330 likes | 495 Views
Spoken language corpora Course overview. goh kawai 2013- 04-09 tue1 week1 spoken language corpora s316. goh do this for tue1. bring and connect laptop, projector, network, bluetooth speaker, clicker arrange desks, chairs show these slides, my website, glexa circulate roster sheet.
E N D
Spoken language corpora Course overview gohkawai2013-04-09 tue1 week1spoken language corporas316
goh do this for tue1 • bring and connectlaptop, projector, network, bluetooth speaker, clicker • arrange desks, chairs • show these slides, my website, glexa • circulate roster sheet
make roster • write • full name • furigana • email address pass sheet
informed consent • your speech and actions maybe 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
contact info • office: office building room s304 • email: grad@kawai.com • web: goh.kawai.com
goh's website • http://goh.kawai.com/ • http://goh.cll.hokudai.ac.jp/ • identical content • hokudai site may be faster
instructor • Goh Kawai (河合 剛かわい ごう) • born in Tokyo, raised in Toronto • came to Sapporo in 2003-04
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
goh’s vocational background • Xerox Palo Alto Research CenterPalo Alto, CA • SRI International Menlo Park, CA • University of Tokyo Tokyo, Japan • University of California Santa Cruz Santa Cruz, CA • Oregon Health & Science University • Beaverton, OR
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)
office hours • drop-inor email for appointment • no phone calls • off campus • see my website
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とめる。☆修士条件 査読のある国際会議で論文発表。☆博士条件 後進の研究指導。☆指導生の発表先 音響学会、音声学会、教育工学会、ASA, AAAL, Calico, Eurocall, Interspeechなど。
alumni • 平野宏子東京大学 博士(科学)東北師範大学 • 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 • 壽崎尚美北海道立高校 • 片桐徳昭札幌開成高校、博士(学術)見込
undergraduate education • english language for freshmen • online course • instructor-led courses
spoken language corpora course • acquire a specific practical skill • not theory • lots of out-of-class work
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
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
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
grading • discussion and project 100% • essential • participate in discussion during class • propose and report your project
schedule attendance mandatory
courseware • everything online • reading material • lecture notes (including this presentation) • http://goh.kawai.com/ • http://goh.cll.hokudai.ac.jp/ • hokudai library catalog of our course's textbooks • view online course offering (シラバス)
Praat • http://www.praat.org/ • built by researchers and engineers in linguistics and speech processing • updated frequently • good support base • Windows, Mac, Linux • free
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 • Praat in action
demo • view praat • time waveform • spectogram • spectral slice • sound sources show praat • vowels • consonants • pure tones (sinusoids)
readings • Jurafsky et al (2000) chapter 4
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
slideshow • if there's time
one-stop website • http://goh.kawai.com/ • link to glexa • course material (these slides) • contact form
see you next week! mailto:grad@kawai.com http://goh.kawai.com/