[1909.06312] Neural Oblivious Decision Ensembles for Deep Learning on Tabular Dataopen searchopen navigation menucontact arXivsubscribe to arXiv mailings

Nowadays, deep neural networks (DNNs) have become the main instrument for machine learning tasks within a wide range of domains, including vision, NLP, and speech. Meanwhile, in an important case of heterogenous tabular data, the advantage of DNNs over shallow counterparts remains questionable. In particular, there is no sufficient evidence that deep learning machinery allows constructing methods that outperform gradient boosting decision trees (GBDT), which are often the top choice for tabular problems. In this paper, we introduce Neural Oblivious Decision Ensembles (NODE), a new deep learning architecture, designed to work with any tabular data. In a nutshell, the proposed NODE architecture generalizes ensembles of oblivious decision trees, but benefits from both end-to-end gradient-based optimization and the power of multi-layer hierarchical representation learning. With an extensive experimental comparison to the leading GBDT packages on a large number of tabular datasets, we demon

1 mentions: @ml_review
Keywords: deep learning
Date: 2020/12/05 08:21

Referring Tweets

@ml_review NODE for Deep Learning on Tabular Data #ICLR2020 By @YandexAI – DNN architecture that substantially outperforms leading GBDT on tabular data – Generalizes CatBoost, making decision tree routing differentiable ArXiv t.co/CEgce5kXV6 PyTorch t.co/gq0hIkkf1y t.co/vvX3F1cDMv

Related Entries

Read more Weights & Biases
0 users, 2 mentions 2020/04/30 00:51
Read more 論文紹介:分布マッチングの観点からみた敵対的不変表現学習の安定化(IJCAI2020) | 東京大学松尾研究室 - Matsuo Lab
0 users, 2 mentions 2020/10/14 03:52
Read more [2010.14742] ElderSim: A Synthetic Data Generation Platform for Human Action Recognition in Eldercar...
0 users, 1 mentions 2020/12/20 08:21
Read more GitHub - leaderj1001/BottleneckTransformers: Bottleneck Transformers for Visual Recognition
0 users, 1 mentions 2021/02/05 15:51
Read more [2010.05273] Federated Learning via Posterior Averaging: A New Perspective and Practical Algorithms
0 users, 1 mentions 2021/02/22 21:51