Joseph Turian

Université de Montréal, Computing, Post-Doc

Followers

177

Following

116

Public Views

Joseph Turian has been working on artificial intelligence research since 1996. His focus is on using sophisticated machine learning techniques to approach large-scale problems in natural language. He is currently a post-doctoral research fellow at the Université de Montréal, studying deep learning methods with Professor Yoshua Bengio, Canada Research Chair in Statistical Learning Algorithms. Dr. Turian defended his dissertation, "Constituent Parsing by Classification" at New York University. He received his AB from Harvard University in Computer Science (cum laude).
Supervisors: I. Dan Melamed

less

Interests

Uploads

Papers by Joseph Turian

Towards Purely Discriminative Training for Tree-Structured Translation Models

Discriminative training methods have recently led to significant advances in the state of the art... more Discriminative training methods have recently led to significant advances in the state of the art of machine translation (MT). Another promising trend is the incorporation of syntactic information into MT systems. Combining these trends is difficult for reasons of system complexity and computational complexity. This study makes progress toward a syntax-aware MT system whose every component is trained discriminatively. Our main innovation is an approach to discriminative learning that is computationally efficient enough for large statistical MT systems, yet whose accuracy on translation subtasks is near the state of the art.

Joseph Turian

Uploads

Papers by Joseph Turian

Log In