Zilong Wang
ABSTRACT
Subsequence search, especially subsequence similarity search, is one of the most important subroutines in time series data mining algorithms, and there is increasing evidence that Dynamic Time Warping (DTW) is the best distance metric. However, in spite of the great effort in software speedup techniques, including early abandoning strategies, lower bound, indexing, computation-reuse, DTW still cost too much time for many applications, e.g. 80% of the total time. Since DTW is a 2-Dimension sequential dynamic search with quite high data dependency, it is hard to use parallel hardware to accelerate it. In this work, we propose a novel framework for FPGA based subsequence similarity search and a novel PE-ring structure for DTW calculation. This framework utilizes the data reusability of continuous DTW calculations to reduce the bandwidth and exploit the coarse-grain parallelism; meanwhile guarantees the accuracy with a two-phase precision reduction. The PE-ring supports on-line updating patterns of arbitrary lengths, and utilizes the hard-wired synchronization of FPGA to realize the fine-grained parallelism. It also achieves flexible parallelism degree to do performance-cost trade-off. The experimental results show that we can achieve several orders of magnitude speedup in accelerating subsequence similarity search compared with the best software and current GPU/FPGA implementations in different datasets.
- Adams, N., Marquez, D., and Wakefield, G. 2005. Iterative deepening for melody alignment and retrieval. ISMIR.Google Scholar
- Alon, J., Athitsos, V., Yuan, Q., and Sclaroff, S. 2009. A unified framework for gesture recognition and spatiotemporal gesture segmentation. IEEE PAMI 31, 9, 1685--1699. Google ScholarDigital Library
- Chadwick, N., McMeekin, D., and Tan, T. 2011. Classifying eye and head movement artifacts in EEG Signals. DEST, 285--291.Google Scholar
- Ding, H., Trajcevski, G., Scheuermann, P., Wang, X., and Keogh, E. J. 2008. Querying and mining of time series data: experimental comparison of representations and distance measures. PVLDB 1, 2, 1542--52. Google ScholarDigital Library
- Fu, A., Keogh, E. J., Lau, L., Ratanamahatana, C., and Wong, R. 2008. Scaling and time warping in time series querying. VLDB J. 17, 4, 899--921 Google ScholarDigital Library
- Grimaldi, M., Albanese, D., Jurman, G., and Furlanello, C. 2009. Mining Very Large Databases of Time-Series: Speeding up Dynamic Time Warping using GPGPU. NIPS WorkshopGoogle Scholar
- Keogh, E. J., Wei, L., Xi, X., Vlachos, M., Lee, S. H., and Protopapas, P. 2009. Supporting exact indexing of arbitrarily rotated shapes and periodic time series under Euclidean and warping distance measures. VLDB J. 18, 3, 611--630. Google ScholarDigital Library
- Kim, S., Park, S., and Chu, W. 2001. An index-based approach for similarity search supporting time warping in large sequence databases. ICDE, 607--61. Google ScholarDigital Library
- Lim, S. H., Park, H., and Kim, S. W. 2007. Using multiple indexes for efficient subsequence matching in time-series databases. Inf. Sci. 177, 24, 5691--5706. Google ScholarDigital Library
- Rakthanmanon, T., Campana, B. J. L., Mueen, A., Batista, G. E. A. P. A., Westover, M. B., Zhu, Q., Zakaria, J. and Keogh, E. J. 2012. Searching and mining trillions of time series subsequences under dynamic time warping. KDD, 262--270. Google ScholarDigital Library
- Sakurai, Y., Faloutsos, C., and Yamamuro, M. 2007. Stream Monitoring under the Time Warping Distance. ICDE, 1046--1055Google Scholar
- Sart, D., Mueen, A., Najjar, W., Niennattrakul, V., and Keogh, E. J. 2010. Accelerating Dynamic Time Warping Subsequence Search with GPUs and FPGAs. ICDM, 1001--1006 Google ScholarDigital Library
- Srikanthan, S., Kumar, A., and Gupta, R. 2011. Implementing the dynamic time warping algorithm in multithreaded environments for real time and unsupervised pattern discovery. IEEE ICCCT, 394--398.Google Scholar
- Takhashi, N., Yoshihisa, T., Sakurai, Y., and Kanazawa, M. 2009. A Parallelized Data Stream Processing System using Dynamic Time Warping Distance. CISIS 1100--1105Google Scholar
- www.terasic.com.twGoogle Scholar
- www.festvox.org/cmu_arctic/Google Scholar
- Zhang, Y., Adl, K., and Glass, J. 2012. Fast spoken query detection using lower-bound Dynamic Time Warping on Graphical Processing Units. ICASSP, 5173 -- 5176.Google Scholar
Index Terms
- Accelerating subsequence similarity search based on dynamic time warping distance with FPGA
Recommendations
Accelerating Dynamic Time Warping Subsequence Search with GPUs and FPGAs
ICDM '10: Proceedings of the 2010 IEEE International Conference on Data MiningMany time series data mining problems require subsequence similarity search as a subroutine. Dozens of similarity/distance measures have been proposed in the last decade and there is increasing evidence that Dynamic Time Warping (DTW) is the best ...
Speeding up similarity search under dynamic time warping by pruning unpromising alignments
Similarity search is the core procedure for several time series mining tasks. While different distance measures can be used for this purpose, there is clear evidence that the Dynamic Time Warping (DTW) is the most suitable distance function for a wide ...
Time Series Subsequence Similarity Search Under Dynamic Time Warping Distance on the Intel Many-core Accelerators
SISAP 2015: Proceedings of the 8th International Conference on Similarity Search and Applications - Volume 9371Subsequence similarity search is one of the most important problems of time series data mining. Nowadays there is empirical evidence that Dynamic Time Warping DTW is the best distance metric for many applications. However in spite of sophisticated ...
Comments