Related Topics
Experts

1

2

3

4

5

6

7

8

9

10
Publications (74)

1Adaptive Development Data Selection for Loglinear Model in Statistical Machine Translationshow that our method can effectively adapt loglinear model parameters to different test data, and consistently achieves good...

2Loglinear weight optimisation via Bayesian Adaptation in Statistical Machine Translationwellknown Bayesian learning paradigm for adapting the model parameters. Since stateoftheart statistical machine translation...

3EMDC: A Semisupervised Approach for Word Alignmenttune weights for discriminative models, while using the model parameters, model scores or alignment links from unsupervised word...

4Phrasal Segmentation Models for Statistical Machine Translationequation (2) by estimating the phrasal segmentation model parameters from naturally occuring phrase sequences in a large...

5Semisupervised Representation Learning for Domain Adaptation using Dynamic Dependency Networksi=1 log P(X i , Z i θ) where θ denotes the set of model parameters. Let q(Y ) be any nonzero distribution over hidden...

6Improving Word Alignment Quality using Morphosyntactic Information2004  Hermann Ney,Maja Popovicf C(f˜ , e) The procedure is similar for the other model parameters, i.e. alignment and fertility probabilities. For models...

7A Unified Approach in SpeechtoSpeech Translation: Integrating Features of Speech recognition and Machine Translationalgorithm (Press et al., 2000) as our tool to optimize model parameters, λM1 , based on different objective translation metrics...

8Symmetric Word Alignments for Statistical Machine Translationcorpus are used as a development corpus to optimize model parameters that are not trained via the EM algorithm, e.g. the...

9Latent Mixture of Discriminative Experts for Multimodal Prediction Modeling∑ h P (y  h, x, θ)P (h  x, θ) (1) where θ is the model parameters that is to be estimated from training data. To keep...

10A comparison of unsupervised methods for PartofSpeech Tagging in Chineseby assigning a probability distribution over the model parameters as a prior distribution, . In HMM, we calculate...