We need more than just deep data analysis (machine learning) techniques to make sense of the wide variety of information available in this age of big data: we also need to utilize databases, cloud computing, security methodologies, and other related technologies while closely collaborating with specialists in the areas where we choose to focus our efforts. At the Machine Learning and Data Science Center, we are engaged in the creation of new values through innovative multidisciplinary research and development projects related to big data analysis.
Machine learning techniques
- We run research and development projects focused on machine learning and data mining techniques that use large-scale data modeling, classification, latent structure analysis, and other techniques to analyze datasets, identify latent data, conduct causality analysis, discover signs of future trends, and interpolate data points.
Important applications related to NTT’s business
- We face important issues—and discover new ones—in response to the potential needs of the NTT group’s core commercial and technological competencies, which include network traffic, security, and customer relationship management (CRM).
Large-scale graph database technologies
- With the expansion of the web and social media, there has been an explosion in the amount of available graph data representing relationships between diverse data points including people, objects, and locations. To further improve and personalize services in these areas, we continue to pursue research that will allow us to establish large-scale graph database technologies for analyzing large-scale graph data at extremely high speeds.
Scalable technologies for distributed data processing with online machine learning
- From detecting damage caused by misinformation, errors, and failures to predicting markets, stock prices, and natural disasters, real-time data analysis is necessary to quickly respond to events in the wide variety of fields covered by big data analysis. We continue to research and develop Jubatus, an online machine learning framework, while integrating it with our machine learning techniques.