Tabular foundation models suffer from test-time adversarial vulnerabilities that degrade accuracy and enable transferable attacks, but incremental adversarial in-context learning improves robustness on multiple benchmarks.
Why in-context learning transformers are tabular data classifiers
6 Pith papers cite this work. Polarity classification is still indexing.
6
Pith papers citing it
citation-role summary
baseline 1
citation-polarity summary
roles
baseline 1polarities
baseline 1representative citing papers
For tabular in-context learning models, recourse is well-defined, its cost is bounded and converges to classical linear recourse as context grows; ASR-ICL finds sparse recourse with fewer queries.
PS-PFN extends posterior sampling to the max k-armed bandit setup using PFNs for in-context posterior estimation of maximal pipeline performance, outperforming other bandit and AutoML strategies on benchmarks.
TabICL scales in-context learning to large tabular data via column-then-row attention for row embeddings followed by a transformer, matching TabPFNv2 speed and performance while outperforming it and CatBoost on datasets over 10K samples.
The synthetic prior for tabular foundation models covers only a narrow part of real table distributions, but this mismatch does not degrade model generalization.
TabSwift introduces an efficient row-wise attention tabular foundation model competitive with TabPFN v2 and TabICL via gated attention stabilization and register tokens, plus adaptive layer-wise early-exit for inference.
citing papers explorer
-
On the Robustness of Tabular Foundation Models: Test-Time Attacks and In-Context Defenses
Tabular foundation models suffer from test-time adversarial vulnerabilities that degrade accuracy and enable transferable attacks, but incremental adversarial in-context learning improves robustness on multiple benchmarks.
-
Algorithmic Recourse of In-Context Learning for Tabular Data
For tabular in-context learning models, recourse is well-defined, its cost is bounded and converges to classical linear recourse as context grows; ASR-ICL finds sparse recourse with fewer queries.
-
In-Context Decision Making for Optimizing Complex AutoML Pipelines
PS-PFN extends posterior sampling to the max k-armed bandit setup using PFNs for in-context posterior estimation of maximal pipeline performance, outperforming other bandit and AutoML strategies on benchmarks.
-
TabICL: A Tabular Foundation Model for In-Context Learning on Large Data
TabICL scales in-context learning to large tabular data via column-then-row attention for row embeddings followed by a transformer, matching TabPFNv2 speed and performance while outperforming it and CatBoost on datasets over 10K samples.
-
Mind the Gap? A Distributional Comparison of Real and Synthetic Priors for Tabular Foundation Models
The synthetic prior for tabular foundation models covers only a narrow part of real table distributions, but this mismatch does not degrade model generalization.
-
TabSwift: An Efficient Tabular Foundation Model with Row-Wise Attention
TabSwift introduces an efficient row-wise attention tabular foundation model competitive with TabPFN v2 and TabICL via gated attention stabilization and register tokens, plus adaptive layer-wise early-exit for inference.