Model-free tracking control of complex dynamical trajectories with machine learning

Nonlinear tracking control enabling a dynamical system to track a desired trajectory is fundamental to robotics, serving a wide range of civil and defense applications. In control engineering, designing tracking control requires complete knowledge of the system model and equations. We develop a model-free, machine-learning framework to control a two-arm robotic manipulator using only partially observed states, where the controller is realized by reservoir computing. Stochastic input is exploited for training, which consists of the observed partial state vector as the first and its immediate future as the second component so that the neural machine regards the latter as the future state of the former. In the testing (deployment) phase, the immediate-future component is replaced by the desired observational vector from the reference trajectory. We demonstrate the effectiveness of the control framework using a variety of periodic and chaotic signals, and establish its robustness against measurement noise, disturbances, and uncertainties.

Decision-making based on an improved visual analytics approach for emotion prediction

Visual Analytics approach allows driving informed and effective decision-making. It assists decision-makers to visually interact with large amount of data and to computationally learn valuable hidden patterns in that data, which improve the decision quality. In this article, we introduce an enhanced visual analytics model combining cognitive-based visual analysis to data mining-based automatic analysis. As emotions are strongly related to human behaviour and society, emotion prediction is widely considered by decision making activities. Unlike speech and facial expressions modalities, EEG (electroencephalogram) has the advantage of being able to record information about the internal emotional state that is not always translated by perceptible external manifestations. For this reason, we applied the proposed cognitive approach on EEG data to demonstrate its efficiency for predicting emotional reaction to films. For automatic analysis, we developed the Echo State Network (ESN) technique considered as an efficient machine learning solution due to its straightforward training procedure and high modelling ability for handling time-series problems. Finally, utility and usability tests were performed to evaluate the developed prototype.

Computational Modeling of Structural Synaptic Plasticity in Echo State Networks

Most existing studies on computational modeling of neural plasticity have focused on synaptic plasticity. However, regulation of the internal weights in the reservoir based on synaptic plasticity often results in unstable learning dynamics. In this article, a structural synaptic plasticity learning rule is proposed to train the weights and add or remove neurons within the reservoir, which is shown to be able to alleviate the instability of the synaptic plasticity, and to contribute to increase the memory capacity of the network as well. Our experimental results also reveal that a few stronger connections may last for a longer period of time in a constantly changing network structure, and are relatively resistant to decay or disruptions in the learning process. These results are consistent with the evidence observed in biological systems. Finally, we show that an echo state network (ESN) using the proposed structural plasticity rule outperforms an ESN using synaptic plasticity and three state-of-the-art ESNs on four benchmark tasks.

Echo Memory-Augmented Network for time series classification

Echo State Networks (ESNs) are efficient recurrent neural networks (RNNs) which have been successfully applied to time series modeling tasks. However, ESNs are unable to capture the history information far from the current time step, since the echo state at the present step of ESNs mostly impacted by the previous one. Thus, ESN may have difficulty in capturing the long-term dependencies of temporal data. In this paper, we propose an end-to-end model named Echo Memory-Augmented Network (EMAN) for time series classification. An EMAN consists of an echo memory-augmented encoder and a multi-scale convolutional learner. First, the time series is fed into the reservoir of an ESN to produce the echo states, which are all collected into an echo memory matrix along with the time steps. After that, we design an echo memory-augmented mechanism employing the sparse learnable attention to the echo memory matrix to obtain the Echo Memory-Augmented Representations (EMARs). In this way, the input time series is encoded into the EMARs with enhancing the temporal memory of the ESN. We then use multi-scale convolutions with the max-over-time pooling to extract the most discriminative features from the EMARs. Finally, a fully-connected layer and a softmax layer calculate the probability distribution on categories. Experiments conducted on extensive time series datasets show that EMAN is state-of-the-art compared to existing time series classification methods. The visualization analysis also demonstrates the effectiveness of enhancing the temporal memory of the ESN.

Recurrent Broad Learning Systems for Time Series Prediction

The broad learning system (BLS) is an emerging approach for effective and efficient modeling of complex systems. The inputs are transferred and placed in the feature nodes, and then sent into the enhancement nodes for nonlinear transformation. The structure of a BLS can be extended in a wide sense. Incremental learning algorithms are designed for fast learning in broad expansion. Based on the typical BLSs, a novel recurrent BLS (RBLS) is proposed in this paper. The nodes in the enhancement units of the BLS are recurrently connected, for the purpose of capturing the dynamic characteristics of a time series. A sparse autoencoder is used to extract the features from the input instead of the randomly initialized weights. In this way, the RBLS retains the merit of fast computing and fits for processing sequential data. Motivated by the idea of "fine-tuning" in deep learning, the weights in the RBLS can be updated by conjugate gradient methods if the prediction errors are large. We exhibit the merits of our proposed model on several chaotic time series. Experimental results substantiate the effectiveness of the RBLS. For chaotic benchmark datasets, the RBLS achieves very small errors, and for the real-world dataset, the performance is satisfactory.

