Time-series Extreme Event Forecasting with Neural Networks at Uber

By Nikolay Pavlovich Laptev, Jason Yosinski, Li Erran Li, Slawek Smyl. 2017

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

Citations in the corpus

Enhancing Time-Series Momentum Strategies Using Deep Neural Networks (2019, methodology): “In recent times, deep neural networks have been increasingly used for time-series prediction and have outperformed traditional benchmarks in applications such as demand forecasting (Laptev et al. 2017), medicine (Lim and van der Schaar 2018), and finance (Zhang, Zohren, and Roberts 2019).”
Probabilistic Forecasting with Temporal Convolutional Neural Network (2019, background): “Both RNNs and CNNs have been shown to be able to model complex nonlinear feature interactions and yield substantial forecasting performances, especially when many related time series are available [2, 9, 18, 27, 29, 34].”
N-HiTS: Neural Hierarchical Interpolation for Time Series Forecasting (2022, methodology): “Over the past few years, deep forecasting methods have become ubiquitous in industrial forecasting systems, with examples in optimal resource allocation and planning in transportation (Laptev et al., 2017), large e-commerce retail (Wen et al.”
CoST: Contrastive Learning of Disentangled Seasonal-Trend Representations for Time Series Forecasting (2022, methodology): “1 INTRODUCTION Time series forecasting has been widely applied to various domains, such as electricity pricing (Cuaresma et al., 2004), demand forecasting (Carbonneau et al., 2008), capacity planning and management (Kim, 2003), and anomaly detection (Laptev et al., 2017).”
Deep and Confident Prediction for Time Series at Uber (2017, methodology): “By providing a large amount of data across numerous dimensions, it has been shown that an LSTM approach can model complex extreme events by allowing nonlinear feature interactions [5], [6].”
Diffusion Convolutional Recurrent Neural Network: Data-Driven Traffic Forecasting (2017, background): “For example, in [25, 12], the authors study time series forecasting using deep Recurrent Neural Networks (RNN).”
GluonTS: Probabilistic Time Series Models in Python (2019, methodology): “…by sharing the parameters Φ. 6 Due to their capability to extract higher-order features, these neural methods can identify complex patterns within and across time series from datasets consisting of raw time series (Flunkert et al., to appear; Laptev et al., 2017) with considerably less human eﬀort.”
Spatio-Temporal Graph Structure Learning for Traffic Forecasting (2020, methodology): “Deep Recurrent Neural Networks were applied to trafﬁc forecasting (Yu et al. 2017; Laptev et al. 2017).”
Deep Learning for Time Series Forecasting: Tutorial and Literature Survey (2020, methodology): “[111] employ an LSTM as a feature extractor (LSTM autoencoder), and use the extracted features, combined with external inputs to generate the forecasts with another LSTM.”
Recurrent Neural Networks for Time Series Forecasting: Current Status and Future Directions (2019, background): “Laptev et al. (2017) incorporate an autoencoder to train a global model across many heterogeneous time series.”
Backpropagation algorithms and Reservoir Computing in Recurrent Neural Networks for the forecasting of complex spatiotemporal dynamics (2019, methodology): “…Kampffmeyer, Rizzi, & Jenssen, 2017) RNN architectures have been bench-marked for short-term load forecasting of demand and consumption of resources in a supply network, while in Laptev, Yosinski, Li, and Smyl (2017) they are utilized for extreme event detection in low dimensional time series.”
Deep State Space Models for Time Series Forecasting (2018, background): “With their capability to extract higher order features, they can identify complex patterns within and across time series, and can do so from datasets of raw time series with considerably less human effort [9, 27, 19].”