You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Sufficient physical activity and restful sleep play a major role in the prevention and cure of many chronic conditions.
Being able to proactively screen and monitor such chronic conditions would be a big step forward for overall health.
The rapid increase in the popularity of wearable devices pro-vides a significant new source, making it possible to track the user’s lifestyle real-time.
In this paper, we propose a novel unsupervised representation learning technique called activ-ity2vecthat learns and “summarizes” the discrete-valued ac-tivity time-series.
It learns the representations with three com-ponents: (i) the co-occurrence and magnitude of the activ-ity levels in a time-segment,
(ii) neighboring context of the time-segment, and (iii) promoting subject-invariance with ad-versarial training.
We evaluate our method on four disorder prediction tasks using linear classifiers. Empirical evaluation demonstrates that our proposed method scales and performs better than many strong baselines.
The adversarial regime helps improve the generalizability of our representations by promoting subject invariant features.
We also show that using the representations at the level of a day works the best since human activity is structured in terms of daily routines.
LabelForest: Non-Parametric Semi-Supervised Learning for Activity Recognition
Activity recognition is central to many motion analysis applications ranging from health assessment to gaming.
However, the need for obtaining sufficiently large amounts of labeled data has limited the development of personalized activity recognition models.
Semi-supervised learning has traditionally been a promising approach in many application domains to alleviate reliance on large amounts of labeled data by learning the label information from a small set of seed labels.
Nonetheless, existing approaches perform poorly in highly dynamic settings, such as wearable systems, because some algorithms rely on predefined hyper-parameters or distribution models that needs to be tuned for each user or context.
To address these challenges, we introduce LabelForest 1, a novel non-parametric semi-supervised learning framework for activity recognition. LabelForest has two algorithms at its core:
(1) a spanning forest algorithm for sample selection and label inference; and (2) a silhouette-based filtering method to finalize label augmentation for machine learning model training.
Our thorough analysis on three human activity datasets demonstrate that LabelForest achieves a labeling accuracy of 90.1% in presence of a skewed label distribution in the seed data.
Compared to self-training and other sequential learning algorithms, LabelForest achieves up to 56.9% and 175.3% improvement in the accuracy on balanced and unbalanced seed data, respectively.
Distribution-Based Semi-Supervised Learning for Activity Recognition
Supervised learning methods have been widely applied to activity recognition. The prevalent success of existing methods, however, has two crucial prerequisites: proper feature extraction and sufficient labeled training data. The former is important to differentiate activities, while the latter is crucial to build a precise learning model. These two prerequisites have become bottlenecks to make existing methods more practical. Most existing feature extraction methods highly depend on domain knowledge, while labeled data requires intensive human annotation effort. Therefore, in this paper, we propose a novel method, named Distribution-based Semi-Supervised Learning, to tackle the aforementioned limitations. The proposed method is capable of automatically extracting powerful features with no domain knowledge required, meanwhile, alleviating the heavy annotation effort through semi-supervised learning. Specifically, we treat data stream of sensor readings received in a period as a distribution, and map all training distributions, including labeled and unlabeled, into a reproducing kernel Hilbert space (RKHS) using the kernel mean embedding technique. The RKHS is further altered by exploiting the underlying geometry structure of the unlabeled distributions. Finally, in the altered RKHS, a classifier is trained with the labeled distributions. We conduct extensive experiments on three public datasets to verify the effectiveness of our method compared with state-of-the-art baselines.
Molecular dynamics (MD) is a powerful computational method for simulating molecular behavior. Deep neural networks provide a novel method of generating MD data efficiently, but there is no architecture that mitigates the well-known exposure bias accumulated by multi-step generations. In this paper, we propose a multi-step time series generator using a deep neural network based on Wasserstein generative adversarial nets. Instead of sparse real data, our model evolves a latent variable z that is densely distributed in a low-dimensional space. This novel framework successfully mitigates the exposure bias. Moreover, our model can evolve part of the system (Feature extraction) with any time step (Step skip), which accelerates the efficient generation of MD data. The applicability of this model is evaluated through three different systems: harmonic oscillator, bulk water, and polymer melts. The experimental results demonstrate that our model can generate time series of the MD data with sufficient accuracy to calculate the physical and important dynamical statistics.