Time series classification with Echo Memory Networks

Echo state networks (ESNs) are randomly connected recurrent neural networks (RNNs) that can be used as a temporal kernel for modeling time series data, and have been successfully applied on time series prediction tasks. Recently, ESNs have been applied to time series classification (TSC) tasks. However, previous ESN-based classifiers involve either training the model by predicting the next item of a sequence, or predicting the class label at each time step. The former is essentially a predictive model adapted from time series prediction work, rather than a model designed specifically for the classification task. The latter approach only considers local patterns at each time step and then averages over the classifications. Hence, rather than selecting the most discriminating sections of the time series, this approach will incorporate non-discriminative information into the classification, reducing accuracy. In this paper, we propose a novel end-to-end framework called the Echo Memory Network (EMN) in which the time series dynamics and multi-scale discriminative features are efficiently learned from an unrolled echo memory using multi-scale convolution and max-over-time pooling. First, the time series data are projected into the high dimensional nonlinear space of the reservoir and the echo states are collected into the echo memory matrix, followed by a single multi-scale convolutional layer to extract multi-scale features from the echo memory matrix. Max-over-time pooling is used to maintain temporal invariance and select the most important local patterns. Finally, a fully-connected hidden layer feeds into a softmax layer for classification. This architecture is applied to both time series classification and human action recognition datasets. For the human action recognition datasets, we divide the action data into five different components of the human body, and propose two spatial information fusion strategies to integrate the spatial information over them. With one training-free recurrent layer and only one layer of convolution, the EMN is a very efficient end-to-end model, and ranks first in overall classification ability on 55 TSC benchmark datasets and four 3D skeleton-based human action recognition tasks.

Hybrid Regularized Echo State Network for Multivariate Chaotic Time Series Prediction

Multivariate chaotic time series prediction is a hot research topic, the goal of which is to predict the future of the time series based on past observations. Echo state networks (ESNs) have recently been widely used in time series prediction, but there may be an ill-posed problem for a large number of unknown output weights. To solve this problem, we propose a hybrid regularized ESN, which employs a sparse regression with the L₁/2 regularization and the L₂ regularization to compute the output weights. The L₁/2 penalty shows many attractive properties, such as unbiasedness and sparsity. The L₂ penalty presents appealing ability on shrinking the amplitude of the output weights. After the output weights are calculated, the input weights, internal weights, and output weights are fine-tuning by a Hessian-free optimization method-conjugate gradient backpropagation algorithm. The fine-tuning helps to bubble up the input information toward the output layer. Besides, the largest Lyapunov exponent is used to calculate the predictable horizon of a chaotic time series. Experimental results on benchmark and real-world datasets show that our proposed method is superior to other ESN-based models, as sparser, smaller-absolute-value, and more informative output weights are obtained. All of the predictions within the predictable horizon of the proposed model are accurate.

Spatio-Temporal Interpolated Echo State Network for Meteorological Series Prediction

Spatio-temporal series prediction has attracted increasing attention in the field of meteorology in recent years. The spatial and temporal joint effect makes predictions challenging. Most of the existing spatio-temporal prediction models are computationally complicated. To develop an accurate but easy-to-implement spatio-temporal prediction model, this paper designs a novel spatio-temporal prediction model based on echo state networks. For real-world observed meteorological data with randomness and large changes, we use a cubic spline method to bridge the gaps between the neighboring points, which results in a pleasingly smooth series. The interpolated series is later input into the spatio-temporal echo state networks, in which the spatial coefficients are computed by the elastic-net algorithm. This approach offers automatic selection and continuous shrinkage of the spatial variables. The proposed model provides an intuitive but effective approach to address the interaction of spatial and temporal effects. To demonstrate the practicality of the proposed model, we apply it to predict two real-world datasets: monthly precipitation series and daily air quality index series. Experimental results demonstrate that the proposed model achieves a normalized root-mean-square error of approximately 0.250 on both datasets. Similar results are achieved on the long short-term memory model, but the computation time of our proposed model is considerably shorter. It can be inferred that our proposed neural network model has advantages on predicting meteorological series over other models.

