Research map / Label-efficient and robust learning / Recognition and long-tail
End-to-End Deep Learning for Predicting Metric Space-Valued Outputs
Yidong Zhou, Su I Iao, Hans-Georg Müller
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 E2M instead of deep Frechet regression | E2M · Simulation, distribution outputs (n=500) | MSPE: 0.869 -> 0.562 | Improved |
| use E2M instead of deep Frechet regression | E2M · Simulation, network outputs (n=2000) | MSPE: 3.018 -> 1.729 | Improved |
| use E2M instead of RFWLFR | E2M · Simulation, SPD outputs power metric (n=2000) | MSPE: 0.637 -> 0.187 | Improved |
| increase sample size from 500 to 2000 | E2M · Simulation, distribution outputs | MSPE: 0.562 -> 0.218 | Improved |
| use anchor-based strategy with n=10,000 | E2M · Simulation, distribution outputs | MSPE: 0.218 -> 0.088 | Improved |
| use E2M instead of GFR with anchors at n=10,000 | E2M · Simulation, network outputs (n=10000) | MSPE: 9.416 -> 1.072 | Improved |
| use E2M instead of DFR | E2M · Human mortality | MSPE: 26.75 -> 22.64 | Improved |
| use E2M instead of DFR | E2M · Taxi network | MSPE: 7.93 -> 6.83 | Improved |
| use E2M instead of GFR when the true relationship is linear | E2M · Simulation, linear distribution outputs (n=2000) | MSPE: 0.188 -> 0.015 | Hurt |
| use E2M instead of IFR at small sample size | E2M · Simulation, distribution outputs (n=50) | MSPE: 1.151 -> 1.103 | Hurt |
| add mild negative entropy regularization lambda=-0.01 instead of 0 (ablation) | E2M · Simulation, distribution outputs (n=2000) | MSPE: 0.444 -> 0.291 | Improved |
| use strong positive entropy regularization lambda=0.1 instead of 0 (ablation) | E2M · Simulation, distribution outputs (n=2000) | MSPE: 0.444 -> 1.033 | Hurt |
Closest papers by meaning
- Deep Divergence Learning (ICML 2020 · similarity 0.84)
- Fréchet Geodesic Boosting (NeurIPS 2025 · similarity 0.84)
- Improving Deep Metric Learning by Divide and Conquer (TPAMI 2021 · similarity 0.84)
- Deep Single-Index Fréchet Regression (ICML 2026 · similarity 0.84)
- Learning Distances from Data with Normalizing Flows and Score Matching (ICML 2025 · similarity 0.83)
- RMLR: Extending Multinomial Logistic Regression into General Geometries (NeurIPS 2024 · similarity 0.83)
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.
