Subscribe to the gold package and get unlimited access to Shamra Academy
Register a new userFaster Retrieval with a Two-Pass Dynamic-Time-Warping Lower Bound
75
0
0.0
(
0
)
Added by
Daniel Lemire
Publication date
2009
fields
Informatics Engineering
and research's language is
English
Authors
Daniel Lemire
Ask ChatGPT about the research
No Arabic abstract
The Dynamic Time Warping (DTW) is a popular similarity measure between time series. The DTW fails to satisfy the triangle inequality and its computation requires quadratic time. Hence, to find closest neighbors quickly, we use bounding techniques. We can avoid most DTW computations with an inexpensive lower bound (LB Keogh). We compare LB Keogh with a tighter lower bound (LB Improved). We find that LB Improved-based search is faster. As an example, our approach is 2-3 times faster over random-walk and shape time series.
rate research
Read More
We prove a lower bound on the space complexity of two-pass semi-streaming algorithms that approximate the maximum matching problem. The lower bound is parameterized by the density of Ruzsa-Szemeredi graphs: * Any two-pass semi-streaming algorithm for maximum matching has approximation ratio at least $(1- Omega(frac{log{RS(n)}}{log{n}}))$, where $RS(n)$ denotes the maximum number of induced matchings of size $Theta(n)$ in any $n$-vertex graph, i.e., the largest density of a Ruzsa-Szemeredi graph. Currently, it is known that $n^{Omega(1/!loglog{n})} leq RS(n) leq frac{n}{2^{O(log^*{!(n)})}}$ and closing this (large) gap between upper and lower bounds has remained a notoriously difficult problem in combinatorics. Under the plausible hypothesis that $RS(n) = n^{Omega(1)}$, our lower bound is the first to rule out small-constant approximation two-pass semi-streaming algorithms for the maximum matching problem, making progress on a longstanding open question in the graph streaming literature.
Dynamic Time Warping (DTW) is widely used for temporal data processing. However, existing methods can neither learn the discriminative prototypes of different classes nor exploit such prototypes for further analysis. We propose Discriminative Prototype DTW (DP-DTW), a novel method to learn class-specific discriminative prototypes for temporal recognition tasks. DP-DTW shows superior performance compared to conventional DTWs on time series classification benchmarks. Combined with end-to-end deep learning, DP-DTW can handle challenging weakly supervised action segmentation problems and achieves state of the art results on standard benchmarks. Moreover, detailed reasoning on the input video is enabled by the learned action prototypes. Specifically, an action-based video summarization can be obtained by aligning the input sequence with action prototypes.
We reduce training time in convolutional networks (CNNs) with a method that, for some of the mini-batches: a) scales down the resolution of input images via downsampling, and b) reduces the forward pass operations via pooling on the convolution filters. Training is performed in an interleaved fashion; some batches undergo the regular forward and backpropagation passes with original network parameters, whereas others undergo a forward pass with pooled filters and downsampled inputs. Since pooling is differentiable, the gradients of the pooled filters propagate to the original network parameters for a standard parameter update. The latter phase requires fewer floating point operations and less storage due to the reduced spatial dimensions in feature maps and filters. The key idea is that this phase leads to smaller and approximate updates and thus slower learning, but at significantly reduced cost, followed by passes that use the original network parameters as a refinement stage. Deciding how often and for which batches the downsmapling occurs can be done either stochastically or deterministically, and can be defined as a training hyperparameter itself. Experiments on residual architectures show that we can achieve up to 23% reduction in training time with minimal loss in validation accuracy.
Models based on neural networks and machine learning are seeing a rise in popularity in space physics. In particular, the forecasting of geomagnetic indices with neural network models is becoming a popular field of study. These models are evaluated with metrics such as the root-mean-square error (RMSE) and Pearson correlation coefficient. However, these classical metrics sometimes fail to capture crucial behavior. To show where the classical metrics are lacking, we trained a neural network, using a long short-term memory network, to make a forecast of the disturbance storm time index at origin time $t$ with a forecasting horizon of 1 up to 6 hours, trained on OMNIWeb data. Inspection of the models results with the correlation coefficient and RMSE indicated a performance comparable to the latest publications. However, visual inspection showed that the predictions made by the neural network were behaving similarly to the persistence model. In this work, a new method is proposed to measure whether two time series are shifted in time with respect to each other, such as the persistence model output versus the observation. The new measure, based on Dynamical Time Warping, is capable of identifying results made by the persistence model and shows promising results in confirming the visual observations of the neural networks output. Finally, different methodologies for training the neural network are explored in order to remove the persistence behavior from the results.
During the last decades there is a continuing international endeavor in developing realistic space weather prediction tools aiming to forecast the conditions on the Sun and in the interplanetary environment. These efforts have led to the need of developing appropriate metrics in order to assess the performance of those tools. Metrics are necessary for validating models, comparing different models and monitoring adjustments or improvements of a certain model over time. In this work, we introduce the Dynamic Time Warping (DTW) as an alternative way to validate models and, in particular, to quantify differences between observed and synthetic (modeled) time series for space weather purposes. We present the advantages and drawbacks of this method as well as applications on WIND observations and EUHFORIA modeled output at L1. We show that DTW is a useful tool that permits the evaluation of both the fast and slow solar wind. Its distinctive characteristic is that it warps sequences in time, aiming to align them with the minimum cost by using dynamic programming. It can be applied in two different ways for the evaluation of modeled solar wind time series. The first way calculates the so-called sequence similarity factor (SSF), a number that provides a quantification of how good the forecast is, compared to a best and a worst case prediction scenarios. The second way quantifies the time and amplitude differences between the points that are best matched between the two sequences. As a result, it can serve as a hybrid metric between continuous measurements (such as, e.g., the correlation coefficient) and point-by-point comparisons. We conclude that DTW is a promising technique for the assessment of solar wind profiles offering functions that other metrics do not, so that it can give at once the most complete evaluation profile of a model.
Log in to be able to interact and post comments
comments
Fetching comments
Sign in to be able to follow your search criteria
