An official website of the United States government
Dynamic time warping estimates the alignment between two sequences and is designed to handle a variable amount of time warping. In many contexts, it performs poorly when confronted with two sequences of different scale, in which the average slope of the true alignment path in the pairwise cost matrix deviates significantly from one. This paper investigates ways to improve the robustness of DTW to such global time warping conditions, using an audio–audio alignment task as a motivating scenario of interest. We modify a dataset commonly used for studying audio–audio synchronization in order to construct a benchmark in which the global time warping conditions are carefully controlled, and we evaluate the effectiveness of several strategies designed to handle global time warping. Among the strategies tested, there is a clear winner: performing sequence length normalization via downsampling before invoking DTW. This method achieves the best alignment accuracy across a wide range of global time warping conditions, and it maintains or reduces the runtime compared to standard usages of DTW. We present experiments and analyses to demonstrate its effectiveness in both controlled and realistic scenarios.
more »
« less
DTWNet: a Dynamic Time Warping Network
Dynamic Time Warping (DTW) is widely used as a similarity measure in various domains. Due to its invariance against warping in the time axis, DTW provides more meaningful discrepancy measurements between two signals than other distance measures. In this paper, we propose a novel component in an artificial neural network. In contrast to the previous successful usage of DTW as a loss function, the proposed framework leverages DTW to obtain a better feature extraction. For the first time, the DTW loss is theoretically analyzed, and a stochastic backpropogation scheme is proposed to improve the accuracy and efficiency of the DTW learning. We also demonstrate that the proposed framework can be used as a data analysis tool to perform data decomposition.
more »
« less
- PAR ID:
- 10181828
- Date Published:
- Journal Name:
- Advances in neural information processing systems
- Volume:
- 32
- ISSN:
- 1049-5258
- Format(s):
- Medium: X
- Sponsoring Org:
- National Science Foundation
More Like this
-
-
null (Ed.)We propose NSYNC, a practical framework to compare side-channel signals for real-time intrusion detection in Additive Manufacturing (AM) systems. The motivation to develop NSYNC is that we find AM systems are asynchronous in nature and there is random variation in timing in a printing process. Although this random variation, referred to as time noise, is very small compared with the duration of a printing process, it can cause existing Intrusion Detection Systems (IDSs) to fail. To deal with this problem, NSYNC incorporates a dynamic synchronizer to find the timing relationship between two signals. This timing relationship, referred to as the horizontal displacement, can not only be used to mitigate the adverse effect of time noise on calculating the (vertical) distance between signals, but also be used as indicators for intrusion detection. An existing dynamic synchronizer is Dynamic Time Warping (DTW). However, we found in experiments that DTW not only consumes an excessive amount of computational resources but also has limited accuracy for processing side-channel signals. To solve this problem, we propose a novel dynamic synchronizer, called Dynamic Window Matching (DWM), to replace DTW. To compare NSYNC against existing IDSs, we built a data acquisition system that is capable of collecting six different types of side-channel signals and performed a total of 302 benign printing processes and a total of 200 malicious printing processes with two printers. Our experiment results show that existing IDSs leveraging side-channel signals in AM systems can only achieve an accuracy from 0.50 to 0.88, whereas our proposed NSYNC can reach an accuracy of 0.99.more » « less
-
This work proposes a framework for tracking a desired path of an object held by an adaptive hand via within-hand manipulation. Such underactuated hands are able to passively achieve stable contacts with objects. Combined with vision-based control and data-driven state estimation process, they can solve tasks without accurate hand-object models or multi-modal sensory feedback. In particular, a data-driven regression process is used here to estimate the probability of dropping the object for given manipulation states. Then, an optimization-based planner aims to track the desired path while avoiding states that are above a threshold probability of dropping the object. The optimized cost function, based on the principle of Dynamic-Time Warping (DTW), seeks to minimize the area between the desired and the followed path. By adapting the threshold for the probability of dropping the object, the framework can handle objects of different weights without retraining. Experiments involving writing letters with a marker, as well as tracing randomized paths, were conducted on the Yale Model T-42 hand. Results indicate that the framework successfully avoids undesirable states, while minimizing the proposed cost function, thereby producing object paths for within-hand manipulation that closely match the target ones.more » « less
-
The data-driven characterization of discretionary lane-changing behaviors has traditionally been hindered by the scarcity of high-resolution data that can precisely record lateral movements. In this study, we conducted an exploratory investigation leveraging the Third Generation Simulation (TGSIM) dataset to advance our understanding of discretionary lane-changing behaviors. In this paper, we developed a discretionary lane-changing extraction pipeline and scrutinized crucial factors such as gaps and relative speeds in leading and following directions. A dynamic time warping (DTW) analysis was performed to quantify the difference between any pair of lane-changing behaviors, and an affinity propagation (AP) clustering, evaluated on normalized DTW distance, was conducted. Our results yielded five clusters based on lead and lag gaps, enabling us to categorize lane-changing behaviors into aggressive, neutral, and cautious for both leading and following directions. Clustering based on relative speeds revealed two distinct groups of lane-changing behaviors, one representing overtaking and the other indicative of transitioning into a lane with stable and homogenous speed. The proposed DTW analysis, in conjunction with AP clustering, demonstrated promising potential in categorizing and characterizing lane-changing behaviors. Additionally, this approach can be readily adapted to analyze any driving behavior.more » « less
-
DTW calculates the similarity or alignment between two signals, subject to temporal warping. However, its computational complexity grows exponentially with the number of time-series. Although there have been algorithms developed that are linear in the number of time-series, they are generally quadratic in time-series length. The exception is generalized time warping (GTW), which has linear computational cost. Yet, it can only identify simple time warping functions. There is a need for a new fast, high-quality multisequence alignment algorithm. We introduce trainable time warping (TTW), whose complexity is linear in both the number and the length of time-series. TTW performs alignment in the continuoustime domain using a sinc convolutional kernel and a gradient-based optimization technique. We compare TTW and GTW on S5 UCR datasets in time-series averaging and classification. TTW outperforms GTW on 67.1% of the datasets for the averaging tasks, and 61.2% of the datasets for the classification tasks.more » « less
- Free Publicly Accessible Full Text
-
Accepted Manuscript1.0
- Conference Paper:
- The DOI is not currently available.
