Recurrent and convolutional neural networks are the most common architectures used for time series forecasting in deep learning literature. These networks use parameter sharing by repeating a set of fixed architectures with fixed parameters over time or space. The result is that the overall architecture is time-invariant (shift-invariant in the spatial domain) or stationary. We argue that time-invariance can reduce the capacity to perform multi-step-ahead forecasting, where modelling the dynamics at a range of scales and resolutions is required. We propose ForecastNet which uses a deep feed-forward architecture to provide a time-variant model. An additional novelty of ForecastNet is interleaved outputs, which we show assist in mitigating vanishing gradients. ForecastNet is demonstrated to outperform statistical and deep learning benchmark models on several datasets.
Huazhong University of Science and Technology（华中科技大学）
Despite the superiority of convolutional neural networks demonstrated in time series modeling and forecasting, it has not been fully explored on the design of the neural network architecture and the tuning of the hyper-parameters. Inspired by the incremental construction strategy for building a random multilayer perceptron, we propose a novel Error-feedback Stochastic Modeling (ESM) strategy to construct a random Convolutional Neural Network (ESM-CNN) for time series forecasting task, which builds the network architecture adaptively. The ESM strategy suggests that random filters and neurons of…
An accurate prediction of crude oil prices over long future horizons is challenging and of great interest to governments, enterprises, and investors. This paper proposes a revised hybrid model built upon empirical mode decomposition (EMD) based on the feed-forward neural network (FNN) modeling framework incorporating the slope-based method (SBM), which is capable of capturing the complex dynamic of crude oil prices. Three commonly used multi-step-ahead prediction strategies proposed in the literature, including iterated strategy, direct strategy, and MIMO (multiple-input multiple-output) strat…
University of Pittsburgh（匹兹堡大学）；Google（谷歌）；JD Finance America Corporation（京东金融美国公司）
Although the distributed machine learning methods can speed up the training of large deep neural networks, the communication cost has become the non-negligible bottleneck to constrain the performance. To address this challenge, the gradient compression based communication-efficient distributed learning methods were designed to reduce the communication cost, and more recently the local error feedback was incorporated to compensate for the corresponding performance loss. However, in this paper, we will show that a new "gradient mismatch" problem is raised by the local error feedback in centraliz…
SA-PEF: Step-Ahead Partial Error Feedback for Efficient Federated Learning
Department of Electronic Systems, Norwegian University of Science and Technology (NTNU)（挪威科学技术大学电子系统系）；Department of Information and Communications Engineering, Aalto University（阿尔托大学信息与通信工程系）；CSIRO’s Data61（CSIRO数据61）
Università della Svizzera Italiana（瑞士意大利语区大学）；Politecnico di Milano（米兰理工大学）
Management and efficient operations in critical infrastructure such as Smart Grids take huge advantage of accurate power load forecasting which, due to its nonlinear nature, remains a challenging task. Recently, deep learning has emerged in the machine learning field achieving impressive performance in a vast range of tasks, from image classification to machine translation. Applications of deep learning models to the electric load forecasting problem are gaining interest among researchers as well as the industry, but a comprehensive and sound comparison among different architectures is not yet…
Universidad Nacional de Colombia（哥伦比亚国立大学）；University of Florida（佛罗里达大学）
Accurate mechanisms for forecasting solar irradiance and insolation provide important information for the planning of renewable energy and agriculture projects as well as for environmental and socio-economical studies. This research introduces a pipeline for the one-day ahead forecasting of solar irradiance and insolation that only requires solar irradiance historical data for training. Furthermore, our approach is able to deal with missing data since it includes a data imputation state. In the prediction stage, we consider four data-driven approaches: Autoregressive Integrated Moving Average …
Single layer Feedforward Neural Network(FNN) is used many a time as a last layer in models such as seq2seq or could be a simple RNN network. The importance of such layer is to transform the output to our required dimensions. When it comes to weights and biases initialization, there is no such specific technique that could speed up the learning process. We could depend on deep network initialization techniques such as Xavier or He initialization. But such initialization fails to show much improvement in learning speed or accuracy. In this paper we propose Zero Initialization (ZI) for weights of…
Inverted AI；University of British Columbia（不列颠哥伦比亚大学）；Mila
