Search
Search Funnelback University
- Refined by:
- Date: 2015
11 -
20 of
30
search results for news |u:mi.eng.cam.ac.uk
Fully-matching results
-
A LANGUAGE SPACE REPRESENTATION FOR SPEECH RECOGNITION
mi.eng.cam.ac.uk/~mjfg/icassp15-ragni.pdf18 May 2015: Such an approach could not only re-duce development costs for each new language but also provide au-tomatic means for language analysis. ... Each language l =1,. ,L defines a new cluster. Trees for non-bias clusters are ini-tialised to have zero mean -
lect4.dvi
mi.eng.cam.ac.uk/~mjfg/local/4F10/lect4.pdf10 Nov 2015: We would. like a new estimate so that for the parameters at the k 1th. ... log-likelihood of the complete dataset in terms of the new. model parameters, θ(k1),. -
Improving Speech Recognition and Keyword Searchfor Low Resource…
mi.eng.cam.ac.uk/~ar527/mendels_is2015a.pdf3 Jun 2015: Wescrape text from multiple genres including blogs, online news,translated TED talks, and subtitles. ... source. Some of the sources were multilingual, suchas TED talks, and others were single language, such as blogsand news sources. -
A LANGUAGE SPACE REPRESENTATION FOR SPEECH RECOGNITION A. Ragni, ...
mi.eng.cam.ac.uk/~ar527/ragni_icassp2015a.pdf6 Apr 2015: Such an approach could not only re-duce development costs for each new language but also provide au-tomatic means for language analysis. ... Each language l =1,. ,L defines a new cluster. Trees for non-bias clusters are ini-tialised to have zero mean -
Improving Speech Recognition and Keyword Searchfor Low Resource…
mi.eng.cam.ac.uk/~mjfg/interspeech15-mendels-babel.pdf9 Oct 2015: Wescrape text from multiple genres including blogs, online news,translated TED talks, and subtitles. ... source. Some of the sources were multilingual, suchas TED talks, and others were single language, such as blogsand news sources. -
Joint Decoding of Tandem and Hybrid Systems for Improved ...
mi.eng.cam.ac.uk/~ar527/wang_is2015a.pdf3 Jun 2015: The acoustic log-likelihoods fromthe tandem GMM acoustic model and hybrid DNN are linearlycombined as new acoustic scores, which are used in the Viterbidecoding process. ... Tyagi, “New entropy based combi-nation rules in HMM/ANN multi-stream ASR,” -
lect4.dvi
mi.eng.cam.ac.uk/~mjfg/local/4F10/lect4_pres.pdf10 Nov 2015: We wouldlike a new estimate so that for the parameters at the k 1th. ... dzi. The second term is not dependent on the new model param-eters, the distribution of zi is known. -
lect8.dvi
mi.eng.cam.ac.uk/~mjfg/local/4F10/lect8_pres.pdf20 Nov 2015: Support Vector Machines For Texture Classification. • ewsRec, a SVM-driven Personal Recommendation Sys-tem for News Websites. • -
Association Between Femur Size and a FocalDefect of the ...
mi.eng.cam.ac.uk/reports/svr-ftp/gee_tr696.pdf3 Feb 2015: InSection 4, we discuss some important details of the methodology and also how the results shed new light onthe mechanisms by which femur size affects fracture risk. -
Joint Decoding of Tandem and Hybrid Systems for Improved ...
mi.eng.cam.ac.uk/~mjfg/interspeech15-wang-babel.pdf9 Oct 2015: The acousticlog-likelihoods from the tandem GMM acoustic model and hy-brid DNN are linearly combined as new acoustic scores, whichare used in the Viterbi decoding process. ... Tyagi, “New entropy based combi-nation rules in HMM/ANN multi-stream ASR,”
Search history
Recently clicked results
Recently clicked results
Your click history is empty.
Recent searches
Recent searches
Your search history is empty.