The paper argues that typical tabular machine learning benchmarks, which aggregate results by averaging scores or ranks, can hide which models are essential for achieving the best performance on specific datasets. It proposes evaluating models against a data‑centric peak performance frontier, classifying them as irreplaceable, sufficient, redundant, or fallible based on their position relative to other models. Applying this to the TabArena benchmark shows that common aggregation metrics mainly capture consistency and failure avoidance, but fail to reflect dataset‑specific strengths, leading to a misalignment between aggregate rewards and true model utility.
By Andrej Tschalzev, Stefan L\"udtke, Heiner Stuckenschmidt, Christian Bartelt
arXiv:2606. 27997v1 Announce Type: new Abstract: Benchmarks of machine learning models often include many datasets, making evaluation expensive.
By Rostislav Gusev, Alexey Zaytsev
arXiv:2605. 17273v3 Announce Type: replace-cross Abstract: State-of-the-Art (SOTA) claims pervade Artificial Intelligence (AI) and Machine Learning (ML) research.
By YongKyung Oh
arXiv:2606. 30410v1 Announce Type: cross Abstract: Foundation models for predictive machine learning on tabular data have recently gained significant traction in academia and industry.
By Lennart Purucker, Andrej Tschalzev, Nick Erickson, Gioia Blayer, David Holzm\"uller, Alan Arazi, Alexander Pfefferle, Mustafa Tajjar, Ga\"el Varoquaux, Frank Hutter
arXiv:2606. 02384v1 Announce Type: new Abstract: Progress in tabular machine learning has largely focused on increasingly sophisticated model architectures.
By Andrej Tschalzev, Nick Erickson, Yuyang Wang, Huzefa Rangwala, Stefan L\"udtke, Heiner Stuckenschmidt, Christian Bartelt
arXiv:2605. 23595v2 Announce Type: replace-cross Abstract: The rapid advancement of machine learning has led to an unprecedented expansion of model ecosystems, making it increasingly difficult to assess the reliability of newly released models on unseen and unlabeled data.
By Trinh Pham, Viet Huynh, Hongzhi Yin, Quoc Viet Hung Nguyen, Thanh Tam Nguyen
The paper discusses how to quantify statistical uncertainty for aggregate performance metrics in machine learning benchmarks, focusing on methods such as bootstrapping, Bayesian hierarchical modeling, and visualizing task weightings with standard errors. It demonstrates that these techniques can uncover insights—for example, revealing that a model may dominate specific task types even if its overall performance is poor. The authors apply their approach to the Visual Task Adaptation Benchmark (VTAB) to illustrate its practical usefulness.
By Rachel Longjohn, Giri Gopalan, Emily Casleton
arXiv:2605. 28418v3 Announce Type: replace Abstract: With the rise of tabular foundation models alongside traditional models still performing well on many tasks, choosing the right model for a tabular dataset remains difficult.
By Markus Herre, Andrej Tschalzev, Sascha Marton, Christian Bartelt
The paper introduces Balance of Benchmarks (BoB), a framework that improves task-conditioned model comparison by weighting benchmark evidence based on semantic density, equating scores across varying difficulty levels, and pooling task-relevant residuals. BoB retains all eligible benchmark data while adjusting its influence, outperforming uniform averaging on the WildScores dataset with higher Spearman correlation, lower MAE, and better shortlist hit rates. The method also reduces ranking instability when benchmarks are repeated or paraphrased, and lowers retrospective regret in model selection.
By Jhen-Ke Lin, Hong-Yun Lin
arXiv:2609.37989v1 Announce Type: new
Abstract: Tabular foundation models achieve strong zero-shot accuracy on structured data by pretraining on synthetic tables, but they ignore the column names, ta...
By Deqing Fu, Huangyuan Su, Rajat Sen, Taman Narayan, Sujay Sanghavi, Abhimanyu Das, Weihao Kong
arXiv:2506. 07673v2 Announce Type: replace Abstract: Large language model (LLM) evaluation is increasingly costly, prompting interest in methods that speed up evaluation by shrinking benchmark datasets.
By Guanhua Zhang, Florian E. Dorner, Moritz Hardt
The paper examines how the choice of anchor model in LLM-as-a-judge evaluations affects reliability. By testing 22 anchors on the Arena-Hard-v2.0 dataset, it shows that extreme anchors (best or worst performers) are poor choices, reducing correlation with human rankings. The study quantifies the anchor effect size, compares it to judge model selection, and offers guidelines and power‑analysis recommendations for more reliable benchmark design.
By Shachar Don-Yehiya, Asaf Yehudai, Leshem Choshen, Omri Abend