We develop a deep generative model built on a fully differentiable simulator for multi-agent trajectory prediction. Agents are modeled with conditional recurrent variational neural networks (CVRNNs), which take as input an ego-centric birdview image representing the current state of the world and output an action, consisting of steering and acceleration, which is used to derive the subsequent agent state using a kinematic bicycle model. The full simulation state is then differentiably rendered for each agent, initiating the next time step. We achieve state-of-the-art results on the INTERACTION…
Wind is slated to become one of the most sought after source of energy in future. Both onshore as well as offshore wind farms are getting deployed rapidly over the world. This paper evaluates a neural network based time series approach to predict wind speed in real time over shorter duration of up to 12 hr based on analysis of three hourly wind data collected through a wave rider buoy deployed off Goa in deep water and far away from the shore. The data were collected for 4 years from February 1998 to February 2002. A simple feed forward type of network trained using a variety of algorithms was…
Due to expensive infrastructure and the difficulties in storage, supply conditions of natural gas are different from those of other traditional energy sources like petroleum or coal. To overcome these challenges, supplier countries require take-or-pay agreements for requested natural gas quantities. These contracts have many pre-clauses; if they are not met due to low/high consumption or other external factors, buyers must completely fulfill them. A similar contract is then imposed on distributors and wholesale consumers. It is thus important for all parties to forecast their daily, monthly, a…
The University of Adelaide（阿德莱德大学）；Sapienza University of Rome（罗马大学）；The Australian Institute for Machine Learning（澳大利亚机器学习研究所）；KTH Royal Institute of Technology（瑞典皇家理工学院）
Accurate short-term wind speed forecasting is essential for large-scale integration of wind power generation. However, the seasonal and stochastic characteristics of wind speed make forecasting a challenging task. This study uses a new hybrid evolutionary approach that uses a popular evolutionary search algorithm, CMA-ES, to tune the hyper-parameters of two Long short-term memory(LSTM) ANN models for wind prediction. The proposed hybrid approach is trained on data gathered from an offshore wind turbine installed in a Swedish wind farm located in the Baltic Sea. Two forecasting horizons includi…
Institute of Geosciences and Exact Sciences（地球科学与精确科学学院）
Navigation and mobility are some of the major problems faced by visually impaired people in their daily lives. Advances in computer vision led to the proposal of some navigation systems. However, most of them require expensive and/or heavy hardware. In this paper we propose the use of convolutional neural networks (CNN), transfer learning, and semi-supervised learning (SSL) to build a framework aimed at the visually impaired aid. It has low computational costs and, therefore, may be implemented on current smartphones, without relying on any additional equipment. The smartphone camera can be us…
IDSIA Dalle Molle Institute for Artificial Intelligence（IDSIA达勒莫勒人工智能研究所）；USI-SUPSI（瑞士意大利语区大学-瑞士南部应用科学与艺术大学）
This paper presents a linear dynamical operator described in terms of a rational transfer function, endowed with a well-defined and efficient back-propagation behavior for automatic derivatives computation. The operator enables end-to-end training of structured networks containing linear transfer functions and other differentiable units {by} exploiting standard deep learning software. Two relevant applications of the operator in system identification are presented. The first one consists in the integration of {prediction error methods} in deep learning. The dynamical operator is included as {t…
Montreal Institute for Learning Algorithms (MILA)（蒙特利尔学习算法研究所（MILA））
Online speech recognition is crucial for developing natural human-machine interfaces. This modality, however, is significantly more challenging than off-line ASR, since real-time/low-latency constraints inevitably hinder the use of future information, that is known to be very helpful to perform robust predictions. A popular solution to mitigate this issue consists of feeding neural acoustic models with context windows that gather some future frames. This introduces a latency which depends on the number of employed look-ahead features. This paper explores a different approach, based on estimati…
Columbia University（哥伦比亚大学）；New York University（纽约大学）
Speaker diarization is a task to label an audio or video recording with the identity of the speaker at each given time stamp. In this work, we propose a novel machine learning framework to conduct real-time multi-speaker diarization and recognition without prior registration and pretraining in a fully online and reinforcement learning setting. Our framework combines embedding extraction, clustering, and resegmentation into the same problem as an online decision-making problem. We discuss practical considerations and advanced techniques such as the offline reinforcement learning, semi-supervisi…
