Forty years of flattening
Where tabular prediction came from, and what changed recently.
1 Aug 2026 · History · 8 min
Tabular prediction has been a solved-enough problem for a decade, so it is worth asking why anything new would appear now. The answer is not a better model. It is a different assumption about the shape of the input.
The tree era
XGBoost's own paper reports the number that made its reputation: of 29 winning Kaggle solutions published in 2015, 17 used XGBoost. The second most popular method, deep neural networks, appeared in 11. That result set the default for the decade that followed.
The flattening tax
By 2017 the flattening step had become infrastructure. Uber's Michelangelo post popularised the feature store — a central place to publish canonical features — and Feast and Tecton turned it into a category. The step nobody liked had acquired a budget, a team and a roadmap.
Deep learning's false start
Deep tabular models arrived on schedule — TabNet in 2019, TabTransformer in 2020, SAINT in 2021 — and each claimed to beat the trees. Then two papers checked. Both found that gradient boosting still won, with far less tuning, on the challengers' own datasets.
Learning without fitting
TabPFN in 2022 changed the question. Instead of fitting a model to your table, it was trained offline on millions of synthetic datasets and then read your table as context — no gradient steps, no tuning, under a second. The 2025 Nature version scaled it; TabICL pushed the context to hundreds of thousands of rows.
Someone kept the joins
The other lineage was quieter and older. Multi-relational data mining framed the problem in 2003. In 2020 a paper made the obvious move explicitly — treat the primary-foreign-key structure as a graph and run a GNN over it — and positioned it against automated feature engineering. It beat DFS on two of three datasets, which is a real result honestly reported.
What actually changed
The Relational Transformer changed the unit. Not the row, not the table — the cell, tagged with its column, table, row and timestamp, with attention masked to follow foreign keys. There is no flattening step because there is nothing to flatten into.
What this buys you
What disappears is a job. No feature table to design, no pipeline to maintain, no model to retrain per question — and no feature store standing between a question and its answer. A first result takes minutes because there is no training run in front of it, and the same checkpoint answers the next question without being asked to learn anything.
What this does not settle
The Relational Transformer is still relatively unoptimised, yet it can already beat XGBoost when the relationships carry signal. It does not win everywhere today, particularly on a single wide table, but that looks more like a starting point than a ceiling. If your data genuinely is a rectangle, forty years of work on rectangles is available to you.
What happens next
The architecture is the part to watch. LLMs keep getting better through changes to attention, retrieval, memory and efficient inference, with an enormous concentration of research and compute behind them. Relational Transformers can inherit that progress while evolving the parts specific to databases — their masks, samplers and objectives. The current model is a first point on that curve, not its final form.
Sources
- 1984Classification and Regression TreesBreiman, Friedman, Olshen & Stone
- 1991Inductive Logic ProgrammingMuggleton
- 2001Random ForestsBreiman
- 2001Greedy Function Approximation: A Gradient Boosting MachineFriedman
- 2003Multi-Relational Data Mining: An IntroductionDžeroski
- 2015Deep Feature SynthesisKanter & Veeramachaneni
- 2016XGBoost: A Scalable Tree Boosting SystemChen & Guestrin
- 2016Entity Embeddings of Categorical VariablesGuo & Berkhahn
- 2017LightGBMKe et al.
- 2017CatBoostProkhorenkova et al.
- 2017MichelangeloHermann & Del Balso, Uber
- 2019TabNetArik & Pfister
- 2020TabTransformerHuang et al.
- 2020Supervised Learning on Relational Databases with GNNsCvitkovic
- 2021SAINTSomepalli et al.
- 2021Tabular Data: Deep Learning is Not All You NeedShwartz-Ziv & Armon
- 2021Revisiting Deep Learning Models for Tabular Data (FT-Transformer)Gorishniy et al.
- 2022TabPFNHollmann et al.
- 2022Why do tree-based models still outperform deep learning on tabular data?Grinsztajn, Oyallon & Varoquaux
- 2023Relational Deep LearningFey et al.
- 2024RelBenchRobinson et al.
- 2025Accurate predictions on small data with a tabular foundation model (TabPFN v2)Hollmann et al.
- 2025TabICLQu et al.
- 2025Relational TransformerRanjan et al.
- 2026PluRelKothapalli et al.