Member-only story
Amazon’s GluonTS is a Deep Learning Framework for Time Series Forecasting
Optimized for PyTorch and MXNet, the framework includes implementations of state of the art deep learning models for time series.
I recently started an AI-focused educational newsletter, that already has over 125,000 subscribers. TheSequence is a no-BS (meaning no hype, no news etc) ML-oriented newsletter that takes 5 minutes to read. The goal is to keep you up to date with machine learning projects, research papers and concepts. Please give it a try by subscribing below:
Time series forecasting has been one of the main use cases for machine learning since its early days and, yet, it lags behind other disciplines in terms of technical breakthroughs. Even though time series forecasting hasn’t had a GPT-3 moment, there hasn’t been a lack of innovation in the space. However, most of the deep learning stacks applied to time series forecasting remained confined in research papers and haven’t been packaged into mainstream data science frameworks. GluonTS is an open source framework released by AWS Labs that tries to address this challenge.
GluonTS enables a simple programming model for time-series forecasting on top of Apache MXNet. Like its name indicates, the framework is based on the Gluon programming interface which abstracts key building blocks of deep learning models. In my opinion, one of the greatest contributions of GluonTS is that is it incorporates deep learning models for time-series forecasting. Despite the proliferation of time-series forecasting libraries in the market, most of them are based on simpler regressive models such as ARIMA or PROPHET. GluonTS extends those capabilities with deep learning architectures for time-series forecasting based on models such as convolutional neural networks(CNNs) or Long-Short Term Memory Networks(LSTMs). However, GluonTS expands beyond that providing a complete end-to-end pipeline for the implementation of time-series forecasting models. Architecturally…
