The Advanced Telecommunication Research Institute

 Path: rwsys!trsvax!utacfd.uta.edu!utagraph.uta.edu!hermes.chpc.utexas.edu!news.utdallas.edu!wupost!zaphod.mps.ohio-state.edu!uakari.primate.wisc.edu!usenet.coe.montana.edu!nntp.uoregon.edu!milton!hlab

From: rick@cs.arizona.edu (Rick Schlichting)

Newsgroups: sci.virtual-worlds

Subject: %-) INDUSTRY: Kahaner Report: Advanced Telecommunication Research

Message-ID: <1992Feb9.014604.6173@milton.u.washington.edu>

Date: 8 Feb 92 03:08:49 GMT

Sender: hlab@milton.u.washington.edu (Human Int. Technology Lab)

Organization: HIT Lab, Seattle WA.

Lines: 724

Approved: cyberoid@milton.u.washington.edu

Crossposted from comp.research.japan.  ATR is the leading virtual-worlds

research site in Japan, according to many sources.  This report may be

useful in understanding why.

  [Dr. David Kahaner is a numerical analyst on sabbatical to the 

   Office of Naval Research-Asia (ONR Asia) in Tokyo from NIST.  The 

   following is the professional opinion of David Kahaner and in no 

   way has the blessing of the US Government or any agency of it.  All 

   information is dated and of limited life time.  This disclaimer should 

   be noted on ANY attribution.]

  [Copies of previous reports written by Kahaner can be obtained from

   host cs.arizona.edu using anonymous FTP.]

To: Distribution

From: David Kahaner ONR Asia [kahaner@cs.titech.ac.jp]

Re: The Advanced Telecommunication Research Institute (ATR)

7 Feb 1992

This file is named "atr.92"

ABSTRACT: A visit to The Advanced Telecommunication Research Institute, 

(ATR) on 21 Jan 1992 is described.  

I first visited ATR more than 18 months ago, shortly after its 16K node 

Connection Machine CM-2 had been installed. See my report "atr", 14 June 

1990 for details. 

My current visit was hosted by 

        Dr. Kazunari Nakane

        Head: Cognitive Processes Department

        ATR

        Auditory and Visual Perception Lab

        Seika-cho, Kyoto 619-02, Japan

         Phone: +81-7749-5-1411, Fax: +81-7749-5-1408

         Email: NAKANE@ATR-HR.ATR.CO.JP

who also provided a great deal of assistance in the editing of this

report, and whose patience is sincerely appreciated.

The ATR organization is complicated.  It was established in October 1985

as a result of the privatization of the Japanese telephone company. At

that time the Japan Key Technology Center (JKTC) was set up, funded mostly by

the dividends from NTT stock that is owned by the Japanese government.

The government now owns 60% of the stock of NTT.  JKTC funds about 40

projects, of which by far the largest is ATR. JKTC top management

includes representatives of various Ministries, MITI, MOF (Ministry of

Finance) and MPT (Ministry of Posts and Telecommunications), although

ATR is viewed as the major laboratory associated with MPT (the ETL lab

is associated with MITI even more directly).  Actually JKTC provides

only 70% of the funding for the four laboratories (or projects) that

represent the science component of ATR.  The labs were established as

7-10 year projects; two will expire in 1993 and two in 1996 although

there is every reason to think that they will be renewed, perhaps with

slightly different emphasis. NTT and 140 other Japanese companies have

also provided private investment and the four labs get the remaining 30%

of their funding from this. The same companies also fund ATR

International, which supports the labs, provides research facilities and

support structure.  ATR International has a president and board of

directors.  It is jointly owned by NTT and the 140 other Japanese

companies with NTT holding about 60% of its stock. To make it more

complicated, the three key people at ATR International are retired from

MPT, the regional Kansai government in which ATR is located, and NTT.

Each of the individual labs has a president, two are from NTT, one from

KDD and one from MPT. In addition, many of the staff within each lab are

part of a particular NTT organization.  Thus there is a great deal of

vertical information flow, and for similar reasons the labs are rather

independent.  Practically though, ATR can be considered as part of the

NTT family.

