A benchmark of 17 methods on 68 tabular datasets found deep‑learning models outperformed gradient‑boosted trees. It compared deep‑learning, gradient‑boosted trees and AutoML. Read more: https://getnews.me/deep-learning-surpasses-traditional-models-on-tabular-data/ #deeplearning #tabulardata

RuleNet, a transformer for tabular data, was tested on eight benchmarks and matched or beat leading tree‑based methods while staying computationally efficient. Read more: https://getnews.me/rulenet-transformer-advances-deep-learning-for-tabular-data/ #tabulardata #transformer

ReFine improves synthetic tabular data generation in low‑data settings, achieving 0.44 R‑squared gain for regression and a 10% relative boost in F1 for classification. Read more: https://getnews.me/refine-boosts-synthetic-tabular-data-generation-in-low-data-settings/ #syntheticdata #tabulardata

One more time (and now with a working link) 😅

Several PhD positions in two brand-new labs at the Lamarr Intitute / TU Dortmund.

Lots of exciting research on #tabulardata, #autom and #benchmarking.

👉 Apply now, there's less than a week left!