Robust representations of oil wells' intervals via sparse attention mechanism

TL;DR

This paper introduces Regularized Transformers (Reguformers), an efficient and robust attention-based model for multivariate time series data, specifically applied to oil well logs, outperforming existing models in classification and representation tasks.

Contribution

The paper proposes a new class of efficient, regularized Transformer models tailored for industrial time series, enhancing robustness to noise and missing data while reducing computational complexity.

Findings

Reguformers outperform RNNs, Transformer, Informer, and Performer in well-interval classification.

They demonstrate high accuracy with a PR~AUC score of 0.983 on well-interval similarity.

Models show significant robustness to missing and noisy data.

Abstract

Transformer-based neural network architectures achieve state-of-the-art results in different domains, from natural language processing (NLP) to computer vision (CV). The key idea of Transformers, the attention mechanism, has already led to significant breakthroughs in many areas. The attention has found their implementation for time series data as well. However, due to the quadratic complexity of the attention calculation regarding input sequence length, the application of Transformers is limited by high resource demands. Moreover, their modifications for industrial time series need to be robust to missing or noised values, which complicates the expansion of the horizon of their application. To cope with these issues, we introduce the class of efficient Transformers named Regularized Transformers (Reguformers). We implement the regularization technique inspired by the dropout ideas to improve robustness and reduce computational expenses. The focus in our experiments is on oil&gas data, namely, well logs, a prominent example of multivariate time series. The goal is to solve the problems of similarity and representation learning for them. To evaluate our models for such problems, we work with an industry-scale open dataset consisting of well logs of more than 20 wells. The experiments show that all variations of Reguformers outperform the previously developed RNNs, classical Transformer model, and robust modifications of it like Informer and Performer in terms of well-intervals' classification and the quality of the obtained well-intervals' representations. Moreover, the sustainability to missing and incorrect data in our models exceeds that of others by a significant margin. The best result that the Reguformer achieves on well-interval similarity task is the mean PR~AUC score equal to 0.983, which is comparable to the classical Transformer and outperforms the previous models.