The four ATR laboratories are co-located in a building in Kansai Science

City, a new area about halfway between Kyoto and Osaka.  I was told that

building costs were about $70M US.  Inside the ATR building there is a

fifth laboratory (called NTT Communication Science Laboratories--CSL)

which started operation in July 1991.  This operation is not part of ATR

but is just leasing the building space.  Right behind ATR, NTT owns

land, where they may establish a new building for this lab.  If this

happens, it will be the only official NTT lab outside of the Kanto area.

Almost all the researchers now at CSL were transferred from NTT Labs in

Mushashino or Yokosuka in Tokyo area.

When I first visited in 1990, ATR was the only building in sight. Now it

is surrounded by others in various stages of construction, including

several hundred residential town-houses. Across the street a large

Sumitomo is going up.  Across from that, a big project called the

Keihanna Convention Center is under construction.  (KEI-HAN-NA is a

combination of the abbreviation of three words, Kyoto, Osaka, and Nara.)

It will be a convention center with hotel and so on.  Those readers who

have been to Japan should imagine what Tsukuba Science City (near Tokyo)

must have looked like during its earliest days to get a sense of the

area around ATR.  Like Tsukuba in those days, many ATR staff feel

isolated from more established parts of Japan, but this will change as

the community develops roots.

Almost half the ATR management are on rotational assignment from a

couple of NTT's labs or divisions. And the other half are mostly from

laboratories of KDD (international telephone company), NHK

(semi-national radio and TV corporation), and CRL (Communication

Research Laboratory of Ministry of Post and Telecommunication).  Most of

the staff are also rotational, coming for a few years.  About one fourth

are from NTT but other Japanese companies are also represented.  There

are also visiting Western researchers. I met four, three Americans and a

Frenchman. Their appointments were temporary, one year at a time, but

apparently could be extended several times by mutual agreement. Each of

the four labs have a few senior scientists who are either permanent or

at least on very long term assignment. These people provide the

technical leadership and continuity.

Rotation of staff in and out of the labs provides a collection of new

ideas, but also leads to some underutilization of equipment that is

ordered for use in one near term research project but is of less

interest to the next.  However, even in such cases there is residual

benefit, as the overall research projects are generally well planned and

are all heading toward some very long term goals, with natural near term

revisions and adjustments. I was told that most equipment will be used

in succeeding projects.  However, staff turnover is definitely ATR's

major organizational concern, and I was told that there are new plans to

develop more permanent positions. This is a difficult issue, because the

ATR labs are not permanent.  If ATR hires any permanent employee they

have to provide for their long-term career plan.  I was told that one

approach is to hire only those that can easily be absorbed back into a

sponsoring company, a university, or a research institute when the ATR

projects end.

Staff experience is also an issue.  As is typical of Japanese companies,

there are relatively few PhDs.  Companies that support ATR clearly use

the labs as a place to put younger employees for training. There was

some difference of opinion among the Japanese I spoke to as to whether

an assignment at ATR was in anticipation of good work to follow (back at

the home company), a reward, or a banishment.  This is very company

dependent. However, most companies send their good engineers because

they are representatives of the company and they want them to look good.

Researchers who come with definite ideas about what to work on seem to

fare quite well.  I was told that each and every researcher knows what

he will do at ATR because there are negotiations on this subject before

he joins.  However, there is the usual trial and error associated with

research topics. Perhaps Western researchers have even more flexibility

in this matter.  I believe though, that many of the younger Japanese are

assigned to projects.  Most new staff will work on a project suggested

by an advisor or department head, by joining an existing project, or by

doing research alone.  There is also the problem of what some of these

Japanese will do when their ATR assignment is complete. Of course, most

will return to their companies, but several have gone on to

universities. One expressed concern to me that the basic research he was

doing at ATR might not be supported in his home lab and hoped that he

would be able to find a good spot when he returned. On the other hand,

some returning employees from Sony, Matsushita and others have been

known to get overseas assignments (e.g.  to US, Germany, UK, etc) after

their ATR assignments. So coming to ATR is a kind of stepping stone for

them. And in general, many researchers have been able to make excellent

use of the experience, results and ideas, etc., from ATR and, in some

cases, were placed in very good positions.  This is especially true for

NTT. On balance, a few years at ATR seems to be a good thing for the

scientists involved.