Data-driven forecasting of high-dimensional chaotic systems with long short-term memory networks

We introduce a data-driven forecasting method for high-dimensional chaotic systems using long short-term memory (LSTM) recurrent neural networks. The proposed LSTM neural networks perform inference of high-dimensional dynamical systems in their reduced order space and are shown to be an effective set of nonlinear approximators of their attractor. We demonstrate the forecasting performance of the LSTM and compare it with Gaussian processes (GPs) in time series obtained from the Lorenz 96 system, the Kuramoto-Sivashinsky equation and a prototype climate model. The LSTM networks outperform the GPs in short-term forecasting accuracy in all applications considered. A hybrid architecture, extending the LSTM with a mean stochastic model (MSM-LSTM), is proposed to ensure convergence to the invariant measure. This novel hybrid method is fully data-driven and extends the forecasting capabilities of LSTM networks.

Laplacian Echo State Network for Multivariate Time Series Prediction

Echo state network is a novel kind of recurrent neural networks, with a trainable linear readout layer and a large fixed recurrent connected hidden layer, which can be used to map the rich dynamics of complex real-world data sets. It has been extensively studied in time series prediction. However, there may be an ill-posed problem caused by the number of real-world training samples less than the size of the hidden layer. In this brief, a Laplacian echo state network (LAESN), is proposed to overcome the ill-posed problem and obtain low-dimensional output weights. First, an echo state network is used to map the multivariate time series into a large reservoir. Then, assuming that an unknown underlying manifold is inside the reservoir, we employ the Laplacian eigenmaps to estimate the manifold by constructing an adjacency graph associated with the reservoir states. Finally, the output weights are calculated by the low-dimensional manifold. In addition, some criteria of transient stability, local controllability, and local observability are given. Experimental results based on two real-world data sets substantiate the effectiveness and characteristics of the proposed LAESN model.

Action perception as hypothesis testing

We present a novel computational model that describes action perception as an active inferential process that combines motor prediction (the reuse of our own motor system to predict perceived movements) and hypothesis testing (the use of eye movements to disambiguate amongst hypotheses). The system uses a generative model of how (arm and hand) actions are performed to generate hypothesis-specific visual predictions, and directs saccades to the most informative places of the visual scene to test these predictions - and underlying hypotheses. We test the model using eye movement data from a human action observation study. In both the human study and our model, saccades are proactive whenever context affords accurate action prediction; but uncertainty induces a more reactive gaze strategy, via tracking the observed movements. Our model offers a novel perspective on action observation that highlights its active nature based on prediction dynamics and hypothesis testing.

Multilayered Echo State Machine: A Novel Architecture and Algorithm

In this paper, we present a novel architecture and learning algorithm for a multilayered echo state machine (ML-ESM). Traditional echo state networks (ESNs) refer to a particular type of reservoir computing (RC) architecture. They constitute an effective approach to recurrent neural network (RNN) training, with the (RNN-based) reservoir generated randomly, and only the readout trained using a simple computationally efficient algorithm. ESNs have greatly facilitated the real-time application of RNN, and have been shown to outperform classical approaches in a number of benchmark tasks. In this paper, we introduce a novel criteria for integrating multiple layers of reservoirs within the ML-ESM. The addition of multiple layers of reservoirs are shown to provide a more robust alternative to conventional RC networks. We demonstrate the comparative merits of this approach in a number of applications, considering both benchmark datasets and real world applications.

You cannot speak and listen at the same time: a probabilistic model of turn-taking

Turn-taking is a preverbal skill whose mastering constitutes an important precondition for many social interactions and joint actions. However, the cognitive mechanisms supporting turn-taking abilities are still poorly understood. Here, we propose a computational analysis of turn-taking in terms of two general mechanisms supporting joint actions: action prediction (e.g., recognizing the interlocutor's message and predicting the end of turn) and signaling (e.g., modifying one's own speech to make it more predictable and discriminable). We test the hypothesis that in a simulated conversational scenario dyads using these two mechanisms can recognize the utterances of their co-actors faster, which in turn permits them to give and take turns more efficiently. Furthermore, we discuss how turn-taking dynamics depend on the fact that agents cannot simultaneously use their internal models for both action (or messages) prediction and production, as these have different requirements-or, in other words, they cannot speak and listen at the same time with the same level of accuracy. Our results provide a computational-level characterization of turn-taking in terms of cognitive mechanisms of action prediction and signaling that are shared across various interaction and joint action domains.

