tier":"T3"},"positioning-platform-vs-product":{"value":"","citation":null,"status":"no-data","tier":"T3"},"taxonomy-three-layer-rollup":{"value":"agent-infra","citation":"https://github.com/MrPeppersDev/agent-infrastructure-landscape/blob/main/docs/schema/section-to-three-layer-rollup.json","status":"real-data","tier":"T1","source_class":"repo"},"developer-activity-opendigger":{"value":"","citation":null,"status":"no-data","tier":"T3"},"freshness-date":{"value":"2026-10-28","citation":"https://mrpeppersdev.github.io/agent-infrastructure-landscape/systems/memformers-gradient-memory","status":"real-data","tier":"T2","source_class":"catalog"},"compliance-certifications":{"value":"","citation":null,"status":"no-data","tier":"T3"},"sdk-languages":{"value":"","citation":null,"status":"no-data","tier":"T3"},"protocol-support":{"value":"","citation":null,"status":"no-data","tier":"T3"},"maturity":{"value":"","citation":null,"status":"no-data","tier":"T3"}},"corroboration":{"cited_cells":19,"third_party_cells":0},"edges":{"outgoing":[{"source":"memformers-gradient-memory","target":"a-mem--gh-wujiangxu-a-mem","type":"cites","evidence":"S2 isInfluential citation","citation":"https://api.semanticscholar.org/graph/v1/paper/d5df3249a372fcd7994fa43fa48a94d322b43161","isInfluential":true},{"source":"memformers-gradient-memory","target":"ait-associative-transformer--arxiv-2309-12862","type":"cites","evidence":"S2 isInfluential citation","citation":"https://api.semanticscholar.org/graph/v1/paper/d5df3249a372fcd7994fa43fa48a94d322b43161","isInfluential":true},{"source":"memformers-gradient-memory","target":"armt-associative-rmt--arxiv-2407-04841","type":"cites","evidence":"S2 isInfluential citation","citation":"https://api.semanticscholar.org/graph/v1/paper/d5df3249a372fcd7994fa43fa48a94d322b43161","isInfluential":true},{"source":"memformers-gradient-memory","target":"atlas--arxiv-2208-03299","type":"cites","evidence":"S2 isInfluential citation","citation":"https://api.semanticscholar.org/graph/v1/paper/d5df3249a372fcd7994fa43fa48a94d322b43161","isInfluential":true},{"source":"memformers-gradient-memory","target":"bai-lab-memoryos--gh-bai-lab-memoryos","type":"cites","evidence":"S2 isInfluential citation","citation":"https://api.semanticscholar.org/graph/v1/paper/d5df3249a372fcd7994fa43fa48a94d322b43161","isInfluential":true},{"source":"memformers-gradient-memory","target":"compressive-transformer--arxiv-1911-05507","type":"cites","evidence":"S2 isInfluential citation","citation":"https://api.semanticscholar.org/graph/v1/paper/d5df3249a372fcd7994fa43fa48a94d322b43161","isInfluential":true},{"source":"memformers-gradient-memory","target":"differentiable-search-index-dsi--arxiv-2202-06991","type":"cites","evidence":"S2 isInfluential citation","citation":"https://api.semanticscholar.org/graph/v1/paper/d5df3249a372fcd7994fa43fa48a94d322b43161","isInfluential":true},{"source":"memformers-gradient-memory","target":"em-llm--arxiv-2407-09450","type":"cites","evidence":"S2 isInfluential citation","citation":"https://api.semanticscholar.org/graph/v1/paper/d5df3249a372fcd7994fa43fa48a94d322b43161","isInfluential":true},{"source":"memformers-gradient-memory","target":"emat--arxiv-2210-16773","type":"cites","evidence":"S2 isInfluential citation","citation":"https://api.semanticscholar.org/graph/v1/paper/d5df3249a372fcd7994fa43fa48a94d322b43161","isInfluential":true},{"source":"memformers-gradient-memory","target":"lm2--arxiv-2502-06049","type":"cites","evidence":"S2 isInfluential citation","citation":"https://api.semanticscholar.org/graph/v1/paper/d5df3249a372fcd7994fa43fa48a94d322b43161","isInfluential":true},{"source":"memformers-gradient-memory","target":"matter--arxiv-2406-04670","type":"cites","evidence":"S2 isInfluential citation","citation":"https://api.semanticscholar.org/graph/v1/paper/d5df3249a372fcd7994fa43fa48a94d322b43161","isInfluential":true},{"source":"memformers-gradient-memory","target":"membart","type":"cites","evidence":"S2 isInfluential citation","citation":"https://api.semanticscholar.org/graph/v1/paper/d5df3249a372fcd7994fa43fa48a94d322b43161","isInfluential":true},{"source":"memformers-gradient-memory","target":"memformer","type":"cites","evidence":"S2 isInfluential citation","citation":"https://api.semanticscholar.org/graph/v1/paper/d5df3249a372fcd7994fa43fa48a94d322b43161","isInfluential":true},{"source":"memformers-gradient-memory","target":"memgpt-v2-agent-tools--arxiv-2310-08560","type":"cites","evidence":"S2 isInfluential citation","citation":"https://api.semanticscholar.org/graph/v1/paper/d5df3249a372fcd7994fa43fa48a94d322b43161","isInfluential":true},{"source":"memformers-gradient-memory","target":"memlong--arxiv-2408-16039","type":"cites","evidence":"S2 isInfluential citation","citation":"https://api.semanticscholar.org/graph/v1/paper/d5df3249a372fcd7994fa43fa48a94d322b43161","isInfluential":true},{"source":"memformers-gradient-memory","target":"memorizing-transformers--openreview-net","type":"cites","evidence":"S2 isInfluential citation","citation":"https://api.semanticscholar.org/graph/v1/paper/d5df3249a372fcd7994fa43fa48a94d322b43161","isInfluential":true},{"source":"memformers-gradient-memory","target":"memory-layers-at-scale--arxiv-2412-09764","type":"cites","evidence":"S2 isInfluential citation","citation":"https://api.semanticscholar.org/graph/v1/paper/d5df3249a372fcd7994fa43fa48a94d322b43161","isInfluential":true},{"source":"memformers-gradient-memory","target":"namms--arxiv-2410-13166","type":"cites","evidence":"S2 isInfluential citation","citation":"https://api.semanticscholar.org/graph/v1/paper/d5df3249a372fcd7994fa43fa48a94d322b43161","isInfluential":true},{"source":"memformers-gradient-memory","target":"titans-google--research-google","type":"cites","evidence":"S2 isInfluential citation","citation":"https://api.semanticscholar.org/graph/v1/paper/d5df3249a372fcd7994fa43fa48a94d322b43161","isInfluential":true},{"source":"memformers-gradient-memory","target":"transformer-xl--arxiv-1901-02860","type":"cites","evidence":"S2 isInfluential citation","citation":"https://api.semanticscholar.org/graph/v1/paper/d5df3249a372fcd7994fa43fa48a94d322b43161","isInfluential":true},{"source":"memformers-gradient-memory","target":"transformerfam--arxiv-2404-09173","type":"cites","evidence":"S2 isInfluential citation","citation":"https://api.semanticscholar.org/graph/v1/paper/d5df3249a372fcd7994fa43fa48a94d322b43161","isInfluential":true}],"incoming":[]},"history":[],"license":"CC-BY-4.0","source":"https://github.com/MrPeppersDev/agent-infrastructure-landscape/blob/main/data/landscape.json"}