The four labs are mostly independent although they are housed in the

same building. The independence is enhanced by giving each lab its own

President as mentioned earlier. This was a bit strange to me, as the

total staffing of 260 is low enough that there could be a great deal of

interaction, although each of the labs is aiming at its own specific

research goals. However, given the complicated funding arrangements it

is not too surprising.

A good example of the independence is the use of the Connection Machine.

This was purchased primarily for the computational studies of cognitive

processes carried out in the Auditory and Visual Perception Research

Lab.  While it is extensively used, I was told that some time is

available (there might be some differences of opinion about this), but

researchers from other labs and even other groups within this lab are

occasionally discouraged from using it.  Dr. Nakane explained to me that

funding for the machine is for specific research jobs (rather than for

other purposes or mere interest). This requires careful accounting and

designation of users. Still, such a important resource ought to be of

great help to scientists throughout ATR.

Lab staffing is as follows. In 1990, total staff was 262, and total

budget was 7.5B Yen, about $50M US. This does not include the space

occupancy fee that ATR International charges to each of the R&D

organizations. I mentioned this figure to a colleague who had spent

three years at ATR, and he felt that the budget figure was severely

underestimated, perhaps because of the complicated funding situation. He

thought that a more realistic figure was twice that.

                                                              

      NAME OF LAB                       

                                           

        Communications Systems                  37

        Interpreting Telephony                  52

        Auditory and Visual Perception          57

        Optical and Radio Communications        40

        ATR International                        6

     RESEARCH STAFF BREAKDOWN

        Invited international researchers       24

        Invited domestic researchers            15

        Other researchers                      144

        Permanent staff researchers              9

     OTHER STAFF

        Assistants                              42

        Clerks                                  28

Publications and patents have been growing since the ATR's founding.

                                         PAPERS          PATENTS

                              International  Domestic

        1986                            5       45          15

        1987                           21      282          70

        1988                           80      414         120

        1989                          140      422         160

        1990                          169      448         223

The figures above were given to me by Dr. Eiji Yodogawa, President of

the Auditory and Visual Perception Lab (AVPL) (Email:

yodogawa@atr-hr.atr.co.jp).  

ATR publishes the ATR Journal, which is mostly in Japanese. However,

they also publish annually, a bound collection of the staff's technical

reports. Many of these are in English, and all have English titles and

abstracts. A list of these reports written in 1991 from AVPL is appended

to this report. English reports are so noted.

The four labs and their primary activities are as follows.

Communications Systems:

  Main goal: Human-oriented intelligent communication system

        Communications with realistic sensations, automatic 3-D shape 

           acquisition, recognition, comprehension, modeling, 

           manipulation and display, 

        Nonverbal interfaces, recognition of facial and eye-gaze 

           directions, understanding gestures and hand movements, 

           integration of visual and speech information 

        3-D image databases

        Cooperative work environment for design of solid objects

        Automatic generation of communication software, extraction of 

           real intentions, use of visual language to give specifications 

           accurately, human deliberative mechanisms in software design, 

           knowledge base, easy to use specification description language 

        Security, cryptographic techniques for large capacity (image) 

           communications, secure telecommunication networks 

Optical and Radio Communications:

        Optical intersatellite communications, optical beam control, 

           optical modulation/demodulation 

        Advanced antennas, active array technology for mobile antennas, 

           methods of mitigating multipath propagation problems, 

           microwave circuit integration, signal processing 

        Optical and electronic devices, growth and characterization of 

           semiconductors with precisely controlled atomic 

           configurations, nonlinear optical devices 

Interpreting Telephony:

  Main goal: Automatic intepreting telephone

        Speech recognition and synthesis, speech database

        Interface between speech and language, spoken language 

           processing, knowledge base, speech and language integration 

        Machine translation, grammar for analysis of Japanese, dialog 

           interpretation, contextual processing 

        Advanced dialogue processing, contextual processing 

My visit was confined to the Auditory and Visual Perception Lab, which 

has three departments. Its main goal is an improved human-machine 

interface.  

