Cycling is a promising sustainable mode for commuting and leisure in cities, however, the fear of getting hit or fall reduces its wide expansion as a commuting mode. In this paper, we introduce a novel method called CyclingNet for detecting cycling near misses from video streams generated by a mounted frontal camera on a bike regardless of the camera position, the conditions of the built, the visual conditions and without any restrictions on the riding behaviour. CyclingNet is a deep computer vision model based on convolutional structure embedded with self-attention bidirectional long-short term memory (LSTM) blocks that aim to understand near misses from both sequential images of scenes and their optical flows. The model is trained on scenes of both safe rides and near misses. After 42 hours of training on a single GPU, the model shows high accuracy on the training, testing and validation sets. The model is intended to be used for generating information that can draw significant conclusions regarding cycling behaviour in cities and elsewhere, which could help planners and policy-makers to better understand the requirement of safety measures when designing infrastructure or drawing policies. As for future work, the model can be pipelined with other state-of-the-art classifiers and object detectors simultaneously to understand the causality of near misses based on factors related to interactions of road-users, the built and the natural environments.

The growing energy and performance costs of deep learning have driven the community to reduce the size of neural networks by selectively pruning components. Similarly to their biological counterparts, sparse networks generalize just as well, if not better than, the original dense networks. Sparsity can reduce the memory footprint of regular networks to fit mobile devices, as well as shorten training time for ever growing networks. In this paper, we survey prior work on sparsity in deep learning and provide an extensive tutorial of sparsification for both inference and training. We describe approaches to remove and add elements of neural networks, different training strategies to achieve model sparsity, and mechanisms to exploit sparsity in practice. Our work distills ideas from more than 300 research papers and provides guidance to practitioners who wish to utilize sparsity today, as well as to researchers whose goal is to push the frontier forward. We include the necessary background on mathematical methods in sparsification, describe phenomena such as early structure adaptation, the intricate relations between sparsity and the training process, and show techniques for achieving acceleration on real hardware. We also define a metric of pruned parameter efficiency that could serve as a baseline for comparison of different sparse networks. We close by speculating on how sparsity can improve future workloads and outline major open problems in the field.

The use of Deep Learning hardware algorithms for embedded applications is characterized by challenges such as constraints on device power consumption, availability of labeled data, and limited internet bandwidth for frequent training on cloud servers. To enable low power implementations, we consider efficient bitwidth reduction and pruning for the class of Deep Learning algorithms known as Discriminative Deep Belief Networks (DDBNs) for embedded-device classification tasks. We train DDBNs with both generative and discriminative objectives under an approximate computing framework and analyze their power-at-performance for supervised and semi-supervised applications. We also investigate the out-of-distribution performance of DDBNs when the inference data has the same class structure yet is statistically different from the training data owing to dynamic real-time operating environments. Based on our analysis, we provide novel insights and recommendations for choice of training objectives, bitwidth values, and accuracy sensitivity with respect to the amount of labeled data for implementing DDBN inference with minimum power consumption on embedded hardware platforms subject to accuracy tolerances.

This paper introduces a novel method to fine-tune handwriting recognition systems based on Recurrent Neural Networks (RNN). Long Short-Term Memory (LSTM) networks are good at modeling long sequences but they tend to overfit over time. To improve the system's ability to model sequences, we propose to drop information at random positions in the sequence. We call our approach Temporal Dropout (TD). We apply TD at the image level as well to internal network representation. We show that TD improves the results on two different datasets. Our method outperforms previous state-of-the-art on Rodrigo dataset.

Time-series is ubiquitous across applications, such as transportation, finance and healthcare. Time-series is often influenced by external factors, especially in the form of asynchronous events, making forecasting difficult. However, existing models are mainly designated for either synchronous time-series or asynchronous event sequence, and can hardly provide a synthetic way to capture the relation between them. We propose Variational Synergetic Multi-Horizon Network (VSMHN), a novel deep conditional generative model. To learn complex correlations across heterogeneous sequences, a tailored encoder is devised to combine the advances in deep point processes models and variational recurrent neural networks. In addition, an aligned time coding and an auxiliary transition scheme are carefully devised for batched training on unaligned sequences. Our model can be trained effectively using stochastic variational inference and generates probabilistic predictions with Monte-Carlo simulation. Furthermore, our model produces accurate, sharp and more realistic probabilistic forecasts. We also show that modeling asynchronous event sequences is crucial for multi-horizon time-series forecasting.

Are you looking to get a better overview of the different Deep Learning research streams currently being pursued? Do you have too many open arXiv tabs to reasonably go through? Too little time to watch full videos? If only there was a quick summary of the key idea and concept of a paper. Then I am happy to introduce the ‚Machine-Learning-Collage' series.

Speech emotion recognition is a vital contributor to the next generation of human-computer interaction (HCI). However, current existing small-scale databases have limited the development of related research. In this paper, we present LSSED, a challenging large-scale english speech emotion dataset, which has data collected from 820 subjects to simulate real-world distribution. In addition, we release some pre-trained models based on LSSED, which can not only promote the development of speech emotion recognition, but can also be transferred to related downstream tasks such as mental health analysis where data is extremely difficult to collect. Finally, our experiments show the necessity of large-scale datasets and the effectiveness of pre-trained models. The dateset will be released on https://github.com/tobefans/LSSED.

Dimension of the encoder output (i.e., the code layer) in an autoencoder is a key hyper-parameter for representing the input data in a proper space. This dimension must be carefully selected in order to guarantee the desired reconstruction accuracy. Although overcomplete representation can address this dimension issue, the computational complexity will increase with dimension. Inspired by non-parametric methods, here, we propose a metalearning approach to increase the number of basis vectors used in dynamic sparse coding on the fly. An actor-critic algorithm is deployed to automatically choose an appropriate dimension for feature vectors regarding the required level of accuracy. The proposed method benefits from online dictionary learning and fast iterative shrinkage-thresholding algorithm (FISTA) as the optimizer in the inference phase. It aims at choosing the minimum number of bases for the overcomplete representation regarding the reconstruction error threshold. This method allows for online controlling of both the representation dimension and the reconstruction error in a dynamic framework.

We propose a novel bootstrap procedure for dependent data based on Generative Adversarial networks (GANs). We show that the dynamics of common stationary time series processes can be learned by GANs and demonstrate that GANs trained on a single sample path can be used to generate additional samples from the process. We find that temporal convolutional neural networks provide a suitable design for the generator and discriminator, and that convincing samples can be generated on the basis of a vector of iid normal noise. We demonstrate the finite sample properties of GAN sampling and the suggested bootstrap using simulations where we compare the performance to circular block bootstrapping in the case of resampling an AR(1) time series processes. We find that resampling using the GAN can outperform circular block bootstrapping in terms of empirical coverage.

In Maurice Merleau-Ponty's phenomenology of perception, analysis of perception accounts for an element of intentionality, and in effect therefore, perception and action cannot be viewed as distinct procedures. In the same line of thinking, Alva No\"{e} considers perception as a thoughtful activity that relies on capacities for action and thought. Here, by looking into psychology as a source of inspiration, we propose a computational model for the action involved in visual perception based on the notion of equilibrium as defined by Jean Piaget. In such a model, Piaget's equilibrium reflects the mind's status, which is used to control the observation process. The proposed model is built around a modified version of convolutional neural networks (CNNs) with enhanced filter performance, where characteristics of filters are adaptively adjusted via a high-level control signal that accounts for the thoughtful activity in perception. While the CNN plays the role of the visual system, the control signal is assumed to be a product of mind.