November 2017 Vol. 15 No. 11

Front-line Researchers

• Naonori Ueda, NTT Fellow, Head of Ueda Research Laboratory and Director of Machine Learning and Data Science Center, NTT Communication Science Laboratories

  

  Overview
  The 2016 White Paper on Information and Communications in Japan issued by the Ministry of Internal Affairs and Communications states that the proactive use of information and communication technology such as the Internet of Things and artificial intelligence has the potential to accelerate economic growth in Japan, and the key to this growth will be the collection and use of big data. Against this background, there are high expectations for research achievements in machine learning. NTT Fellow Naonori Ueda of NTT Communication Science Laboratories has announced a string of Japan-first and world-first achievements in machine learning analysis and technology. We asked him about important aspects of research and his frame of mind as a researcher.

Feature Articles: Communication Science that Enables corevo^®―Artificial Intelligence that Gets Closer to People

• Basic Research in the Era of Artificial Intelligence, Internet of Things, and Big Data―New Research Design through the Convergence of Science and Engineering

  

  Abstract
  To build a productive relationship between humans and artificial intelligence (AI), we must grasp the current situation as accurately as possible and make investments in the future toward developing such a relationship. This article introduces how we see and interpret the AI, Internet of Things, and big data era from the standpoint of promoting research and development of basic technologies.

• Generative Personal Assistance with Audio and Visual Examples

  

  Abstract
  The rapid progress of deep learning is affecting the world we live in. Media generation (i.e., image and audio generation) is a typical example of this progress, and impressive research results are being reported around the world. In this article, we first overview this very active research field. Then we introduce our efforts in developing a generative personal assistance system with audio and visual examples. Specifically, we explain how our new deep-learning approach will overcome the limitations encountered in existing studies on personal assistance systems. Finally, we discuss future directions in the media generation field.

• Efficient Algorithm for Enumerating All Solutions to an Exact Cover Problem

  

  Abstract
  We introduce an algorithm that finds all solutions to an exact cover problem. Many real-world tasks including designing apartment layouts and electric circuits can be formulated and solved as exact cover problems. Our algorithm can solve exact cover problems up to 10,000 times faster than the previous method. Moreover, our method compresses and stores all solutions and so can efficiently find the solutions that satisfy several constraints. Therefore, our algorithm can efficiently find good solutions to exact cover problems found in the real world.

• Memory-efficient Word Embedding Vectors

  

  Abstract
  Word embedding is a technique for identifying the semantic relationships between words by computer. Word embedding vectors enable computers to provide a guess similar to the intuition or common sense of human beings. This article introduces a method for reducing the required memory consumption of this important fundamental operation of word embedding vectors while maintaining the ability to calculate semantic relationships, which is an important property when this technique is applied to real world systems.

• Synthesizing Ghost-free Stereoscopic Images for Viewers without 3D Glasses

  

  Abstract
  When a conventional stereoscopic display is viewed without three-dimensional (3D) glasses, image blurs, or ghosts, are visible due to the fusion of stereo image pairs. This artifact severely degrades 2D image quality, making it difficult to simultaneously present clear 2D and 3D content. To overcome this limitation, we recently proposed a method to synthesize ghost-free stereoscopic images. Our method gives binocular disparity to a 2D image and drives human binocular disparity detectors by the addition of a quadrature-phase pattern that induces spatial subband phase shifts. The disparity-inducer patterns added to the left and right images are identical except for the contrast polarity. Physical fusion of the two images cancels out the disparity-inducer components and makes only the original 2D pattern visible to viewers without glasses.

• Personalizing Your Speech Interface with Context Adaptive Deep Neural Networks

  

  Abstract
  This article introduces our recent progress in speaker adaptation of neural network based acoustic models for automatic speech recognition. Deep neural networks have greatly improved the performance of speech recognition systems, enabling the recent widespread use of speech interfaces. However, recognition performance still greatly varies from one speaker to another. To address this issue, we are pursuing research on novel deep neural network architectures that enable rapid adaptation of network parameters to the acoustic context, for example, the speaker voice characteristics. The proposed network architecture is general and can potentially be used to solve other problems requiring adaptation of neural network parameters to some context or domain.

Regular Articles

• Optimization of Harvest Time in Microalgae Cultivation Using an Image Processing Algorithm for Color Restoration

  

  Abstract
  Algae have been attracting attention as a next-generation alternative energy source to fossil fuels, but the high cost of cultivation remains an issue. This article introduces technology that reduces the cultivation cost through the use of an image-processing algorithm for color restoration to quickly determine the optimal time to harvest algae.

Global Standardization Activities

• Report on First Meeting of ITU-T TSAG (Telecommunication Standardization Advisory Group) for the Study Period 2017 to 2020

  

  Abstract
  The first meeting of the Telecommunication Standardization Advisory Group (TSAG) of the International Telecommunication Union - Telecommunication Standardization Sector (ITU-T) for the study period 2017 to 2020 was held at the ITU headquarters in Geneva, Switzerland, May 1−4, 2015, with some 120 delegates from 40 countries attending. The 11-person Japanese contingent comprised representatives of the Ministry of Internal Affairs and Communications, NTT, KDDI, Hitachi, Fujitsu, Mitsubishi Electric, NEC, OKI, and TTC (The Telecommunication Technology Committee). The new organization of TSAG is overviewed, and topics discussed during the meeting are described in this article.

Information

• Event Report: NTT Communication Science Laboratories Open House 2017

  

  Abstract
  NTT Communication Science Laboratories Open House 2017 was held in Keihanna Science City, Kyoto, on June 1 and 2, 2017. Nearly 1800 visitors enjoyed 6 talks and 29 exhibits, which focused on our latest research activities and efforts in the fields of information and human sciences.

Short Reports

• Arkadin Brings Businesses into the Future with New Cloud Unified Communications Services and Digital Operations Strategies

  

  Abstract
  As an NTT Communications company, Arkadin is at the forefront of digital cloud communications technologies. We are transforming and adapting rapidly to take advantage of new business opportunities in this fast growing and dynamic market sector. Several services were introduced and a new partnership struck in the past few months, which we believe will help enhance our market leadership and drive greater success for our customers.

• World’s Largest Transmission Capacity with Standard Diameter Multi-core Optical Fiber―Accelerated Multi-core Fiber Application Using Current Standard Technology

External Awards/Papers Published in Technical Journals and Conference Proceedings

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