Auditory and Visual Perception:

   Visual Perception Department

            Visual perception mechanisms:

              Motion perception, binocular stereopsis, 

              Character and word perception

              Image concept formation

            Brain activity measurement (eye movement, etc)

            Pattern recognition:

              Neural network models for handwritten characters,

                spatiotemporal patterns

              3D object recognition

   Cognitive Processes Department

            Cognitive processes for visual information:

              Neural computing models of pattern vision and spatial vision

            Parallel processing mechanism:

              Mathematical analysis and synthesis of neural net models

              Information processing using neural net models

            Learning and motor control mechanisms

              Neural net models of human motor control

              Integrated learning of somato sensory and visual information

 

   Hearing and Speech Perception Department

            Hearing mechanism:

              Auditory peripheral models

              Auditory preprocessor for speech recognition

            Speech Perception mechanism

              Co-articulation model and application to speech recognition

              Speech recognition using neural nets

              Speech prosody

In this lab one of the most active researchers was

        Dr. Mitsuo Kawato

        ATR (address as above)

         Tel: +81-7749-51452, Fax: +81-7749-51408

         Email: KAWATO@ATR-HR.ATR.CO.JP

Kawato's background is in biophysical engineering and was previously a

faculty member at Osaka University.  Kawato is one of the few

semi-permanent members of the research staff and he is collaborating

with almost a dozen younger researchers on an astonishing variety of

projects.  This research has generated nearly 50 papers since 1987, most

in Western journals or in Proceedings of international conferences.

Kawato gave me a long list of the projects that he is currently working

on (outlined below), but I had an opportunity to discuss only a few of

these with his co-workers.

Learning trajectory control.

   * Learning control experiment of rubbertuator Softarm by feedback-error 

       learning neural net. Inverse statics and inverse dynamics. Six-

       muscle arm model. Sensor fusion for object manipulation 

   * Feedback error learning for closed loop system. Models of regions of 

       cerebellum. Recognition of manipulated objects by motor learning.  

Trajectory formation for arm and articulator.

   * Explanation of Fitts law by neural net. Learning acquisition of 

       forward dynamics of speech articulator on CM-2 using EMG as 

       control input.  

   * Combination of cross validation method and new information 

       criterion to estimate generalization capability of neural net 

       models.  Pattern generation and recognition using neural net.  

   * Minimum-muscle-tension-change and minimum-motor-command models based 

       on biomechanical data of monkey musculoskeletal system with 17 

       muscles.  

   * Human arm movement using Optotrack and recording EMG. 

   * Measurement of EMG, articulatory movement and acoustics for speech 

       articulator and neural net model.  

Computational model of visual cortices and sensor fusion.

   * Use of CM-2 to implement Markov random field (MRF) model of images, 

       and learning of potential energies for hidden variables without 

       teachers.  

   * Hierarchical MRF based on conditional probability and application to 

       pattern recognition 

   * Shape from shading based on forward and inverse models of optics

   * Integration of visual, somatosensory, and motor control information

       for object manipulation.  