Sensorimotor Coarticulation in the Execution and Recognition of Intentional Actions

Humans excel at recognizing (or inferring) another's distal intentions, and recent experiments suggest that this may be possible using only subtle kinematic cues elicited during early phases of movement. Still, the cognitive and computational mechanisms underlying the recognition of intentional (sequential) actions are incompletely known and it is unclear whether kinematic cues alone are sufficient for this task, or if it instead requires additional mechanisms (e.g., prior information) that may be more difficult to fully characterize in empirical studies. Here we present a computationally-guided analysis of the execution and recognition of intentional actions that is rooted in theories of motor control and the coarticulation of sequential actions. In our simulations, when a performer agent coarticulates two successive actions in an action sequence (e.g., "reach-to-grasp" a bottle and "grasp-to-pour"), he automatically produces kinematic cues that an observer agent can reliably use to recognize the performer's intention early on, during the execution of the first part of the sequence. This analysis lends computational-level support for the idea that kinematic cues may be sufficiently informative for early intention recognition. Furthermore, it suggests that the social benefits of coarticulation may be a byproduct of a fundamental imperative to optimize sequential actions. Finally, we discuss possible ways a performer agent may combine automatic (coarticulation) and strategic (signaling) ways to facilitate, or hinder, an observer's action recognition processes.

Adaptive Elastic Echo State Network for Multivariate Time Series Prediction

Echo state network (ESN) is a new kind of recurrent neural network with a randomly generated reservoir structure and an adaptable linear readout layer. It has been widely employed in the field of time series prediction. However, when high-dimensional reservoirs are utilized to predict multivariate time series, there may be a collinearity problem. In this paper, to overcome the collinearity problem and obtain a sparse solution, we propose a new model-adaptive elastic ESN, in which adaptive elastic net algorithm is used to calculate the unknown weights. It combines the strengths of the quadratic regularization and the adaptively weighted lasso shrinkage. Hence, the proposed model can deal with the collinearity problem and enjoy the oracle property with an unbiased estimation. We exhibit the merits of our model on two benchmark multivariate chaotic datasets and two real-world applications. Experimental results substantiate the effectiveness and characteristics of the proposed model.

Quaternion-valued echo state networks

Quaternion-valued echo state networks (QESNs) are introduced to cater for 3-D and 4-D processes, such as those observed in the context of renewable energy (3-D wind modeling) and human centered computing (3-D inertial body sensors). The introduction of QESNs is made possible by the recent emergence of quaternion nonlinear activation functions with local analytic properties, required by nonlinear gradient descent training algorithms. To make QENSs second-order optimal for the generality of quaternion signals (both circular and noncircular), we employ augmented quaternion statistics to introduce widely linear QESNs. To that end, the standard widely linear model is modified so as to suit the properties of dynamical reservoir, typically realized by recurrent neural networks. This allows for a full exploitation of second-order information in the data, contained both in the covariance and pseudocovariances, and a rigorous account of second-order noncircularity (improperness), and the corresponding power mismatch and coupling between the data components. Simulations in the prediction setting on both benchmark circular and noncircular signals and on noncircular real-world 3-D body motion data support the analysis.

Spatio-temporal learning with the online finite and infinite echo-state Gaussian processes

Successful biological systems adapt to change. In this paper, we are principally concerned with adaptive systems that operate in environments where data arrives sequentially and is multivariate in nature, for example, sensory streams in robotic systems. We contribute two reservoir inspired methods: 1) the online echostate Gaussian process (OESGP) and 2) its infinite variant, the online infinite echostate Gaussian process (OIESGP) Both algorithms are iterative fixed-budget methods that learn from noisy time series. In particular, the OESGP combines the echo-state network with Bayesian online learning for Gaussian processes. Extending this to infinite reservoirs yields the OIESGP, which uses a novel recursive kernel with automatic relevance determination that enables spatial and temporal feature weighting. When fused with stochastic natural gradient descent, the kernel hyperparameters are iteratively adapted to better model the target system. Furthermore, insights into the underlying system can be gleamed from inspection of the resulting hyperparameters. Experiments on noisy benchmark problems (one-step prediction and system identification) demonstrate that our methods yield high accuracies relative to state-of-the-art methods, and standard kernels with sliding windows, particularly on problems with irrelevant dimensions. In addition, we describe two case studies in robotic learning-by-demonstration involving the Nao humanoid robot and the Assistive Robot Transport for Youngsters (ARTY) smart wheelchair.

Relative entropy minimizing noisy non-linear neural network to approximate stochastic processes

A method is provided for designing and training noise-driven recurrent neural networks as models of stochastic processes. The method unifies and generalizes two known separate modeling approaches, Echo State Networks (ESN) and Linear Inverse Modeling (LIM), under the common principle of relative entropy minimization. The power of the new method is demonstrated on a stochastic approximation of the El Niño phenomenon studied in climate research.

Human sensorimotor communication: a theory of signaling in online social interactions

Although the importance of communication is recognized in several disciplines, it is rarely studied in the context of online social interactions and joint actions. During online joint actions, language and gesture are often insufficient and humans typically use non-verbal, sensorimotor forms of communication to send coordination signals. For example, when playing volleyball, an athlete can exaggerate her movements to signal her intentions to her teammates (say, a pass to the right) or to feint an adversary. Similarly, a person who is transporting a table together with a co-actor can push the table in a certain direction to signal where and when he intends to place it. Other examples of "signaling" are over-articulating in noisy environments and over-emphasizing vowels in child-directed speech. In all these examples, humans intentionally modify their action kinematics to make their goals easier to disambiguate. At the moment no formal theory exists of these forms of sensorimotor communication and signaling. We present one such theory that describes signaling as a combination of a pragmatic and a communicative action, and explains how it simplifies coordination in online social interactions. We cast signaling within a "joint action optimization" framework in which co-actors optimize the success of their interaction and joint goals rather than only their part of the joint action. The decision of whether and how much to signal requires solving a trade-off between the costs of modifying one's behavior and the benefits in terms of interaction success. Signaling is thus an intentional strategy that supports social interactions; it acts in concert with automatic mechanisms of resonance, prediction, and imitation, especially when the context makes actions and intentions ambiguous and difficult to read. Our theory suggests that communication dynamics should be studied within theories of coordination and interaction rather than only in terms of the maximization of information transmission.

Nonparametric mixtures of gaussian processes with power-law behavior

Gaussian processes (GPs) constitute one of the most important Bayesian machine learning approaches, based on a particularly effective method for placing a prior distribution over the space of regression functions. Several researchers have considered postulating mixtures of GPs as a means of dealing with nonstationary covariance functions, discontinuities, multimodality, and overlapping output signals. In existing works, mixtures of GPs are based on the introduction of a gating function defined over the space of model input variables. This way, each postulated mixture component GP is effectively restricted in a limited subset of the input space. In this paper, we follow a different approach. We consider a fully generative nonparametric Bayesian model with power-law behavior, generating GPs over the whole input space of the learned task. We provide an efficient algorithm for model inference, based on the variational Bayesian framework, and prove its efficacy using benchmark and real-world datasets.

Chaotic time series prediction based on a novel robust echo state network

In this paper, a robust recurrent neural network is presented in a Bayesian framework based on echo state mechanisms. Since the new model is capable of handling outliers in the training data set, it is termed as a robust echo state network (RESN). The RESN inherits the basic idea of ESN learning in a Bayesian framework, but replaces the commonly used Gaussian distribution with a Laplace one, which is more robust to outliers, as the likelihood function of the model output. Moreover, the training of the RESN is facilitated by employing a bound optimization algorithm, based on which, a proper surrogate function is derived and the Laplace likelihood function is approximated by a Gaussian one, while remaining robust to outliers. It leads to an efficient method for estimating model parameters, which can be solved by using a Bayesian evidence procedure in a fully autonomous way. Experimental results show that the proposed method is robust in the presence of outliers and is superior to existing methods.

Nonlinear system modeling with random matrices: echo state networks revisited

Echo state networks (ESNs) are a novel form of recurrent neural networks (RNNs) that provide an efficient and powerful computational model approximating nonlinear dynamical systems. A unique feature of an ESN is that a large number of neurons (the "reservoir") are used, whose synaptic connections are generated randomly, with only the connections from the reservoir to the output modified by learning. Why a large randomly generated fixed RNN gives such excellent performance in approximating nonlinear systems is still not well understood. In this brief, we apply random matrix theory to examine the properties of random reservoirs in ESNs under different topologies (sparse or fully connected) and connection weights (Bernoulli or Gaussian). We quantify the asymptotic gap between the scaling factor bounds for the necessary and sufficient conditions previously proposed for the echo state property. We then show that the state transition mapping is contractive with high probability when only the necessary condition is satisfied, which corroborates and thus analytically explains the observation that in practice one obtains echo states when the spectral radius of the reservoir weight matrix is smaller than 1.