Research map / Probabilistic and scientific modeling / Molecules and proteins
Identifying Learnwares via Reduced Neural Conditional Mean Embedding
Zi-Yu Mao, Ming Li
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 |
|---|---|---|---|
| train the neural feature map instead of leaving it untrained (ablation) | RNCME · Synthetic complex conditional distribution | MMD: 1.37 -> 0.87 | Improved |
| use the finite-dimensional neural feature map instead of an infinite-dimensional RBF kernel for CME | CME · NICO++ compression stage (n=40,000) | time (s): > 10,000 -> 0.516 | Improved |
| increase reduced set size m (ablation) | RNCME · California Housing Prices | NDCG@1: converges beyond m of about 800, matching or surpassing unreduced 0.9652 | Improved |
| increase reduced set size m (ablation) | RNCME · California Housing Prices | generation time | Hurt |
| model output conditional distribution (RNCME) instead of input-only RKME | RKME specification · California Housing Prices (regression) | NDCG@k | Improved |
| use RNCME instead of RKME, HF, cRKME and LP | learnware specification · NICO++ (classification) | NDCG@k and EA@k: k=1: best learnware for 16/20 tasks, top-3 for all | Improved |
Closest papers by meaning
- Learnware Specification via Label-Aware Neural Embedding (AAAI 2025 · similarity 0.92)
- On the Ability of Developers' Training Data Preservation of Learnware (NeurIPS 2024 · similarity 0.91)
- Identifying Useful Learnwares for Heterogeneous Label Spaces (ICML 2023 · similarity 0.89)
- Neural-Kernel Conditional Mean Embeddings (ICML 2024 · similarity 0.89)
- Handling Learnwares from Heterogeneous Feature Spaces with Explicit Label Exploitation (NeurIPS 2024 · similarity 0.85)
- Breaking the Batch Barrier (B3) of Contrastive Learning via Smart Batch Mining (NeurIPS 2025 · similarity 0.84)
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.
