Humans can learn the use of language through physical interaction with their environment and semiotic communication other people. It is very important to obtain a computational understanding how humans form symbol system skills autonomous mental development. Recently, many studies have been conducted on construction robotic systems machine-learning methods that embodied multimodal systems. Understanding human social interactions developing robot smoothly communicate users in long term,...

In this paper, we propose a visualization method for driving behavior that helps people to recognize distinctive patterns in continuous data. Driving can be measured using various types of sensors connected control area network. The multi-dimensional time series data are called many cases, each dimension the is not independent other statistical sense. For example, accelerator opening rate and longitudinal acceleration mutually dependent. We hypothesize only small number hidden features...

Humans use signs, e.g., sentences in a spoken language, for communication and thought. Hence, symbol systems like language are crucial our with other agents adaptation to real-world environment. The we human society adaptively dynamically change over time. In the context of artificial intelligence (AI) cognitive systems, grounding problem has been regarded as one central problems related {\it symbols}. However, was originally posed connect symbolic AI sensorimotor information did not...

Understanding information processing in the brain-and creating general-purpose artificial intelligence-are long-standing aspirations of scientists and engineers worldwide. The distinctive features human intelligence are high-level cognition control various interactions with world including self, which not defined advance vary over time. challenge building human-like intelligent machines, as well progress brain science behavioural analyses, robotics, their associated theoretical...

Creating autonomous robots that can actively explore the environment, acquire knowledge and learn skills continuously is ultimate achievement envisioned in cognitive developmental robotics. Importantly, if aim to create develop through interactions with their learning processes should be based on physical social world manner of human development. Based this context, paper, we focus two concepts models predictive coding. Recently, have attracted renewed attention as a topic considerable...

Constructive studies on symbol emergence systems seek to investigate computational models that can better explain human language evolution, the creation of systems, and construction internal representations. Specifically, emergent communication aims formulate a model enables agents build efficient sign This study provides new for communication, which is based probabilistic generative (PGM) instead discriminative deep reinforcement learning. We define Metropolis-Hastings (MH) naming game by...

Deep generative models (DGM) are increasingly employed in emergent communication systems. However, their application multimodal data contexts is limited. This study proposes a novel model that combines DGM with the Metropolis-Hastings (MH) naming game, enabling two agents to focus jointly on shared subject and develop common vocabularies. The proves it can handle data, even cases of missing modalities. Integrating MH game variational autoencoders (VAE) allows form perceptual categories...

A sequence prediction method for driving behavior data is proposed in this paper. The can predict a longer latent state of than conventional methods. derived by focusing on the double articulation structure latently embedded data. two-layer hierarchical originally found spoken language, i.e., sentence words and word letters. Analogously, we assume that comprise obtained extending nonparametric Bayesian unsupervised morphological analyzer using nested Pitman-Yor language model (NPYLM), which...

To realize human-like robot intelligence, a large-scale cognitive architecture is required for robots to understand the environment through variety of sensors with which they are equipped. In this paper, we propose novel framework named Serket that enables construction generative model and its inference easily by connecting sub-modules allow acquire various capabilities interaction their environments others. We consider models can be constructed smaller fundamental hierarchically while...

This paper describes a framework for the development of an integrative cognitive system based on probabilistic generative models (PGMs) called Neuro-SERKET. Neuro-SERKET is extension SERKET, which can compose elemental PGMs developed in distributed manner and provide scheme that allows composed to learn throughout unsupervised way. In addition head-to-tail connection supported by supports tail-to-tail head-to-head connections, as well neural network-based modules, i.e., deep models. As...

The understanding and acquisition of a language in real-world environment is an important task for future robotics services. Natural processing cognitive have both been focusing on the problem decades using machine learning. However, many problems remain unsolved despite significant progress learning (such as deep probabilistic generative models) during past decade. remaining not systematically surveyed organized, most them are highly interdisciplinary challenges robotics. This study...

