Description
SAINT is a hybrid deep learning approach to solving tabular data problems. SAINT performs attention over both rows and columns, and it includes an enhanced embedding method. The architecture, pre-training and training pipeline are as follows:
- layers with 2 attention blocks each, one self-attention block, and a novel intersample attention blocks that computes attention across samples are used.
- For pre-training, this involves minimizing the contrastive and denoising losses between a given data point and its views generated by CutMix and mixup. During finetuning/regular training, data passes through an embedding layer and then the SAINT model. Lastly, the contextual embeddings from SAINT are used to pass only the embedding corresponding to the CLS token through an MLP to obtain the final prediction.
Papers Using This Method
SAINT: Attention-Based Modeling of Sub-Action Dependencies in Multi-Action Policies2025-05-17Similarity-Aware Token Pruning: Your VLM but Faster2025-03-14ASCenD-BDS: Adaptable, Stochastic and Context-aware framework for Detection of Bias, Discrimination and Stereotyping2025-02-04Computational Analysis of Yaredawi YeZema Silt in Ethiopian Orthodox Tewahedo Church Chants2024-12-25A Survey on Deep Tabular Learning2024-10-15Semi-Quantitative Analysis and Seroepidemiological Evidence of Past Dengue Virus Infection among HIV-infected patients in Onitsha, Anambra State, Nigeria2024-03-23Tabular Machine Learning Methods for Predicting Gas Turbine Emissions2023-07-17Challenges and Opportunities in Information Manipulation Detection: An Examination of Wartime Russian Media2022-05-24SAINT: Improved Neural Networks for Tabular Data via Row Attention and Contrastive Pre-Training2021-06-02