Conferences system

11 мая 2011 года, Москва, офис Яндекса.

11 мая, 18.30  в московском офисе Яндекса в рамках Цикла научных семинаров выступит Dirk Van Compernolle, эксперт в области технологий распознавания речи.

Тема доклада: From Speech Recognition to Audio Search.

Web search technology has been a main contributor to the success of the internet.  That a single query may generate millions of hits is sometimes annoying; but despite this seemingly poor result 'searching on the web' has become one of the top daily activities of many people. Capabilities of today's systems are still largely restricted to text based search, while searching via images or audio is still in its infancy.
Dirk Van Compernolle about his presentation: «In this talk we will focus on the problem of audio search and the underlying speech recognition technology.  In the first part of the talk we will review the principles of statistical speech recognition. 
In the second part of the talk we will review some new developments in the speech recognition field that allow for searching audio files with no and/or incomplete transcriptions. This is either achieved by performing the search at an intermediate level (phonetic transcription level) or by doing a search directly at the acoustic level».

Dirk Van Compernolle is a Professor at the K.U.Leuven where he is teaching on speech recognition; his current research focuses on novel paradigms in speech recognition, most notably the usage of example based techniques.

ВНИМАНИЕ! Докладчик будет читать на английском языке, но свои вопросы на английском или русском языке вы можете задать уже сегодня. 
Во время семинара будет доступна онлайн-трансляция, которую можно будет посмотреть тут.

Начало мероприятия в 18:30, а подтвердить регистрацию вы можете с 18.00 (лучше подойти к этому времени, чтобы успеть к началу лекции).