Atrial fibrillation is a clinically important arrhythmia. There are some reports on machine learning models for AF diagnosis using electrocardiogram data. However, few have proposed an eXplainable Artificial Intelligence (XAI) model to enable physicians easily understand the model's results.

Building a human-like integrative artificial cognitive system, that is, an general intelligence (AGI), is the holy grail of (AI) field. Furthermore, computational model enables system to achieve development will be excellent reference for brain and science. This paper describes approach develop architecture by integrating elemental modules enable training as whole. based on two ideas: (1) brain-inspired AI, learning human build human-level intelligence, (2) probabilistic generative...

Understanding the emergence of symbol systems, especially language, requires construction a computational model that reproduces both developmental learning process in everyday life and evolutionary dynamics throughout history. This study introduces collective predictive coding (CPC) hypothesis, which emphasizes models interdependence between forming internal representations through physical interactions with environment sharing utilizing meanings social semiotic within system. The total...

Humans develop their concept of an object by classifying it into a category, and acquire language interacting with others at the same time. Thus, meaning word can be learnt connecting recognized concept. We consider such ability to important in allowing robots flexibly knowledge concepts. Accordingly, we propose method that enables knowledge. The is formed multimodal information acquired from objects, model human speech describing features. stochastic concepts, estimating parameters. point...

An unsupervised learning method, called double articulation analyzer with temporal prediction (DAA-TP), is proposed on the basis of original DAA model. The method will enable future advanced driving assistance systems to determine context and predict possible scenarios behavior by segmenting modeling incoming driving-behavior time series data. In previous studies, we applied model data argued that contextual changing points can be estimated as chunks. A sequence which predicts next hidden...

Human–robot interaction during general service tasks in home or retail environment has been proven challenging, partly because (1) robots lack high-level context-based cognition and (2) humans cannot intuit the perception state of as they can for other humans. To solve these two problems, we present a complete robot system that given highest evaluation score at Customer Interaction Task Future Convenience Store Challenge World Robot Summit 2018, which implements several key technologies:...

Due to their flexibility, soft-bodied robots can potentially achieve rich and various behaviors within a single body. However, date, no methodology has effectively harnessed these such diverse desired functionalities. Controllers that accomplish only limited range of in have been handcrafted. Moreover, the should be determined through body–environment interactions because an appropriate behavior may not always manifested even if body dynamics are given. Therefore, we proposed SenseCPG-PGPE,...

In the present paper, we propose a decoder-free extension of Dreamer, leading model-based reinforcement learning (MBRL) method from pixels. Dreamer is sample- and cost-efficient solution to robot learning, as it used train latent state-space models based on variational autoencoder conduct policy optimization by trajectory imagination. However, this autoencoding approach often causes object vanishing, in which fails perceives key objects for solving control tasks, thus significantly limiting...

In this paper, we propose an online algorithm for multimodal categorization based on the autonomously acquired information and partial words given by human users. For concept formation, latent Dirichlet allocation (MLDA) using Gibbs sampling is extended to version. We introduce a particle filter, which significantly improve performance of MLDA, keep tracking good models among various with different parameters. also unsupervised word segmentation method hierarchical Pitman-Yor Language Model...

In this paper, we propose a novel semiotic prediction method for driving behavior based on double articulation structure. It has been reported that predicting from its multivariate time series data by using machine learning methods, e.g., hybrid dynamical system, hidden Markov model and Gaussian mixture model, is difficult because driver's affected various contextual information. To overcome problem, assume information structure develop extending nonparametric Bayesian unsupervised...

In this paper, we propose an online learning algorithm based on a Rao-Blackwellized particle filter for spatial concept acquisition and mapping. We have proposed nonparametric Bayesian model (SpCoA). novel method (SpCoSLAM) integrating SpCoA FastSLAM in the theoretical framework of generative model. The can simultaneously learn place categories lexicons while incrementally generating environmental map. Furthermore, has scene image features language added to SpCoA. experiments, tested...