There is a great deal of neural net research in progress, including new

algorithms, and applications to physical situations of speech and

vision.  The traveling salesman problem is another application being

studied.  The CM-2 is used heavily now for simulations. (M. Hirayama

showed me several very interesting demonstrations.) During my first visit

the ATR staff was still learning about this machine and usage was low. I

was told, however, that the machine now much busier.  It is extensively

used as a neural network simulator for the previously mentioned studies

of human perception, motor control and some other related topics.  Some

of the computation/simulation on CM-2 require enormous amount of machine

use, say, ten days of continuous time for example.  Presently the CM-2

is working in 'exclusive mode' but not in 'TSS mode'.  This is partly

because of some problems with the TSS Operating System, and partly

because of some decrease of computational performance in TSS mode, which

seem to decrease the efficiency of the CM-2 usage. However, the newest

version of TSS OS seems to have resolved these problems and is to be

installed very soon.  However, it has been used heavily enough so that

request are being heard that it be replaced with more powerful one. The

current machine was purchased on a 3-year installment plan, and included

32-bit floating point units. ATR is hoping to get 64-bit floating point

in future, but at the current time this has not happened.  (I believe

that there are three Connection Machines in Japan--one at ATR, a second

at the Institute for Computational Fluid Dynamics in Tokyo, and a third

at an undisclosed location, probably a private company.)

One of the most interesting aspects of ATR's research is how tightly 

coupled it is to biological models of real sophistication.  A significant 

aspect of the work seems to be associated with modeling and then 

verifying models with complicated physical measurements. One example of 

this is related to understanding the mechanism behind the articulatory to 

acoustic transform, i.e., to understand the process by which linguistic 

intentions become speech. Speech production entails extraordinary 

coordination among diverse neurophysiological and anatomical structures. 

These unfold through time to produce a complex acoustic signal that 

conveys to listeners something of the speakers intentions. ATR's research 

has used measured movements of the articulator (upper and lower lips and 

horizontal and vertical jaws) and EMG data from four separate muscle 

groups and used these to train a neural net to generate motor commands to 

the articulators. 

Another especially elegant example is work by Kawato on studying the

trajectory followed by human arms. Kawato proposed that this minimizes

the time integral of the square rate of change of the torque. Based on

this model he has developed a neural network that reproduced Fitts law

(sometimes called speed-accuracy trade-off of arm movement) and also

used it to study feedback-error-learning for ill-posed problems. Human

sensory-motor control, such as arm movement involves a number of

ill-posed problems.

I think that it is unusual to find so much biophysical expertise being

put to use in modeling, especially coupled with advanced computer

hardware and software.  In the area of vision, there are projects to

automatically generate 3D images from a series of stereo- pairs, and

another fascinating project to analyse, quantify, and predict our

response to optical illusions.  Projects like these require careful

eye-movement measurements and distance monitoring, as well as deep

understanding of current ideas in cognition. These are difficult

mathematical problems too. Our brain essentially solves the inverse

problem of reconstructing a 3D visible surface from the data of a 2D

image projected onto the retina. (Solutions of this problem are related

to extensions of regularization techniques, some of which are very well

known in computational mathematics.) There certainly is plenty of

hardware available for experimenting. While I was there two technicians

from Canada were visiting ATR to help set-up a pair of Optotracks,

sophisticated distance measuring equipment costing in the neighborhood

of $60K US each. Workstations are everywhere, along with associated

peripherals. The ATR building and labs are large,  spacious, and very

well equipped.  One of the only places in the US that I know of, where

comparably subtle work is going on is at MIT's brain research

department, and there is a substantial flow of visitors between ATR and

MIT.  (Immediately after my visit, T. Poggio from MIT arrived to give

several days of lectures.)

My visits to ATR have been cordial and revealing. About ten scientists I

met were enthusiastically willing to describe their research, although

the independent laboratory structure made it difficult for me to get a

full picture. My visits have been highly organized and tightly time-

constrained.  The lab I visited is doing very excellent basic research,

but I still haven't seen the work on interpreting telephony or

communications systems in other laboratories, and I would like to go

back and learn about these.  What I have seen is a group of capable

Japanese, supplemented by a substantial number of long term Western

visitors.  There are still some birth pains both at ATR and in the

surrounding area; ATR is barely five years old.  Finally, procedures

need to be established to insure that ATR is not only a training ground

for young Japanese scientists and a temporary or sabbatical station for

Western ones, but also to insure that ATR will develop the continuity

and world class stature that will enable it to accomplish the major

long-term goals it has set for itself.

