Deep Declarative Dynamic Time Warping for End-to-End Learning of Alignment Paths

03/19/2023
by   Ming Xu, et al.
0

This paper addresses learning end-to-end models for time series data that include a temporal alignment step via dynamic time warping (DTW). Existing approaches to differentiable DTW either differentiate through a fixed warping path or apply a differentiable relaxation to the min operator found in the recursive steps used to solve the DTW problem. We instead propose a DTW layer based around bi-level optimisation and deep declarative networks, which we name DecDTW. By formulating DTW as a continuous, inequality constrained optimisation problem, we can compute gradients for the solution of the optimal alignment (with respect to the underlying time series) using implicit differentiation. An interesting byproduct of this formulation is that DecDTW outputs the optimal warping path between two time series as opposed to a soft approximation, recoverable from Soft-DTW. We show that this property is particularly useful for applications where downstream loss functions are defined on the optimal alignment path itself. This naturally occurs, for instance, when learning to improve the accuracy of predicted alignments against ground truth alignments. We evaluate DecDTW on two such applications, namely the audio-to-score alignment task in music information retrieval and the visual place recognition task in robotics, demonstrating state-of-the-art results in both.

READ FULL TEXT
research
03/05/2017

Soft-DTW: a Differentiable Loss Function for Time-Series

We propose in this paper a differentiable learning loss between time ser...
research
05/16/2017

Optimal Warping Paths are unique for almost every pair of Time Series

An optimal warping path between two time series is generally not unique....
research
08/10/2023

Stabilizing Training with Soft Dynamic Time Warping: A Case Study for Pitch Class Estimation with Weakly Aligned Targets

Soft dynamic time warping (SDTW) is a differentiable loss function that ...
research
10/16/2020

Differentiable Divergences Between Time Series

Computing the discrepancy between time series of variable sizes is notor...
research
08/04/2020

Exact, Parallelizable Dynamic Time Warping Alignment with Linear Memory

Audio alignment is a fundamental preprocessing step in many MIR pipeline...
research
06/16/2022

Closed-Form Diffeomorphic Transformations for Time Series Alignment

Time series alignment methods call for highly expressive, differentiable...
research
04/11/2023

Soft Dynamic Time Warping for Multi-Pitch Estimation and Beyond

Many tasks in music information retrieval (MIR) involve weakly aligned d...

Please sign up or login with your details

Forgot password? Click here to reset