2 papers
cs.LG2026
Valid Feature-Level Inference for Tabular Foundation Models via the Conditional Randomization Test
Mohamed Salem
Modern machine learning models are highly expressive but notoriously difficult to analyze statistically. In particular, while black-box predictors can achieve strong empirical perf…
stat.ML2025
Likelihood-guided Regularization in Attention Based Models
Mohamed Salem, Inyoung Kim
The transformer architecture has demonstrated strong performance in classification tasks involving structured and high-dimensional data. However, its success often hinges on large-…