-------------------------------------------------------------------------

A Collection of Technical Publications

1991 Vol. 1 of 2 (Jan--June 1991)

ATR Auditory and Visual Perception Research Laboratories

Note: Authors listed simply as ATR refer to 

        ATR Auditory and Visual Perception Labs

        Seika-cho, Kyoto 619-02, Japan

         Phone: +81-7749-5-1411, Fax: +81-7749-5-1408

For email access, try "lastname"@ATR-HR.ATR.CO.JP

Unless noted all papers are in Japanese with English abstracts.

1.  Visual Perception

Hemispheric difference between Japanese Kanji and Kana in a Picture-word

Matching Task

     Chisato Aoki (ATR)

     Adam Reeves (North-Eastern Univ.)

A Camera Calibration Method without a Special Object to be Measured

     Youji Fukada (ATR)

A Discontinuity Detector Based on the Pervasive Noise in Surface Property Data

     E.B. Gamble, Jr. (ATR) (In English)

A New Gaze Analyzer Which Limits the Viewer's Gaze Time and Its Application to

        the Analysis of a Visual Search

     Hitoshi Hongo, Mitsuho Yamada (ATR)

Eye Movement Analysis of Visual Search Process - How We Count Patterns

        Quickly?

     Hitoshi Hongo, Mitsuho Yamada, Keiichi Ueno (ATR)

Objective Evaluation of Hand-Written Character Quality

     Takahito Kato, Mitsuho Yamada (ATR)

System for Predicting the Quality of Hand-Written Characters

     Takahito Kato (ATR)

Quality Factors of Hand-Written Characters Based on Human Visual Perception

     Takahito Kato, Mitsuho Yamada (ATR) (In English)

Approximation of Nonlinear Dynamics by Recurrent Network and Learning Chaos

     Yoshihiko Murakami, Masa-aki Sato (ATR)

A Recurrent Network which Learns Chaotic Dynamics

     Yoshihiko Murakami, Masa-aki Sato (In English)

Positive Motion After-Effect Induced by Bandpass Filtered Random-Dot

        Kinematograms

     Shin'ya Nishida (ATR)

     Takao Sato (NTT Basic Research Laboratories)

Positive Motion Aftereffect Induced by a Bandpass Random Dot Kinematogram

     Shin'ya Nishida (ATR)

     Takao Sato (NTT Basic Research Laboratories)

Visual Motion Perception and Eye Movement: A Two Stage Model for Visual Motion

        Perception Including Applicability to Bionocular Depth Perception

     Masami Ogata, Takao Sato (ATR)

Context Dependent Effect in the Image Concepts Formation and its Application

     Masaomi Oda (ATR)

Mathematical Foundation of Wavelets I: Non-orthonormal Wavelets

     Masa-aki Sato (ATR)

Mathematical Foundation of Wavelets II: Orthonormal Wavelets

     Masa-aki Sato (ATR)

A Recurrent Neural Network which Learns Fluctuations in Voice Waveforms

     Masa-aki Sato, Kazuki Joe, Tatsuya Hirahara (ATR)

Learning Nonlinear Dynamics by Recurrent Neural Networks

     Masa-aki Sato, Yoshihiko Murakami (ATR)

Monocular Stereopsis After Motion Adaptation

     Hiroshi Ono (Department of Psychology, York University)

     Satoshi Shioiri, Takao Sato (ATR) (In English)

Third International Conference on Computer Vision

     Keiichi Ueno

Human Visual Perception and Cognitive Processes, Modeling and its Application

     Keiichi Ueno, Kazunari Nakane (ATR)

Binocular Eye Movement Analysis of Stereoscopic Images - In the Case of

        Showing a Moving Stimulus only to One Eye

     Kenya Uomori, Mitsuho Yamada (ATR)

Binocular Eye Movement in the Case of Showing a Moving Stimulus Only to One

        Eye

     Lenya Uomori, Mitsuho Yamada (ATR)

