SERS

https://openreview.net/forum?id=jR1lvwexLt

Dynamic regulariser driven by Wasserstein distance between task distributions; automatically relaxes or strengthens constraints in proportion to task similarity.

At a glance

Type
Self-evolving pseudo-rehearsal
Tier
T4
Created
2025-09-18 (submitted to NeurIPS 2025 Sept 18 2025; Jun Wang / Liang Ding et al.)
Latest release
not applicable — not OSS
License
not applicable — not OSS
GitHub
not applicable — no GitHub repo
Pricing
not applicable — not commercial
Funding
not applicable — not commercial

Taxonomy

storage
parametric
retrieval
parametric-recall
persistence
parametric-permanent
update
parametric-edit
unit
episode
governance
inspectable
conflict
llm-arbitrate

When to use

Optimised for: not applicable - research paper

Anti-fit: not applicable - research paper

Pros & cons

Pros

Self-evolving rehearsal-scheduling for continual learning.

Cons

Preprint; eval scope limited.

Claims & capabilities

Lightweight continual-learning framework with three mechanisms: pseudo-input synthesis (decouples prompt and label generation via semantic masking and templates), label self-evolution (blends base and fine-tuned outputs to prevent over-specialization), and dynamic regularization using Wasserstein distance between task distributions; reduces forgetting by over 2 percentage points vs strong pseudo-rehearsal baselines; NeurIPS 2025 poster

Technical surface

API surface
not applicable — research paper
Backend storage
not applicable — research paper
Deployment
not applicable — not a deployable product
Embedding model
not applicable — research paper
Multi-tenancy
not applicable — research paper
MCP
not applicable — research paper, no deployed product
A2A
not applicable — research paper, no deployed product
OpenTelemetry
not applicable — research paper, no deployed product

Similar systems

Other recent method papers — theorized, no distinct product in the catalog, ranked by inbound references.

  • Compressive Transformer T3

    Maintains recent states in full resolution while compressing older memories with learned compression functions. DeepMind.

  • MemGPT v2 / agent-tools T3

    Already in catalog as the foundational MemGPT paper. Note: Letta is the productionised successor (cross-listed).

  • Transformer-XL T3

    Extends context through segment-level recurrence + caching of hidden states from prior segments. Foundational long-context architecture.

  • Generative Agents T3

    Park et al. — landmark agent-simulation paper. Reflection + memory stream + retrieval enable believable agent behavior.

  • MemoryBank T3

    Enhances LLMs with long-term memory. Early influential paper.

  • Reflexion T3

    Language agents with verbal reinforcement learning.

Row last verified 2026-05-14. Catalog data is CC-BY-4.0 — see how to read this.