Attend and Diagnose: Clinical Time Series Analysis Using Attention Models
With widespread adoption of electronic health records, there is an increased emphasis for predictive models that can effectively deal with clinical time-series data. Powered by Recurrent Neural Network (RNN) architectures with Long Short-Term Memory (LSTM) units, deep neural networks have achieved state-of-the-art results in several clinical prediction tasks. Despite the success of RNN, its sequential nature prohibits parallelized computing, thus making it inefficient particularly when processing long sequences. Recently, architectures which are based solely on attention mechanisms have shown remarkable success in transduction tasks in NLP, while being computationally superior. In this paper, for the first time, we utilize attention models for clinical time-series modeling, thereby dispensing recurrence entirely. We develop the SAnD (Simply Attend and Diagnose) architecture, which employs a masked, self-attention mechanism, and uses positional encoding and dense interpolation strategies for incorporating temporal order. Furthermore, we develop a multi-task variant of SAnD to jointly infer models with multiple diagnosis tasks. Using the recent MIMIC-III benchmark datasets, we demonstrate that the proposed approach achieves state-of-the-art performance in all tasks, outperforming LSTM models and classical baselines with hand-engineered features.
Fourier Feature Approximations for Periodic Kernels in Time-Series Modelling
Gaussian Processes (GPs) provide an extremely powerful mechanism to model a variety of problems but incur an O(N3) complexity in the number of data samples. Common approximation methods rely on what are often termed inducing points but still typically incur an O(NM2) complexity in the data and corresponding inducing points. Using Random Fourier Feature (RFF) maps, we overcome this by transforming the problem into a Bayesian Linear Regression formulation upon which we apply a Bayesian Variational treatment that also allows learning the corresponding kernel hyperparameters, likelihood and noise parameters. In this paper we introduce an alternative method using Fourier series to obtain spectral representations of common kernels, in particular for periodic warpings, which surprisingly have a convergent, non-random form using special functions, requiring fewer spectral features to approximate their corresponding kernel to high accuracy. Using this, we can fuse the Random Fourier Feature spectral representations of common kernels with their periodic counterparts to show how they can more effectively and expressively learn patterns in time-series for both interpolation and extrapolation. This method combines robustness, scalability and equally importantly, interpretability through a symbolic declarative grammar that is both functionally and humanly intuitive — a property that is crucial for explainable decision making. Using probabilistic programming and Variational Inference we are able to efficiently optimise over these rich functional representations. We show significantly improved Gram matrix approximation errors, and also demonstrate the method in several time-series problems comparing other commonly used approaches such as recurrent neural networks.
Non-Parametric Outliers Detection in Multiple Time Series A Case Study: Power Grid Data Analysis
In this study we consider the problem of outlier detection with multiple co-evolving time series data. To capture both the temporal dependence and the inter-series relatedness, a multi-task non-parametric model is proposed, which can be extended to data with a broader exponential family distribution by adopting the notion of Bregman divergence. Albeit convex, the learning problem can be hard as the time series accumulate. In this regards, an efficient randomized block coordinate descent (RBCD) algorithm is proposed. The model and the algorithm is tested with a real-world application, involving outlier detection and event analysis in power distribution networks with high resolution multi-stream measurements. It is shown that the incorporation of inter-series relatedness enables the detection of system level events which would otherwise be unobservable with traditional methods.
T-C3D: Temporal Convolutional 3D Network for Real-Time Action Recognition
Video-based action recognition with deep neural networks has shown remarkable progress. However, most of the existing approaches are too computationally expensive due to the complex network architecture. To address these problems, we propose a new real-time action recognition architecture, called Temporal Convolutional 3D Network (T-C3D), which learns video action representations in a hierarchical multi-granularity manner. Specifically, we combine a residual 3D convolutional neural network which captures complementary information on the appearance of a single frame and the motion between consecutive frames with a new temporal encoding method to explore the temporal dynamics of the whole video. Thus heavy calculations are avoided when doing the inference, which enables the method to be capable of real-time processing. On two challenging benchmark datasets, UCF101 and HMDB51, our method is significantly better than state-of-the-art real-time methods by over 5.4% in terms of accuracy and 2 times faster in terms of inference speed (969 frames per second), demonstrating comparable recognition performance to the state-of-the-art methods. The source code for the complete system as well as the pre-trained models are publicly available at https://github.com/tc3d.
