AI Paper Atlas

Research map / Vision-language and multimodal / Human and embodied modeling

Benchmarking Multimodal AutoML for Tabular Data with Text Fields

Xingjian Shi, Jonas Mueller, Nick Erickson, Mu Li, Alex Smola

NeurIPS 2021 · Vision-language and multimodal › Human and embodied modeling · approach: Multimodal and affective interaction modeling · 3 citations (OpenAlex)

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 changedWhereMeasuredEffect
switch text backbone from RoBERTa to ELECTRA (ablation)Text-Net · 18-dataset multimodal benchmark, averageavg score: 0.572 -> 0.605Improved
add exponential layer-wise learning rate decay (tau=0.8) (ablation)ELECTRA · 18-dataset multimodal benchmark, averageavg score: 0.605 -> 0.632Improved
add averaging of top-3 checkpoints (ablation)ELECTRA + decay · 18-dataset multimodal benchmark, averageavg score: 0.632 -> 0.635Improved
Fuse-Late instead of Text-Net (add numeric/categorical inputs) (ablation)Multimodal-Net · 18-dataset multimodal benchmark, averageavg score: 0.635 -> 0.664Improved
Fuse-Late instead of Fuse-Early (ablation)Multimodal-Net · 18-dataset multimodal benchmark, averageavg score: 0.658 -> 0.664Improved
Fuse-Late instead of All-Text (ablation)Multimodal-Net · 18-dataset multimodal benchmark, averageavg score: 0.662 -> 0.664Improved
stack ensemble Multimodal-Net with tabular models instead of weighted ensemble (ablation)Multimodal-Net · 18-dataset multimodal benchmark, averageavg score: 0.676 -> 0.684Improved
use fine-tuned text embeddings instead of pretrained (Text-Embedding vs Pre-Embedding) (ablation)tabular models · 18-dataset multimodal benchmark, averageavg score: 0.593 -> 0.638Improved
contextualize text embeddings with numeric/categorical info (Multimodal-Embedding vs Text-Embedding) (ablation)tabular models · 18-dataset multimodal benchmark, averageavg score: 0.638 -> 0.644Improved
ensemble Multimodal-Net with tabular models instead of using Transformer as embedding (ablation)Stack-Ensemble · 18-dataset multimodal benchmark, averageavg score: 0.644 -> 0.684Improved
use text features instead of ignoring themStack-Ensemble vs AG-Stack · mercariscore: 0.098 vs 0.605Improved
use text features instead of ignoring themStack-Ensemble vs AG-Stack · kickscore: 0.670 vs 0.797Improved

Closest papers by meaning

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.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy