Recurrent Neural Network With Parametric Bias Research Articles

A Model of Spatial Cell Development in Rat Hippocampus Based on Artificial Neural Network.

Physiological studies have shown that the hippocampal structure of rats develops at different stages, in which the place cells continue to develop during the whole juvenile period of rats and mature after the juvenile period. As the main information source of place cells, grid cells should mature earlier than place cells. In order to make better use of the biological information exhibited by the rat brain hippocampus in the environment, we propose a position cognition model based on the spatial cell development mechanism of rat hippocampus. The model uses a recurrent neural network with parametric bias (RNNPB) to simulate changes in the discharge characteristics during the development of a single stripe cell. The oscillatory interference mechanism is able to fuse the developing stripe waves, thus indirectly simulating the developmental process of the grid cells. The output of the grid cells is then used as the information input of the place cells, whose development process is simulated by BP neural network. After the place cells matured, the position matrix generated by the place cell group was used to realize the position cognition of rats in a given spatial region. The experimental results show that this model can simulate the development process of grid cells and place cells, and it can realize high precision positioning in the given space area. Moreover, the experimental effect of cognitive map construction using this model is basically consistent with the effect of RatSLAM, which verifies the validity and accuracy of the model.

Learning for Goal-Directed Actions Using RNNPB: Developmental Change of “What to Imitate”

“What to imitate” is one of the most important and difficult issues in robot imitation learning. A possible solution from an engineering approach involves focusing on the salient properties of actions. We investigate the developmental change of what to imitate in robot action learning in this paper. Our robot is equipped with a recurrent neural network with parametric bias (RNNPB), and learned to imitate multiple goal-directed actions in two different environments (i.e., simulation and real humanoid robot). Our close analysis of the error measures and the internal representation of the RNNPB revealed that actions’ most salient properties (i.e., reaching the desired end of motor trajectories) were learned first, while the less salient properties (i.e., matching the shape of motor trajectories) were learned later. Interestingly, this result was analogous to the developmental process of human infant’s action imitation. We discuss the importance of our results in terms of understanding the underlying mechanisms of human development.

RGBD camera monitoring system for Alzheimer's disease assessment using Recurrent Neural Networks with Parametric Bias action recognition

The present paper proposes a computer vision system to diagnose the stage of illness in patients affected by Alzheimer's disease. In the context of Ambient Assisted Living (AAL), the system monitors people in home environment during daily personal care activities. The aim is to evaluate the dementia stage, observing actions listed in the Direct Assessment of Funcional Status (DAFS) index and detecting anomalies during the performance, in order to assign a score explaining if the action is correct or not. In this work brushing teeth and grooming hair by a hairbrush are analysed. The technology consists of the application of a Recurrent Neural Network with Parametric Bias (RNNPB) that is able to learn movements connected with a specific action and recognize human activities by parametric bias that work like mirror neurons. This study has been conducted using Microsoft Kinect to collect data about the actions observed and oversee the user tracking and gesture recognition. Experiments prove that the proposed computer vision system can learn and recognize complex human activities and evaluates DAFS score.

Multiple Action Sequence Learning and Automatic Generation for a Humanoid Robot Using RNNPB and Reinforcement Learning

This paper proposes how to learn and generate multiple action sequences of a humanoid robot. At first, all the basic action sequences, also called primitive behaviors, are learned by a recurrent neural network with parametric bias (RNNPB) and the value of the internal nodes which are parametric bias (PB) determining the output with different primitive behaviors are obtained. The training of the RNN uses back propagation through time (BPTT) method. After that, to generate the learned behaviors, or a more complex behavior which is the combination of the primitive behaviors, a reinforcement learning algorithm: Q-learning (QL) is adopt to determine which PB value is adaptive for the generation. Finally, using a real humanoid robot, the proposed method was confirmed its effectiveness by the results of experiment.

人工神経回路モデルと声道物理モデルを用いた母音模倣モデルに基づく音素獲得シミュレーション

This paper proposes a continuous vowel imitation system that explains the process of phoneme acquisition by infants from the dynamical systems perspective. Almost existing models concerning this process dealt with discrete phoneme sequences. Human infants, however, have no knowledge of phoneme innately. They perceive speech sounds as continuous acoustic signals. The imitation target of this study is continuous acoustic signals including unknown numbers and kinds of phonemes. The key ideas of the model are (1) the use of a physical vocal tract model called the Maeda model for embodying the motor theory of speech perception, (2) the use of a dynamical system called the Recurrent Neural Network with Parametric Bias (RNNPB) trained with both dynamics of the acoustic signals and articulatory movements of the Maeda model, and (3) the segmenting method of a temporal sequence using the prediction error of the RNNPB model. The experiments of our model demonstrated following results: (a) the self-organization of the vowel structure into attractors of RNNPB model, (b) the improvement of vowel imitation using movement of the Maeda model, and (c) the generation of clear vowels based on the bubbling process trained with a few random utterances. These results suggest that our model reflects the process of phoneme acquisition.

Self-organization of Dynamic Object Features Based on Bidirectional Training

This paper presents a method to self-organize object features that describe object dynamics using bidirectional training. The model is composed of a dynamics learning module and a feature extraction module. Recurrent Neural Network with Parametric Bias (RNNPB) is utilized for the dynamics learning module, learning and self-organizing the sequences of robot and object motions. A hierarchical neural network is linked to the input of RNNPB as the feature extraction module for self-organizing object features that describe the object motions. The two modules are simultaneously trained through bidirectional training using image and motion sequences acquired from the robot's active sensing with objects. Experiments are performed with the robot's pushing motion with a variety of objects to generate sliding, falling over, bouncing and rolling motions. The results have shown that the model is capable of self-organizing object dynamics based on the self-organized features.

Analysis of Reliable Predictability based Motion Generation using RNNPB

Reliable predictability, which is tightly connected to consistency of environmental changes, is one of the main factors that determine human behaviors. As a constructive approach to understanding this mechanism, the authors have developed a method to generate autonomous object pushing motions based on consistency of object motions using a humanoid robot. The method consists of constructing a dynamics prediction model using Recurrent Neural Network with Parametric Bias (RNNPB), and motion searching based on an object consistency evaluation function using Steepest Descent Method. First, RNNPB is trained using the observed object dynamics and robot motion sequences, acquired during active sensing with objects. Next, the Steepest Descent Method is applied for searching the reliably predictable motion through the constructed dynamics model. Finally, the obtained motion is linked to the initial object image using a hierarchical neural network. The model inputs the object image outputting the reliably predictable robot motion which induces consistent object motions. The model was analyzed through two experiments, pushing cylindrical objects with a humanoid robot. The analysis has shown the method's effectivity and limitations to generate consistent object motions. I. I NTRODUCTION

Predicting Object Dynamics From Visual Images Through Active Sensing Experiences

Prediction of dynamic features is an important task for determining the manipulation strategies of an object. This paper presents a technique for predicting dynamics of objects relative to the robot's motion from visual images. During the training phase, the authors use the recurrent neural network with parametric bias (RNNPB) to self-organize the dynamics of objects manipulated by the robot into the PB space. The acquired PB values, static images of objects and robot motor values are input into a hierarchical neural network to link the images to dynamic features (PB values). The neural network extracts prominent features that each induce object dynamics. For prediction of the motion sequence of an unknown object, the static image of the object and robot motor value are input into the neural network to calculate the PB values. By inputting the PB values into the closed loop RNNPB, the predicted movements of the object relative to the robot motion are calculated recursively. Experiments were conducted with the humanoid robot Robovie-IIs pushing objects at different heights. The results of the experiment predicting the dynamics of target objects proved that the technique is efficient for predicting the dynamics of the objects.

2P1-G03 Segmenting Sound Signals and Articulatory Movement using Recurrent Neural Network toward Phoneme Acquisition

This paper proposes a computational model for phoneme acquisition by infants. Infants perceive speech not as discrete phoneme sequences but as continuous acoustic signals. One of critical problems in phoneme acquisition is the design for segmenting these continuous speech. The key idea to solve this problem is that articulatory mechanisms such as the vocal tract help human beings to perceive sound units corresponding to phonemes. To segment acoustic signal with articulatory movement, our system was implemented by using a physical vocal tract model, called the Maeda model, and applying a segmenting method using Recurrent Neural Network with Parametric Bias (RNNPB). This method determines segmentation boundaries in a sequence using the prediction error of the RNNPB model, and the PB values obtained by the method can be encoded as kind of phonemes. Experimental results demonstrated that our system could self-organize the same phonemes in different continuous sounds. This suggests that our model reflects the process of phoneme acquisition.

Experience-based imitation using RNNPB

Robot imitation is a useful and promising alternative to robot programming. Robot imitation involves two crucial issues. The first is how a robot can imitate a human whose physical structure and properties differ greatly from its own. The second is how the robot can generate various motions from finite programmable patterns (generalization). This paper describes a novel approach to robot imitation based on its own physical experiences. We considered the target task of moving an object on a table. For imitation, we focused on an active sensing process in which the robot acquires the relation between the object's motion and its own arm motion. For generalization, we applied the RNNPB (recurrent neural network with parametric bias) model to enable recognition/generation of imitation motions. The robot associates the arm motion which reproduces the observed object's motion presented by a human operator. Experimental results proved the generalization capability of our method, which enables the robot to imitate not only motion it has experienced, but also unknown motion through nonlinear combination of the experienced motions.

Dynamic and interactive generation of object handling behaviors by a small humanoid robot using a dynamic neural network model

This study presents experiments on the learning of object handling behaviors by a small humanoid robot using a dynamic neural network model, the recurrent neural network with parametric bias (RNNPB). The first experiment showed that after the robot learned different types of ball handling behaviors using human direct teaching, the robot was able to generate adequate ball handling motor sequences situated to the relative position between the robot's hands and the ball. The same scheme was applied to a block handling learning task where it was shown that the robot can switch among learned different block handling sequences, situated to the ways of interaction by human supporters. Our analysis showed that entrainment of the internal memory structures of the RNNPB through the interactions of the objects and the human supporters are the essential mechanisms for those observed situated behaviors of the robot