Search
Search Funnelback University
- Refined by:
- Date: 2015
21 -
30 of
30
search results for news |u:mi.eng.cam.ac.uk
Fully-matching results
-
lect9_pres
mi.eng.cam.ac.uk/~mjfg/local/4F10/lect9_pres.2up.pdf10 Nov 2015: Support Vector Machines For Texture Classification. • ewsRec, a SVM-driven Personal Recommendation Sys-tem for News Websites. • -
Annotating large lattices with the exact word error Rogier ...
mi.eng.cam.ac.uk/~mjfg/Kernel/van_dalen-2015-exact_error.pdf16 Jun 2015: This approximation becomes particularly problematicwith new types of acoustic models that require flexible align-ments. ... 4. ResultsTo test the new algorithm for error-marking lattices, a realistictraining set with long utterances was chosen. -
IMPROVING MULTIPLE-CROWD-SOURCED TRANSCRIPTIONS USING A…
mi.eng.cam.ac.uk/~mjfg/ALTA/publications/van_dalen-2015-improving.pdf11 Feb 2015: and a new system is built, applying minimum phone errortraining and speaker-adaptive training. ... 14] Daniel Jurafsky and James H. Martin, Speech and LanguageProcessing, Prentice-Hall, New Jersey, 2000. -
Department of Engineering 1 Monoids: e�cient segmental features…
mi.eng.cam.ac.uk/~mjfg/Kernel/van_dalen-2013-tr-monoid_features17 Apr 2015: Amore realistic strategy for decoding would prune the hypothesis space. is reporttherefore proposes a new, more exible class of features. ... e new features are as ecient as features used in Ragniand Gales (2012); van Dalen et al. -
main
mi.eng.cam.ac.uk/~cipolla/publications/article/2016-CVIU-EVTTS.pdf10 Sep 2015: New interactive technologies provide an opportunity for ASC individuals61to practice their communication skills. ... Errorsare decreased with each new iteration of the model. (a) (b) (c) (d) (e) (f) (g). -
main
mi.eng.cam.ac.uk/~cipolla/archive/Publications/article/2016-CVIU-EVTTS.pdf10 Sep 2015: New interactive technologies provide an opportunity for ASC individuals61to practice their communication skills. ... Errorsare decreased with each new iteration of the model. (a) (b) (c) (d) (e) (f) (g). -
Department of Engineering 1 Generative Kernels and Score-Spaces…
mi.eng.cam.ac.uk/~mjfg/Kernel/rcv25_2015_y3.pdf1 Jun 2015: e new features are as ecient as features used in Ragniand Gales (2012); van Dalen et al. ... Also, as new models for speechrecognition are becoming more complex and features richer, the xed time alignmentscould restrict performance. -
Model-based Approaches to Robust SpeechRecognition in Diverse…
mi.eng.cam.ac.uk/~mjfg/thesis_yw293.pdf27 Oct 2015: Two new. approaches to enforcing the independence constraint are proposed in Chapter 6. ... detect the beginning and the end of speech events; in a broadcast news transcription system,. -
"Refinements in Hierarchical Phrase-Based Translation…
mi.eng.cam.ac.uk/~wjb31/ppubs/jpino2015HieroRefinementsThesis.pdf6 Feb 2015: a new cabinet,responsible for handling daily affairs of the new government.”. ... The president accepted it and at the same time alsoasked him to form a new cabinet government to be responsiblefor routine administration.”. -
PROBABILISTIC ACOUSTIC MODELLING FOR PARAMETRIC SPEECH SYNTHESIS Sean …
mi.eng.cam.ac.uk/~wjb31/ppubs/shannon2014probabilistic-thesis.pdf2 Feb 2015: As we develop the new model and generation method we also investigate and develop. ... method. To address the second issue, we introduce a new GV-like generation method.
Search history
Recently clicked results
Recently clicked results
Your click history is empty.
Recent searches
Recent searches
Your search history is empty.