November 2018 Vol. 16 No. 11

View from the Top

• Kazuhiro Yoshizawa, President and Chief Executive Officer, NTT DOCOMO

  

  Overview
  With its sights firmly on the 5G (fifth-generation mobile communications networks) era, NTT DOCOMO has achieved a gigabit-capable mobile network by increasing its effective download speed by 1.5 times over the previous year. Breaking away from its traditional image of a mobile and smartphone company, NTT DOCOMO has strived to improve its long-term corporate value and to create a safe and secure, comfortable, and richer society, for which it has gained a high reputation, including earning a No. 1 ranking in total score in the 2018 Toyo Keizai CSR (Corporate Social Responsibility) Ranking. We asked Kazuhiro Yoshizawa, President and Chief Executive Officer of NTT DOCOMO, about his outlook and aspirations based on the Medium-Term Strategy 2020 known as “Declaration beyond.”

Front-line Researchers

• Ryuichiro Higashinaka, Senior Distinguished Researcher, NTT Media Intelligence Laboratories

  

  Overview
  The spread of smartphones and artificial intelligence (AI) systems that speak means that the opportunities for general users to come into contact with AI are increasing. Users are now expecting improved ease of use and more new technologies from research and development. Under these circumstances, NTT has continuously gained attention―both domestically and globally―in announcing pioneering technologies in the fields of question answering functions and language processing. We visited Ryuichiro Higashinaka, Senior Distinguished Researcher at NTT Media Intelligence Laboratories, and asked him about the progress and outlook of his research, his attitude as a researcher, and how the development of AI and a future in which robots and humans can talk smoothly will bring about changes in our society.

Feature Articles: Further Exploring Communication Science

• Shift to New Dimensions―Further Initiatives to Deepen Communication Science

  

  Abstract
  NTT Communication Science Laboratories aims to realize communication that reaches the heart, from person to person, and between people and computers. We are building fundamental theories in pursuit of the essence of people and of information, and working to create core technologies that will transform society. This article introduces some of our initiatives to push deeper in communication science, in areas including speech and audio processing, dialogue processing, human information science, sports brain science, and machine learning and optimization.

• SpeakerBeam: A New Deep Learning Technology for Extracting Speech of a Target Speaker Based on the Speaker’s Voice Characteristics

  

  Abstract
  In a noisy environment such as a cocktail party, humans can focus on listening to a desired speaker, an ability known as selective hearing. Current approaches developed to realize computational selective hearing require knowing the position of the target speaker, which limits their practical usage. This article introduces SpeakerBeam, a deep learning based approach for computational selective hearing based on the characteristics of the target speaker’s voice. SpeakerBeam requires only a small amount of speech data from the target speaker to compute his/her voice characteristics. It can then extract the speech of that speaker regardless of his/her position or the number of speakers talking in the background.

• Network Reliability Optimization by Using Binary Decision Diagrams

  

  Abstract
  We introduce an algorithm that can automatically identify communication network topologies that are robust against failures. Robustness is usually assessed by the metric of network reliability. Since communication networks are a critical infrastructure, designing networks with high network reliability values is essential. However, the problem of finding a network topology that offers the maximum network reliability is a computationally difficult problem, and previous methods therefore restrict their application area to very small networks. Our proposed method exploits the novel data structure called binary decision diagrams, which makes it possible to find the most reliable network topology for communication networks with more than 10 times as many nodes (100) than is possible with previous methods.

• Ukuzo―A Projection Mapping Technique to Give Illusory Depth Impressions to Two-dimensional Real Objects

  

  Abstract
  Ukuzo is a light projection technique that gives illusory depth impressions to two-dimensional real objects by projecting cast shadow images onto them. The technique can not only give depth impressions but also manipulate material impressions of the object with the projected shadow patterns. The technique is promising for enhancing the expression in paper-based advertisements and visual arts.

• Cross-media Scene Analysis: Estimating Objects’ Visuals Only from Audio

  

  Abstract
  Human beings can get a visual image of the surrounding environment from sounds they hear. Can we give similar capabilities to computers? In this article, we introduce our recent efforts in cross-media scene analysis applied to estimate the type, location, and visual shape of objects in a scene based only on sound sources recorded with multiple microphones.

• Measuring, Understanding, and Cultivating Wellbeing in the Age of Technology

  

  Abstract
  The development of information and communication technology (ICT) has brought efficiency and convenience to our daily lives. However, it has also been observed to have a negative impact on the emotional state of users. The concept of design for wellbeing is currently drawing a lot of attention. This approach seeks to explore how ICT can support human psychological wellbeing. In this article, we address several questions related to the issue of how to exploit technology and design to improve the wellbeing of humans.

Regular Articles

• High-speed Avalanche Photodiodes toward 100-Gbit/s per Lambda Era

  

  Abstract
  We have developed an avalanche photodiode (APD) in an effort to achieve 100-Gbit/s operation with a single wavelength (100-Gbit/s/lambda). The APD features a gap-grading layer and was set between the absorber and avalanche layers. It achieved an operating bandwidth of 42 GHz in low-gain conditions. An optical receiver made with the APD demonstrated 106-Gbit/s 4-level pulse amplitude modulation operation and 40-km optical amplifier-free transmission over a single-mode fiber under assumption of a KP4 forward error correction threshold.

Global Standardization Activities

• Telecom Infra Project―Its Structure and Activities

  

  Abstract
  Telecom Infra Project (TIP) was launched in February 2016 through the initiative of Facebook. It is aimed at bringing about innovations in telecom hardware, software, and operations by introducing such technical trends as openness and disaggregation, which have become increasingly important in recent years. TIP is similar to open source projects in that communities play an essential role.

Information

• Event Report: NTT Communication Science Laboratories Open House 2018

  

  Abstract
  NTT Communication Science Laboratories Open House 2018 was held in Keihanna Science City, Kyoto, on May 31 and June 1, 2018. Over 1600 visitors enjoyed 6 talks and 29 exhibits, which included our latest research efforts in the fields of information and human sciences.

Short Reports

• Ultrahigh-speed Integrated Circuit Capable of Wireless Transmission of 100 Gbit/s in the 300-GHz Band

External Awards/Papers Published in Technical Journals and Conference Proceedings

• External Awards/Papers Published in Technical Journals and Conference Proceedings