Research map / Continual learning / Online and replay-based
Lifelong Infinite Mixture Model Based on Knowledge-Driven Dirichlet Process
Fei Ye, Adrian G. Bors
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (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 LIMix mixture instead of single-model LGM | LGM · MSFIR unsupervised (avg) | MSE: 190.38 -> 28.97 | Improved |
| use LIMix instead of BatchEnsemble | BatchEnsemble · MSFIR unsupervised (avg) | MSE: 31.91 -> 28.97 | Improved |
| use LIMix instead of BatchEnsemble | BatchEnsemble · MSFIR unsupervised (avg) | PSNR: 19.31 -> 19.86 | Improved |
| use full LIMix instead of compact Student (ablation) | Student · MSFIR unsupervised (avg) | MSE: 163.45 -> 28.97 | Improved |
| use LIMix instead of BatchEnsemble | BatchEnsemble · MSFIRC classification (avg) | accuracy: 83.47 -> 84.18 | Improved |
| use LIMix instead of BatchEnsemble | BatchEnsemble · MSFIRC classification on MNIST | accuracy: 99.40 -> 91.16 | Hurt |
| use LIMix instead of BatchEnsemble | BatchEnsemble · MSFIRC classification on SVHN | accuracy: 74.46 -> 82.60 | Improved |
| use LIMix instead of CURL | CURL · MSFIRC classification on CIFAR10 | accuracy: 67.34 -> 54.66 | Hurt |
| use LIMix instead of FROMP | FROMP · Permuted MNIST | accuracy: 94.9 -> 96.46 | Improved |
| use LIMix instead of FROMP | FROMP · Split MNIST | accuracy: 99.0 -> 99.21 | Improved |
| reduce LIMix components from 10 to 7 (ablation) | LIMix · Permuted MNIST | accuracy: 96.46 -> 88.78 | Hurt |
| reduce LIMix components from 5 to 3 (ablation) | LIMix · Split MNIST | accuracy: 99.21 -> 91.37 | Hurt |
Closest papers by meaning
- Lifelong Compression Mixture Model via Knowledge Relationship Graph (AAAI 2023 · similarity 0.89)
- Dynamic Expansion Diffusion Learning for Lifelong Generative Modelling (AAAI 2025 · similarity 0.86)
- Scalable Language Model with Generalized Continual Learning (ICLR 2024 · similarity 0.85)
- Online Task-Free Continual Learning via Dynamic Expansionable Memory Distribution (CVPR 2025 · similarity 0.85)
- A Unified and General Framework for Continual Learning (ICLR 2024 · similarity 0.85)
- Learn more, but bother less: parameter efficient continual learning (NeurIPS 2024 · 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.
