Research map / Continual learning / Online and replay-based
Labeled Memory Networks for Online Model Adaptation
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| Use adaptive RNN-based weighting instead of a fixed PCN-memory weight (ablation) | LMN · Brightkite, FSQNYC, FSQTokyo, and Yoochoose | log perplexity: 3.88 -> 3.57; 6.13 -> 5.54; 5.40 -> 5.32; 5.01 -> 4.96 | Helped |
| Use adaptive RNN-based weighting instead of a fixed PCN-memory weight (ablation) | LMN · Brightkite, FSQNYC, FSQTokyo, Geolife, and Yoochoose | MRR: (0.51) -> (0.55); (0.25) -> (0.27); (0.26) -> (0.28); (0.83) -> (0.83); (0.24) -> (0.25) | Mixed |
| Add labeled memory with online updates for examples with non-zero loss | LMN · Omniglot online adaptation | overall accuracy: 56% -> 86% | Helped |
| Add labeled memory with label-specific storage and retrieval | LMN · Omniglot online adaptation | seen-label accuracy: 59% -> 92% | Helped |
| Use label-organized memory rather than the compared memory organization | LMN · Omniglot 4k-way 1-shot | accuracy: 48.2% -> 52.6% | Helped |
| Use label-organized memory rather than the compared memory organization | LMN · Omniglot 4k-way 2-shot | accuracy: 58.0% -> 62.5% | Helped |
| Use label-organized memory rather than the compared memory organization | LMN · Omniglot 4k-way 3-shot | accuracy: 60.3% -> 64.1% | Helped |
| Adapt with LMN memory rather than parameter updates that may alter the shared representation | Omniglot online adaptation · new labels | accuracy: 49% -> 71% | Helped |
| Use online memory adaptation rather than the neural cache | LMN · Text8 (2000) | perplexity: 99.9 -> 91.1 | Helped |
| Use online memory adaptation rather than Pointer LSTM | LMN · Wikitext2 (100) | perplexity: 80.8 -> 77.6 | Helped |
Closest papers by meaning
- Memory-based Parameter Adaptation (ICLR 2018 · similarity 0.89)
- Online Meta-Learning via Learning with Layer-Distributed Memory (NeurIPS 2021 · similarity 0.88)
- Dual Memory Networks: A Versatile Adaptation Approach for Vision-Language Models (CVPR 2024 · similarity 0.88)
- On the Duality between Gradient Transformations and Adapters (ICML 2025 · similarity 0.86)
- Online Adaptation of Language Models with a Memory of Amortized Contexts (NeurIPS 2024 · similarity 0.86)
- Memory Decoder: A Pretrained, Plug-and-Play Memory for Large Language Models (NeurIPS 2025 · similarity 0.85)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.