Akihito Sudo

Associative Memory for Online Learning in Noisy Environments Using Self-Organizing Incremental Neural Network

Associative memory operating in a real environment must perform well in online incremental learning and be robust to noisy data because noisy associative patterns are presented sequentially in a real environment. We propose a novel associative memory that satisfies these requirements. Using the proposed method, new associative pairs that are presented sequentially can be learned accurately without forgetting previously learned patterns. The memory size of the proposed method increases adaptively with learning patterns. Therefore, it suffers neither redundancy nor insufficiency of memory size, even in an environment in which the maximum number of associative pairs to be presented is unknown before learning. Noisy inputs in real environments are classifiable into two types: noise-added original patterns and faultily presented random patterns. The proposed method deals with two types of noise. To our knowledge, no conventional associative memory addresses noise of both types. The proposed associative memory performs as a bidirectional one-to-many or many-to-one associative memory and deals not only with bipolar data, but also with real-valued data. Results demonstrate that the proposed methods features are important for application to an intelligent robot operating in a real environment. The originality of our work consists of two points: employing a growing self-organizing network for an associative memory, and discussing what features are necessary for an associative memory for an intelligent robot and proposing an associative memory that satisfies those requirements.

An online incremental learning pattern-based reasoning system

An architecture for reasoning with pattern-based if-then rules is proposed. By processing patterns as real-valued vectors and classifying similar if-then rules into clusters in long-term memory, the proposed system can store pattern-based if-then rules of propositional logic, including conjunctions, disjunctions, and negations. Moreover, it achieves some important properties for intelligent systems such as incremental learning, generalization, avoidance of duplicate results, and robustness to noise. Results of experiments demonstrate that the proposed method is effective for intelligent systems for solving various tasks autonomously in a real environment.

Associative Memory for Online Incremental Learning in Noisy Environments

Associative memory operating in a real environment must perform well on online incremental learning and be robust to noisy data because noisy associative patterns are presented sequentially in a real environment. We propose a novel associative memory that satisfies these needs. Using the proposed method, new associative pairs that are presented sequentially can be learned accurately without forgetting previously learned patterns. The memory size of the proposed method increases adaptively when learning patterns. Therefore, it suffers neither redundancy nor insufficiency of memory size, even in an environment where the maximum number of associative pairs to be presented is unknown before learning. The proposed method deals with two types of noise. To our knowledge, no conventional associative memory deals with both types. The proposed associative memory performs as a bidirectional one-to-many or many-to-one associative memory and deals not only with bipolar data, but also real-valued data. We infer that the proposed methods features are important for application to an intelligent robot operating in a real environment.

Real-time people movement estimation in large disasters from several kinds of mobile phone data

Recently, an understanding of mass movement in urban areas immediately after large disasters, such as the Great East Japan Earthquake (GEJE), has been needed. In particular, mobile phone data is available as time-varying data. However, much more detailed movement that is based on network flow instead of aggregated data is needed for appropriate rescue on a real-time basis. Hence, our research aims to estimate real-time human movement during large disasters from several kinds of mobile phone data. In this paper, we simulate the movement of people in the Tokyo metropolitan area in a large disaster situation and obtain several kinds of fragmentary movement observation data from mobile phones. Our approach is to use data assimilation techniques combining with simulation of population movement and observation data. The experimental results confirm that the improvement in accuracy depends on the observation data quality using sensitivity analysis and data processing speed to satisfy each condition for real-time estimation.

Particle filter for real-time human mobility prediction following unprecedented disaster

Real-time estimation of human mobility following a massive disaster will play a crucial role in disaster relief. Because human mobility in massive disasters is quite different from their usual mobility, real-time human location data is necessary for precise estimation. Due to privacy concerns, real-time data is anonymized and a popular form of anonymization is population distribution. In this paper, we aim to estimate human mobility following an unprecedented disaster using such population distribution data. To overcome technical obstacles including high dimensionality, we propose novel particle filter by devising proposal distribution. Our proposal distribution provides states considering both prediction model and acquired observation. Therefore, particles maintain high likelihood. In the experiments, our methods realized more accurate estimation than the baselines, and its estimated mobility was consistent with the survey researches. The computational cost is significantly low enough for real-time operations. The GPS data collected on the day of the Great East Japan Earthquake is used for the evaluation.

