Japanese

Read about R&D activities at NTT Laboratories by clicking on the following links to this year's R&D Annual Report.


“corevo” —Evolving the Generation of New Values

AI supporting people

recommend-mark : recommended exhibits

  • B-5
  • Scalable knowledge discovery from large-scale “relationships”
  • pdf
  • Efficient parallel graph processing toward large-scale graph mining
  • We are developing a fundamental technology that enables large-scale graph mining of connected data elastically in accordance with the amount of data. We aim at on-demand traffic control and timely decision making through the high-speed analysis of ever-growing data.
  • recommend-markB-6
  • Network AI that finds useful devices from somewhere and connects them with you
  • pdf
  • Tacit computing : territorially-distributed resource utilization AI by tacit computing
  • We aim to share and use a variety of devices connected to a network with tacit computing. Network AI understands the value of everyone's devices all over the world. Then, it finds useful devices for you at that moment and connects them with you.
  • B-7
  • Predictive automatic control proactively prevents QoE degradation
  • pdf
  • Coordinated prediction and control techniques enable proactive network operation
  • corevo technology is applied to various types of data acquired inside/outside of networks such as network resources, traffic, communication quality, and event information. With the detection/analysis of potential risks, such as failures and congestion, preliminary autonomous control, and early automatic restoration, we realize network services that do not degrade QoE.
  • B-8
  • Early detection of network anomalies with NW-AI
  • pdf
  • Deep learning based network anomaly detection
  • To detect network anomalies such as silent failures in advance, we developed a deep-learning based technology as a part of our research on applying AI technology to solve network issues (NW-AI). In this demonstration, we show that we can detect rare network failures, which are difficult to detect with existing methods, in advance of customer calls.
  • B-9
  • Enabling easy localization of ongoing/potential failures by adaptively changing rules of monitoring and control on network equipment
  • pdf
  • Maintenance-free-oriented network technologies for proactive maintenance
  • Our engine effectively collects various types of data from network devices and sensors, enabling autonomous network monitoring and control. By adopting AI to cooperate with this engine, autonomous diagnosis done depending on the status of network/environment can identify ongoing or potential failures without human operation, which reduces OPEX.
  • B-10
  • Large-scale 3D point cloud analysis technology enables construction of real-world database
  • pdf
  • Enhanced assistance of equipment maintenance with space-state estimation technology
  • Our aim is to achieve technology for space-state estimation that enables automatic detection of infrastructure equipment and estimation of deterioration to support inspection operations and maintenance that need visual inspection by experts. We present the technology to estimate the state of objects by detecting specific objects with 3D point cloud and image analyses and superimposing point cloud data on images.