Related Topics
Experts
Publications (54)

1Top Accuracy and Fast Dependency Parsing is not a Contradiction2010  Bernd BohnetKernel is F (x, y) = −→w ∗ φ(x, y) where −→w is the weight vector and the size of −→w is n. Algorithm 2 shows the update...

2Random Restarts in Minimum Error Rate Training for Statistical Machine Translation2008  Robert C. Moore,Chris Quirkseries of onedimensional optimizations of the feature weight vector, using an innovative line search that takes advantage...

3Loglinear weight optimisation via Bayesian Adaptation in Statistical Machine Translationbecome unstable and fail in obtaining an appropriate weight vector Λ. However, it is quite common to have a great amount...

4Mining Words in the Minds of Second Language Learners: LearnerSpecific Word Difficultyproperties we want predictors to have. interpretable weight vector Most predictors use weight vectors trained with data....

5A Lazy Learning Model for Entity Linking using QuerySpecific Informationare represented as a feature vector Xi ∈ χ , and u is a weight vector. Estimate u on Aq. A popular method for finding u is empirical...

6A HighPerformance Syntactic and Semantic Dependency Parserx mapping that mapped the features to indices of the weight vector by a random function. Usually, the featureindex mapping...

7Robust Learning in Random Subspaces: Equipping NLP for OOV Effects2012  Anders Søgaard,Anders Johannsenvariable is never instantiated in our test data, the learned weight vector will give us the decision boundary TEST(2DSVC) (the dashed...

8Kernel Slicing: Scalable Online Training with Conjunctive Featuresfor (φd(x))i 6= 0, and w is a weight vector in the expanded feature space, Fd. The weight vector w is computed from S and α:...

9Inducing Crosslingual Distributed Representations of Wordskeeps a weight vector for each task. Assuming that at time t the algorithm has made s mistakes, the compound weight vector at...

10EasyFirst Chinese POS Tagging and Dependency Parsing⃗⃗ Here, ⃗⃗ is the model’s parameter or weight vector and is the feature vector extracted from ( )...