Research map / Vision-language and multimodal / Cross-modal learning
No Data? No Problem: Robust Vision-Tabular Learning with Missing Values
Marta Hasny, Laura Alexandra Daza, Keno Bressem, Maxime Di Folco, Julia Schnabel
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 |
|---|---|---|---|
| add missingness augmentation in downstream tuning (ablation) | RoVTL (trainable) · UK Biobank BMI | MAE: 4.53 -> 1.69 | Improved |
| add missingness augmentation in downstream tuning (ablation) | RoVTL (frozen) · UK Biobank CAD | AUC: 77.07 -> 84.68 | Improved |
| add TabMoFe loss (ablation) | RoVTL (trainable) · UK Biobank CAD | AUC: 83.09 -> 84.91 | Improved |
| add TabMoFe loss (ablation) | RoVTL (trainable) · UK Biobank BMI | MAE: 1.98 -> 1.69 | Improved |
| add gate to cross-attention (ablation) | RoVTL (trainable) · UK Biobank CAD | AUC: 82.13 -> 84.91 | Improved |
| add gate to cross-attention (ablation) | RoVTL (trainable) · UK Biobank BMI | MAE: 1.87 -> 1.69 | Improved |
| add gate to cross-attention (ablation) | RoVTL (frozen) · UK Biobank CAD | AUC: 84.60 -> 84.68 | No clear effect |
| add contrastive pretraining (ablation) | RoVTL (trainable) · UK Biobank CAD | AUC: 82.94 -> 84.91 | Improved |
| add contrastive pretraining (ablation) | RoVTL (trainable) · M&M's disease classification | AUC: 66.50 -> 77.56 | Improved |
| use missingness augmentation instead of no augmentation in pretraining (ablation) | RoVTL (trainable) · UK Biobank CAD | AUC: 83.44 -> 84.91 | Improved |
| use missingness instead of marginal-sampling corruption (ablation) | RoVTL (trainable) · UK Biobank CAD | AUC: 83.59 -> 84.91 | Improved |
| use missingness instead of Gaussian perturbation (ablation) | RoVTL (trainable) · UK Biobank CAD | AUC: 78.64 -> 84.91 | Improved |
Closest papers by meaning
- Tabular Insights, Visual Impacts: Transferring Expertise from Tables to Images (ICML 2024 · similarity 0.84)
- ICE-T: Interactions-aware Cross-column Contrastive Embedding for Heterogeneous Tabular Datasets (AAAI 2025 · similarity 0.83)
- TRL-Bench: Standardizing Cross-Paradigm Representation-Level Evaluation of Tabular Encoders (NeurIPS 2026 · similarity 0.83)
- Large Scale Transfer Learning for Tabular Data via Language Modeling (NeurIPS 2024 · similarity 0.83)
- BiSHop: Bi-Directional Cellular Learning for Tabular Data with Generalized Sparse Modern Hopfield Model (ICML 2024 · similarity 0.82)
- TabularBench: Benchmarking Adversarial Robustness for Tabular Deep Learning in Real-world Use-cases (NeurIPS 2024 · similarity 0.82)
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.