A framework for evacuation hotspot detection after large scale disasters using location data from smartphones: case study of Kumamoto earthquake

Large scale disasters cause severe social disorder and trigger mass evacuation activities. Managing the evacuation shelters efficiently is crucial for disaster management. Kumamoto prefecture, Japan, was hit by an enormous (Magnitude 7.3) earthquake on 16th of April, 2016. As a result, more than 10,000 buildings were severely damaged and over 100,000 people had to evacuate from their homes. After the earthquake, it took the decision makers several days to grasp the locations where people were evacuating, which delayed of distribution of supply and rescue. This situation was made even more complex since some people evacuated to places that were not designated as evacuation shelters. Conventional methods for grasping evacuation hotspots require on-foot field surveys that take time and are difficult to execute right after the hazard in the confusion. We propose a novel framework to efficiently estimate the evacuation hotspots after large disasters using location data collected from smartphones. To validate our framework and show the useful analysis using our output, we demonstrated the framework on the Kumamoto earthquake using GPS data of smartphones collected by Yahoo Japan. We verified that our estimation accuracy of evacuation hotspots were very high by checking the located facilities and also by comparing the population transition results with newspaper reports. Additionally, we demonstrated analysis using our framework outputs that would help decision makers, such as the population transition and function period of each hotspot. The efficiency of our framework is also validated by checking the processing time, showing that it could be utilized efficiently in disasters of any scale. Our framework provides useful output for decision makers that manage evacuation shelters after various kinds of large scale disasters.

Pattern-Based Reasoning System Using Self-incremental Neural Network for Propositional Logic

We propose an architecture for reasoning with pattern-based if-then rules that is effective for intelligent systems like robots solving varying tasks autonomously in a real environment. The proposed system can store pattern-based if-then rules of propositional logic, including conjunctions, disjunctions, negations, and implications. The naive pattern-based reasoning can store pattern-based if-then rules and make inferences using them. However, it remains insufficient for intelligent systems operating in a real environment. The proposed system uses an algorithm that is inspired by self-incremental neural networks such as SONIN and SOINN-AM in order to achieve incremental learning, generalization, avoidance of duplicate results, and robustness to noise, which are important properties for intelligent systems

PredictGIS 2017 workshop report held in conjunction with ACM SIGSPATIAL 2017

The prediction of human and vehicle mobility in a city is becoming attracting field. This topic attracts researchers in broad field from the behavioral science, where understanding the complexity of the human mobility behavior is one of the hot topic, to industrial field, which apply the result to many beneficial applications. Recent progress to sensing human mobility via smartphones is boosting this trend. However, due to the complexity and context-dependence of human behavior and the incompleteness and noise of geospatial data collecting from various sensors, the prediction of human and vehicle mobility is still far from solved. This workshop aimed at collecting contributions on the cutting-edge studies in human mobility description, modeling, intelligent computational method which can advance the human and vehicle prediction research. Potential topics included, but were not limited to 1) The next location prediction of individual mobility, 2) The crowd or population mobility prediction, 3) Dynamics of pedestrians, 4) commute flow and migration flow, 5) Traffic congestion, road usage forecast and optimal vehicle routing, 6) Social event forecast using geospatial data, 7) Novel agent mobility simulators, and 8) Case studies of mobility estimation in academia as well as in industrial field.

Predicting Indoor Crowd Density using Column-Structured Deep Neural Network

This work proposes a deep neural network approach known as the column-structured deep neural network (COL-DNN-R) for predicting crowd density in an indoor environment using historical Wi-Fi traces of individual visitors. With a structure designed to minimize feature engineering, COL-DNN accepts raw features such as crowd density, opening and closing hours and peak visitor counts for extracting features. The extracted features are used by a regression model R for predicting the crowd densities. Standard regression models such as MLP, RF and SVM can be used as R. Experiments are performed to investigate the effect of feature representation and model structure on the prediction accuracy. Experiment results show the best prediction accuracy is obtained using features extracted by COL-DNN and using MLP as the regression model, i.e., R = MLP.