Objective Evaluation of the Feeling of Depth in 2D or 3D Images Using the

        Convergence Angle of the Eyes

     Mitsuho Yamada (ATR)

     Nobuyuki Hiruma, Haruo Hoshino (NHK Science and Technical Research

             Laboratories) (In English)

Accuracy of Binocular Disparity Analysis

     Mitsuho Yamada, Kenya Uomori (ATR)

Analysis of Line of Sight During Athletic Sports

     Mitsuho Yamada (ATR)

Fractal Dimension Analysis on Binocular Small Involuntary Eye Movements

     Hiroshi Yoshimatsu, Mitsuho Yamada (ATR) (In English)

Fractal Dimension of Drift Components of Small Involuntary Eye Movement

     Hiroshi Yoshimatsu, Mitsuho Yamada (ATR)

Fractal Dimension Analysis of Small Involuntary Movement

     Hiroshi Yoshimatsu, Mitsuho Yamada (ATR)

A Fractal Dimension of Binocular Small Involuntary Movement

     Hiroshi Yoshimatsu, Mitsuho Yamada (ATR)

2.  Cognitive Processes and Behavior

Control of Movement, Postural Stability, and Muscle Angular Stiffness - A 17-

        Muscle Model of the Monkey's Arm

     Menashe Dornay (ATR) (In English)

Characteristics of Moving Distance Estimation in a CG-Space

     Hideo Fujii, Toshio Inui (ATR)

Computational Models for Learning Motor Control in Four Regions of Cerebellum

     Hiroaki Gomi, Mitsuo Kawato (ATR)

Learning Control of a Closed Loop System Using Feedback-Error-Learning

     Hiroaki Gomi, Mitsuo Kawato (ATR)

A Model of Perceiving Shape from Shading in Monocular Depth Perception

     Hideki Hayakawa, Toshio Inui, Mitsuo Kawato (ATR)

A Computational Model of Perceiving Shape from Shading in Monocular Depth

        Perception

     Hideki Hayakawa, Toshio Inui, Mitsuo Kawato (ATR)

Information Processing using Fine-Grained Parallel Computer

     Makoto Hirayama (ATR)

Contour Extraction by Local Parallel and Stochastic Algorithm which has Energy

        Learning Faculty (In English)

     Sadayuki Hongo, Mitsuo Kawato, Toshio Inui (ATR)

     Sei Miyake (NHK Science and Technical Research Laboratories)

Contour Extraction of Natural Images Based on a Multi-Layered MRF Model -

        A Two-Resolution Model

     Sadayuki Hongo, Mitsuo Kawato, Toshio Inui, D. Litt (ATR)

Computational Theory of Vision

     Toshio Inui (ATR)

Virtual Trajectory and Stiffness Ellipse During Force-Trajectory Control Using

        a Parallel-Hierarchical Neural Network Model (In English)

     Masazumi Katayama, Mitsuo Kawato (ATR)

Learning Trajectory and Force Control of an Artificial Muscle Arm -

        Neural Network Control with Hierarchical Objective Functions

     Masazumi Katayama, Mitsuo Kawato (ATR) (In English)

A Neural Network Model which Recognizes Shape of a Grasped Object and Decides

        Hand Configuration

     Naohiro Fukumura, Yoji Uno Ryoji Suzuki (Faculty of Engineering,

             University of Tokyo)

     Mitsuo Kawato (ATR)

Motor Control by Neural Network Models

     Mitsuo Kawato (ATR)

Model of Four Regions of Cerebellum (In English)

     Mitsuo Kawato, Hiroaki Gomi (ATR)

Optimization and Learning in Neural Networks for Formation and Control of

        Coordinated Movement

     Mitsuo Kawato (ATR)

Chronobiology

     Mitsuo Kawato (ATR)

Computational Theory and Visual Cortices

     Mitsuo Kawato, Toshio Inui (ATR)

Self-Similarity Analysis of Images using Schauder Expansion

     Tsuyoshi Ogura, Noboru Sonehara (ATR)

Neural Network Applications to Image Processing

     Noboru Sonehara (ATR)