Enhancing Machine Learning Classification for Electrical Time Series Applications
Machine learning applications to electrical time series data will have wide-ranging impacts in the near future. Electricity disaggregation holds the promise of reducing billions of dollars of electrical waste every year. In the power grid, automatic classification of disturbance events detected by phasor measurement units could prevent cascading blackouts before they occur. Additional applications include better market segmentation by utility companies, improved design of appliances, and reliable incorporation of renewable energy resources into the power grid. However, existing machine learning methods remain unimplemented in the real world because of limiting assumptions that hinder performance. My research contributions are summarized as follows: In electricity disaggregation, I introduced the first label correction approach for supervised training samples. For unsupervised disaggregation, I introduced event detection that does not require parameter tuning and appliance discovery that makes no assumptions on appliance types. These improvements produce better accuracy, faster computation, and more scalability than any previously introduced method and can be to applied natural gas disaggregation, water disaggregation, and other source separation domains. My current work challenges long-held assumptions in time series shapelets, a classification tool with applicability in electrical time series and dozens of additional domains.
Learning Datum-Wise Sampling Frequency for Energy-Efficient Human Activity Recognition
Continuous Human Activity Recognition (HAR) is an important application of smart mobile/wearable systems for providing dynamic assistance to users. However, HAR in real-time requires continuous sampling of data using built-in sensors (e.g., accelerometer), which significantly increases the energy cost and shortens the operating span. Reducing sampling rate can save energy but causes low recognition accuracy. Therefore, choosing adaptive sampling frequency that balances accuracy and energy efficiency becomes a critical problem in HAR. In this paper, we formalize the problem as minimizing both classification error and energy cost by choosing dynamically appropriate sampling rates. We propose Datum-Wise Frequency Selection (DWFS) to solve the problem via a continuous state Markov Decision Process (MDP). A policy function is learned from the MDP, which selects the best frequency for sampling an incoming data entity by exploiting a datum related state of the system. We propose a method for alternative learning the parameters of an activity classification model and the MDP that improves both the accuracy and the energy efficiency. We evaluate DWFS with three real-world HAR datasets, and the results show that DWFS statistically outperforms the state-of-the-arts regarding a combined measurement of accuracy and energy efficiency.
Sensor-Based Activity Recognition via Learning From Distributions
Sensor-based activity recognition aims to predict users' activities from multi-dimensional streams of various sensor readings received from ubiquitous sensors. To use machine learning techniques for sensor-based activity recognition, previous approaches focused on composing a feature vector to represent sensor-reading streams received within a period of various lengths. With the constructed feature vectors, e.g., using predefined orders of moments in statistics, and their corresponding labels of activities, standard classification algorithms can be applied to train a predictive model, which will be used to make predictions online. However, we argue that in this way some important information, e.g., statistical information captured by higher-order moments, may be discarded when constructing features. Therefore, in this paper, we propose a new method, denoted by SMMAR, based on learning from distributions for sensor-based activity recognition. Specifically, we consider sensor readings received within a period as a sample, which can be represented by a feature vector of infinite dimensions in a Reproducing Kernel Hilbert Space (RKHS) using kernel embedding techniques. We then train a classifier in the RKHS. To scale-up the proposed method, we further offer an accelerated version by utilizing an explicit feature map instead of using a kernel function. We conduct experiments on four benchmark datasets to verify the effectiveness and scalability of our proposed method.
Extreme Low Resolution Activity Recognition With Multi-Siamese Embedding Learning
This paper presents an approach for recognizing human activities from extreme low resolution (e.g., 16x12) videos. Extreme low resolution recognition is not only necessary for analyzing actions at a distance but also is crucial for enabling privacy-preserving recognition of human activities. We design a new two-stream multi-Siamese convolutional neural network. The idea is to explicitly capture the inherent property of low resolution (LR) videos that two images originated from the exact same scene often have totally different pixel values depending on their LR transformations. Our approach learns the shared embedding space that maps LR videos with the same content to the same location regardless of their transformations. We experimentally confirm that our approach of jointly learning such transform robust LR video representation and the classifier outperforms the previous state-of-the-art low resolution recognition approaches on two public standard datasets by a meaningful margin.
Mobile Network Failure Event Detection and Forecasting With Multiple User Activity Data Sets
As the demand for mobile network services increases, immediate detection and forecasting of network failure events have become important problems for service providers. Several event detection approaches have been proposed to tackle these problems by utilizing social data. However, these approaches have not tried to solve event detection and forecasting problems from multiple data sets, such as web access logs and search queries. In this paper, we propose a machine learning approach that incorporates multiple user activity data into detecting and forecasting failure events. Our approach is based on a two-level procedure. First, we introduce a novel feature construction method that treats both the imbalanced label problem and the data sparsity problem of user activity data. Second, we propose a model ensemble method that combines outputs of supervised and unsupervised learning models for each data set and gives accurate predictions of network service outage. We demonstrate the effectiveness of the proposed models by extensive experiments with real-world failure events occurred at a network service provider in Japan and three user activity data sets.
