Search

1 Mar 2000: That system was constructed using HMMs trained on the Wall Street Journal (WSJ) corpus as a base and then adapted to individual data types of broadcast news data using supervised maximum ... The rest of the paper is arranged as follows. We first give

7 Oct 2001: A new scheme was developed for finding speaker clusters in found speech such as broadcast news. ... This consisted of two parts. Firstly the automatic transcription of 100 hours of broadcast news data and secondly the retrieval of documents relevant to

22 Nov 2006: This is especially thecase when there is no other additional constraints such as a lan-guage model, e.g. ... Major limitations of this paper are thatexperiments are conducted on artificially corrupted data and as-sume noise stationarity; however, recent

1 Mar 2000: That system was constructed using HMMs trained on the Wall Street Journal (WSJ) corpus as a base and then adapted to individual data types of broadcast news data using supervised maximum ... The rest of the paper is arranged as follows. We first give

3 Dec 2008: 1. IntroductionSpeech recognition systems are increasingly being built withfound data such as broadcast news and conversational tele-phone speech recordings. ... As DMTs arespeaker-independent, the same transforms can be used for thetraining and test data

19 May 2005: 1. INTRODUCTION. For many years automatic transcription of broadcast news (BN)and conversational telephone speech (CTS) data have been the twomain tasks for the research community of large vocabulary contin-uous ... How-ever such a simple linear

10 Oct 2007: 3. STT POST PROCESSING. In processing data such as Broadcast News (BN) or Broadcast Con-versations (BCs) for an STT system, the first stage is to segment thedata into homogeneous blocks, ... The language models for each of the systems were trainedon over

19 Dec 2006: Front-end CMLLRAs it is useful to compare the uncertainty decoding schemes toapproaches such as CMLLR, CMLLR can be modified to usea GMM front-end selection process. ... Future work will examine real found data, suchas broadcast news, and the application

10 Apr 2000: The rest of the paper is arranged as follows. We first give de-tails of the broadcast news data used, and then describe our workon segment processing (segmentation, classification and ... This corpus will be referred to as BN-train96. A further tranche

8 Mar 2000: The rest of the paper is arranged as follows. We first givedetails of the broadcast news data used in the experiments,and briefly describe our work on segment processing whichsplits the ... We havefound that using the full system with adaptation results