J
Jacob Devlin
Researcher at Google
Publications - 41
Citations - 58905
Jacob Devlin is an academic researcher from Google. The author has contributed to research in topics: Computer science & Machine translation. The author has an hindex of 23, co-authored 35 publications receiving 31122 citations. Previous affiliations of Jacob Devlin include Carnegie Mellon University & BBN Technologies.
Papers
More filters
Posted Content
Generating Natural Questions About an Image
TL;DR: In this article, the authors introduce the task of Visual Question Generation (VQG), where the system is tasked with asking a natural and engaging question when shown an image, and provide three datasets which cover a variety of images from object-centric to event-centric.
Posted Content
Language Model Pre-training for Hierarchical Document Representations
TL;DR: This work proposes algorithms for pre-training hierarchical document representations from unlabeled data which include fixed-length sentence/paragraph representations which integrate contextual information from the entire documents.
Posted Content
Semantic Code Repair using Neuro-Symbolic Transformation Networks
TL;DR: In this paper, the authors propose a two-stage approach where first a large set of repair candidates are generated by rule-based processors, and then these candidates are scored by a statistical model using a novel neural network architecture which they refer to as Share, Specialize, and Compete.
Proceedings ArticleDOI
Statistical Machine Translation Features with Multitask Tensor Networks
Hendra Setiawan,Zhongqiang Huang,Jacob Devlin,Thomas Lamar,Rabih Zbib,Richard Schwartz,John Makhoul +6 more
TL;DR: A three-pronged approach to improving Statistical Machine Translation (SMT), building on recent success in the application of neural networks to SMT, that augment the architecture of the neural network with tensor layers that capture important higher-order interaction among the network units.
Proceedings Article
Factored Soft Source Syntactic Constraints for Hierarchical Machine Translation
TL;DR: The proposed approach significantly improves a strong string-to-dependency translation system on multiple evaluation sets and keeps translation rules intact and factorizes the use of syntactic constraints through two separate models.