Search
Search Funnelback University
- Refined by:
- Date: 2012
Did you mean economiespast |u:mi.eng.cam.ac.uk?
1 -
10 of
10
search results for Economics test |u:mi.eng.cam.ac.uk
where 0
match all words and 10
match some words.
Results that match 1 of 2 words
-
paper.dvi
mi.eng.cam.ac.uk/~mjfg/yw293_ASRU11.pdf19 Jan 2012: Theoriginal AURORA4 task is derived from Wall Street Journal(WSJ0) 5k-word closed vocabulary task, with 330 utterancesfrom 8 speakers in the test set. ... Schemes Est. test Avg.01 04 08. VTS ML 43.8 48.9 55.7 49.5. -
paper.dvi
mi.eng.cam.ac.uk/~mjfg/sxz20_inter11.pdf19 Jan 2012: Test set A was usedas the development set for tuning parameters for all systems,such as the penalty factorC for the structured SVMs. ... Model Train/Test Set A Set B Set C Avg.HMM — 9.84 9.11 9.53 9.49. ( -
Department of Engineering 1 Generative Kernels and Score-Spaces…
mi.eng.cam.ac.uk/~mjfg/rcv25_2012_y1.pdf29 May 2012: ese scores then form features for the discriminative classiers. is. Test data O. ... Matched training data is available for these test conditions, but not used inthis work. -
SSVM_LVCSR_ASRU11.dvi
mi.eng.cam.ac.uk/~mjfg/sxz20_ASRU11.pdf19 Jan 2012: Thereare three test sets each includes0 20dB five SNRs. Set Awas used as the development set for tuning parameters forall systems, such as the penalty factorC in structured SVMs.The joint ... TheHMMs are state-clustered triphones (3140 states) with -
SSVM_LVCSR.dvi
mi.eng.cam.ac.uk/~mjfg/sxz_TASLP12.pdf19 Oct 2012: RES TRICTED BY LATTICES GENERATED BYHMM-VTS. Model Param. Criterion Test Set WER (%) AvgA B C D. ... SSVM LM (1-slack-µ) 7.5 13.3 11.1 19.6 15.4. training and test data. -
INFERENCE ALGORITHMS FOR GENERATIVE SCORE-SPACES A. Ragni and M. ...
mi.eng.cam.ac.uk/~ar527/20120117113826_193449_2122.pdf29 Jul 2012: The maximum word accuracy criterion was used to train dis-criminative models on multi-style data using suboptimal alignments.Test set A was used as the validation set to stop training. ... However, in this case the SDM givesimprovement on test set B only. -
A Variational Perspective onNoise-Robust Speech Recognition R. C. van …
mi.eng.cam.ac.uk/~mjfg/rcv25_ASRU11.pdf19 Jan 2012: All results are averaged over three of thefour available test sets, Feb 89, Oct 89, and Feb 91 (Sep 92is not used), a total of 30 test speakers and 900 utterances. -
chapter_revise.dvi
mi.eng.cam.ac.uk/~mjfg/noise_review10.pdf10 Feb 2012: In practice this is rarely the case. Originally the backgroundnoise was simply estimated from periods of “silence” in the test conditions. ... In particular the AURORA2 test set [27] has been used to evaluate a number of systems within a -
paper.dvi
mi.eng.cam.ac.uk/~mjfg/segdisc_2012.pdf19 Oct 2012: Whether these. 5In the machine learning literature the problem of handling amismatch between training and test conditions is sometimesreferredto as sample selection bias or covariat shift. ... As the length of the observation and word sequencesvary over -
1 Semi-Supervised Video Segmentation usingTree Structured Graphical…
mi.eng.cam.ac.uk/~cipolla/archive/Publications/article/2013-PAMI-Video-segmentation.pdf22 Dec 2012: We demonstrate the efficacy of our proposed algorithm usingseveral qualitative and quantitative tests on both foreground/background and multi-class video segmentation problems usingpublicly available and our own datasets. ... Dataset: Our test sequences
Search history
Recently clicked results
Recently clicked results
Your click history is empty.
Recent searches
Recent searches
Your search history is empty.