Neural Network Model Processing on Massively Parallel Computers

     Noboru Sonehara, Makoto Hirayama (ATR)

Relaxation Neural Network Model for Optimal Binary Representation of Images

        and Its Implementation on a Parallel Computer

     Noboru Sonehara (ATR)

3.  Hearing and Speech Perception

Auditory/Speech-Perception Modeling and Its Applications to Speech Processing

     Shigeru Katagiri, Yoh-ichi Tohkura (ATR)

Internal Speech Spectrum Representation by Spatio-Temporal Masking Pattern

     Tatsuya Hirahara (ATR) (In English)

A Nonlinear Cochlear Filter with Adaptive Q Circuits

     Tatsuya Hirahara (ATR)

LVQ2 Phoneme Recognition using Adaptive Q Cochlear Filter-Banks

     Tatsuya Hirahara, Hitoshi Iwamida (ATR)

Auditory Peripheral Models in Speech Recognition

     Tatsuya Hirahara (ATR)

Automatic Speech Recognition and Auditory Peripheral Models

     Tatsuya Hirahara (ATR)

Current Research Status of Auditory Peripheral Models

     Tatsuya Hirahara (ATR)

Acoustic Characteristics of the ATR Variable Reverberation Room

     Tatsuya Hirahara, Carl Muller, Yoh'ichi Tohkura (ATR)

Frequency Response of Headphones Measured in Free Field and Diffuse Field by

        Loudness Comparison (In English)

     Kazuo Ueda, Tatsuya Hirahara (ATR)

LVQ Phoneme Recognition Experiments with Spectrum Target Prediction Model

     Toshiyuki Aritsuka (ATR)

     Masato Akagi (NTT Basic Research Laboratories)

Temporal Characteristics of Effects of "Prototypes" on Within-Category

        Discrimination for Speech Sounds

     Minoru Tsuzaki (ATR)

     Jorge A. Gurlekian (Univ. Buenos Aires)

Effects of Typicality and Interstimulus Interval on the Discrimination of

        Speech Stimuli (In English)

     Minoru Tsuzaki (ATR)

     Jorge A. Gurlekian (Univ. Buenos Aires)

On the Validity of the Models for Speech Perception

     Minoru Tsuzaki (ATR)

Reviews of the 121st Meeting of Acoustical Society of America

     Minoru Tsuzaki, Yoh'ichi Tohkura, Yoshinori Sagisaka (ATR)

LVQ-Based Shift-Tolerant Phoneme Recognition  (In English)

     Erik McDermott, Shigeru Katagiri (ATR)

Large Vocabulary Word Recognition Using an LVQ-HMM Hybrid

     Hitoshi Iwamida, Shigeru Katagiri, Erik McDermott (ATR)

Speaker-Independent Large Vocabulary Word Recognition Using an LVQ/HMM Hybrid

        Algorithm (In English)

     Hitoshi Iwamida, Shigeru Katagiri, Erik McDermott (ATR)

A New Connected Word Recognition Algorithm Based on HMM/LVQ Segmentation and

        LVQ Classification (In English)

     Padma Ramesh, Shigeru Katagiri (ATR)

     Chin-Hui Lee (Speech Research Department, AT&T Bell Laboratories)

Review on Recent Discriminative Training Algorithms - Generalized

        Probabilistic Descent Method

     Shigeru Katagiri (ATR)

A New Descriminative Training Algorithm for Dynamic Time Warping-Based Speech

        Recognition (In English)

     Takashi Komori, Shigeru Katagiri (ATR)

Speech Recognition Using a New Minimum-Distortion Segmentation/Discriminative

        Classification Hybrid Algorithm (In English)

     Andrew Duchon, Shigeru Katagiri (ATR)

Discriminative Training for Various Speech Units (In English)

     Erik McDermott, Shigeru Katagiri (ATR)

Neural Networks as a Model of Parallel Computation

     Kaoru Nakazono (ATR)

-----------------------------END OF REPORT---------------------------------