110412 spoken language corpora week 1

Document Sample
110412 spoken language corpora week 1 Powered By Docstoc
					Spoken Language Corpora
2011-04-12 course overview



2011 spring semester,
Tue1
elective course for IMCTS graduate
students
ITE-SE
           make roster
write
 full name
 furigana
 email address

                                         pass sheet


updated 2011-04-11 11:30 utc goh kawai
           informed consent
your speech and actions may be recorded,
 archived and, without revealing your identity,
 used and made public for research and
 education purposes
if you disagree, I will neither record nor
 retaliate
学生の言動を録音し、保存し、匿名としたうえで
 研究と教育のために利用したり公開する可能性
 がある


updated 2011-04-11 11:30 utc goh kawai
           welcome to boot camp


          師師                               語
          範範                               道
          代                              剛 師
         白紀                              道 範
         黒子剛                             場
updated 2011-04-11 11:30 utc goh kawai
           instructor
Goh Kawai (河合 剛 かわい ごう)
 born in Tokyo, raised in Toronto
 came to Sapporo in 2003-04




updated 2011-04-11 11:30 utc goh kawai
           consider showing website
http://goh.kawai.com/




updated 2011-04-11 11:30 utc goh kawai
           goh’s academic background
Univ of Tokyo
  BA linguistics, 1984
ICU
  MA educational technology, 1986
Stanford Univ
  linguistics (dropout)
Univ of Tokyo
  PhD information and
   communication engineering, 1999
updated 2011-04-11 11:30 utc goh kawai
           goh’s vocational background
Xerox Palo Alto Research Center
 Palo Alto, CA
SRI International
 Menlo Park, CA
University of Tokyo
 Tokyo, Japan
University of California Santa Cruz
 Santa Cruz, CA
Oregon Graduate Institute
 Beaverton, OR

updated 2011-04-11 11:30 utc goh kawai
           goh’s interests
research
 spoken and written language
   processing technology applied
   to language learning
personal interests
 flying, kayaking, cycling,
   snowshoeing, amateur radio,
   sado (way of tea)
updated 2011-04-11 11:30 utc goh kawai
           contact info
office: jouhou-kyouiku-kan 3rd
 floor server room
email: grad@kawai.com
web: goh.kawai.com




updated 2011-04-11 11:30 utc goh kawai
           office hours
drop-in or email for appointment
 no phone calls
off campus
       from 2011-04-29 to 2011-05-05
       from 2011-05-15 to 2011-05-24
       from 2011-08-02 to 2011-09-21

updated 2011-04-11 11:30 utc goh kawai
           class periods




updated 2011-04-11 11:30 utc goh kawai
           grad school catalog blurb
担当分野/マルチメディア言語情報処理論
   研究領域、学歴(言語学学士、教育学修士、
 電子情報工学博士)、職歴(研究所2社、大学4校)
 、業績一覧、所属学会、授業資料、教え子の匿名
 コメント(全ての学部授業)などをwebに掲載。メー
 ルで面会予約。電話不可。私の評価を元指導生
 に直接たずねるとよい。
言語情報処理、教育工学☆領域 言語学と情報処
 理技術を利用した非母語学習。☆手法 学習シス
 テムや教材を制作し、学習効果を定量的に評価す
 る。☆指導方法 協同プロジェクトを共著論文にま
updated 2011-04-11 11:30 utc goh kawai
           alumni
平野宏子                                    東京大学 博士(科学)
                                         吉林華僑外国語学院
歌代崇史                                    東京工業大学 博士(工学)
                                         北海学園大学
三角美樹                                    札幌開成高校
壽崎尚美                                    教材制作
片桐徳昭                                    札幌開成高校、博士進学



updated 2011-04-11 11:30 utc goh kawai
           undergraduate education
english language for freshmen
 online course
 instructor-led courses




updated 2011-04-11 11:30 utc goh kawai
           english online




updated 2011-04-11 11:30 utc goh kawai
           instructor-led course




updated 2011-04-11 11:30 utc goh kawai
           pronunciation lunch




updated 2011-04-11 11:30 utc goh kawai
           spoken language corpora course

acquire a specific practical skill
not theory
lots of out-of-class work




updated 2011-04-11 11:30 utc goh kawai
           objectives
re: spoken language corpora, explain:
  basic concepts (definitions, features)
  uses (analysis, engineering, learning)
  design and development strategies
re: speech analysis, perform:
  design and collect corpus
  label and analyze speech
  interpret analyses

updated 2011-04-11 11:30 utc goh kawai
           prerequisites

phonetics and phonology
 sound system of English and/or
   Japanese
 IPA desirable
audio input and output using computers
 bring your laptop (Linux, Windows,
   Mac)
statistics
 mean, standard deviation
updated 2011-04-11 11:30 utc goh kawai
           format of each class period
explain concepts and theory
collect and analyze speech
  learn software tools
  transcribe and analyze
  design corpus
learn about research and
 academia
explain next week's assignment
updated 2011-04-11 11:30 utc goh kawai
           grading
discussion and project   100%
essential
 classroom participation
 project




updated 2011-04-11 11:30 utc goh kawai
           schedule                                 •purple means assignment

wk          date                         activity    wk     date            activity

1     2011-04-12 install software                    8    2011-06-07 design L1 script

2     2011-04-19 transcribe speech                   9    2011-06-14 design L1 script

3     2011-04-26 record read speech                  10 2011-06-21 design L2 script

      2011-05-03 holiday                             11 2011-06-28 design L2 script
                             record spontaneous      12 2011-07-05 project report
4     2011-05-10
                             speech
5     2011-05-17 no class
                                                     13 2011-07-12 project report

6     2011-05-24 no class
                                                     14 2011-07-19 critique
                                                                     probably no class
7     2011-05-31 record read speech                  15 2011-07-26
                                                                     (make up day)



                                 attendance mandatory
updated 2011-04-11 11:30 utc goh kawai
           courseware
everything online
  reading material
  lecture notes (including this
    presentation)
  etc
http://goh.kawai.com/
http://goh.cll.hokudai.ac.jp/
 (inprog)
updated 2011-04-11 11:30 utc goh kawai
           Praat
http://www.praat.org/                   PRAAT

built by researchers and
 engineers in linguistics and
 speech processing
updated frequently
good support base
Windows, Mac, Linux
free
updated 2011-04-11 11:30 utc goh kawai
           what can Praat do?
record and play speech
display waveforms,
 spectrograms, pitch and more
label speech at various levels
  phone, mora, syllable, word,
   phrase and utterance levels
SIL fonts SIL
Praat in action PRAAT
updated 2011-04-11 11:30 utc goh kawai
           demo
view praat
 time waveform
 spectogram
 spectral slice
sound sources show praat
 vowels
 consonants
 pure tones (sinusoids)
updated 2011-04-11 11:30 utc goh kawai
           readings
Jurafsky et al (2000) chapter 4




updated 2011-04-11 11:30 utc goh kawai
           next week
install Praat
TIMIT sentences
       download from my website
       extract speech files from archive
       read files into Praat
       play speech
       view waveforms and spectograms
       label at the word level

updated 2011-04-11 11:30 utc goh kawai
           slideshow
if there's time




updated 2011-04-11 11:30 utc goh kawai
            see you next week!




                                         mailto:grad@kawai.com
updated 2011-04-11 11:30 utc goh kawai
                                           http://goh.kawai.com/

				
DOCUMENT INFO
Shared By:
Categories:
Tags:
Stats:
views:9
posted:10/4/2012
language:Unknown
pages:32