# Seq-U-Net: A One-Dimensional Causal U-Net for Efficient Sequence Modelling

arXiv: https://arxiv.org/abs/1911.06393 · HTML: https://mlsanity.com/paper/1911.06393/

## Key facts

- Cited by 16 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 14 Nov 2019.

## Built on by (most cited first, 16 of 16)

- [Triformer: Triangular, Variable-Specific Attentions for Long Sequence Multivariate Time Series Forecasting--Full Version](https://mlsanity.com/paper/2204.13767/) (2022)
- [Fredformer: Frequency Debiased Transformer for Time Series Forecasting](https://mlsanity.com/paper/2406.09009/) (2024)
- [Yformer: U-Net Inspired Transformer Architecture for Far Horizon Time Series Forecasting](https://mlsanity.com/paper/2110.08255/) (2021)
- [TCCT: Tightly-Coupled Convolutional Transformer on Time Series Forecasting](https://mlsanity.com/paper/2108.12784/) (2021)
- [Low-Rank Constraints for Fast Inference in Structured Models](https://mlsanity.com/paper/2201.02715/) (2022)
- [TILDE-Q: A Transformation Invariant Loss Function for Time-Series Forecasting](https://mlsanity.com/paper/2210.15050/) (2022)
- [FDNet: Focal Decomposed Network for Efficient, Robust and Practical Time Series Forecasting](https://mlsanity.com/paper/2306.10703/) (2023)
- [Low-latency Real-time Voice Conversion on CPU](https://mlsanity.com/paper/2311.00873/) (2023)
- [LORD: Lower-Dimensional Embedding of Log-Signature in Neural Rough Differential Equations](https://mlsanity.com/paper/2204.08781/) (2022)
- [Temporal Saliency Detection Towards Explainable Transformer-based Timeseries Forecasting](https://mlsanity.com/paper/2212.07771/) (2022)
- [ArterialNet: Reconstructing Arterial Blood Pressure Waveform with Wearable Pulsatile Signals, a Cohort-Aware Approach](https://mlsanity.com/paper/2410.18895/) (2024)
- [QuantFormer: Learning to Quantize for Neural Activity Forecasting in Mouse Visual Cortex](https://mlsanity.com/paper/2412.07264/) (2024)
- [HeatGen: A Guided Diffusion Framework for Multiphysics Heat Sink Design Optimization](https://mlsanity.com/paper/2511.09578/) (2025)
- [WaveBound: Dynamic Error Bounds for Stable Time Series Forecasting](https://mlsanity.com/paper/2210.14303/) (2022)
- [Forecasting VIX using Bayesian Deep Learning](https://mlsanity.com/paper/2401.17042/) (2024)
- [A physics-informed U-Net-LSTM network for nonlinear structural response under seismic excitation](https://mlsanity.com/paper/2511.21276/) (2025)

## Abstract (arXiv)

Convolutional neural networks (CNNs) with dilated filters such as the Wavenet or the Temporal Convolutional Network (TCN) have shown good results in a variety of sequence modelling tasks. However, efficiently modelling long-term dependencies in these sequences is still challenging. Although the receptive field of these models grows exponentially with the number of layers, computing the convolutions over very long sequences of features in each layer is time and memory-intensive, prohibiting the use of longer receptive fields in practice. To increase efficiency, we make use of the "slow feature" hypothesis stating that many features of interest are slowly varying over time. For this, we use a U-Net architecture that computes features at multiple time-scales and adapt it to our auto-regressive scenario by making convolutions causal. We apply our model ("Seq-U-Net") to a variety of tasks including language and audio generation. In comparison to TCN and Wavenet, our network consistently saves memory and computation time, with speed-ups for training and inference of over 4x in the audio generation experiment in particular, while achieving a comparable performance in all tasks.