Generalized Agreement for Bidirectional Word AlignmentGeneralized Agreement • August 12th, 2015
While agreement-based joint training has proven to deliver state-of-the-art align- ment accuracy, the produced word align- ments are usually restricted to one-to- one mappings because of the hard cons- traint on agreement. We propose a ge- neral framework to allow for arbitrary loss functions that measure the disagreement between asymmetric alignments. The loss functions can not only be defined between asymmetric alignments but al- so between alignments and other latent structures such as phrase segmentations. We use a Viterbi EM algorithm to train the joint model since the inference is intractable. Experiments on Chinese- English translation show that joint training with generalized agreement achieves sig- nificant improvements over two state-of- the-art alignment methods.
Generalized Agreement Underlying a Multicasting EnvironmentGeneralized Agreement • July 27th, 2002
The BA problem is solved in a multicasting environment. The proposed protocol uses the minimum number of message exchanges to reach an agreement while tolerating the maximum number of faulty components in the distributed system. It makes all the fault-free processors reach a common value to keep the system from the influences of processor failures and transmission medium failures.
