Improving sentence compression by learning to predict gaze
Sigrid Klerke, Yoav Goldberg and Anders Søgaard
This paper shows how to use gaze measurements obtained via eye-tracking to improve models for deletion-based sentence compression. It uses information from different data sources, eliminating the need to reply on data that is doubly annotated with sentence-compression and eye-tracking information, and points to data utilization methodology that could lead to larger improvements in the future.
Patterns of Wisdom: Discourse-Level Style in Multi-Sentence Quotations
Kyle Booten and Marti A. Hearst
A Joint Model of Orthography and Morphological Segmentation
Ryan Cotterell, Tim Vieira and Hinrich Schütze
Feuding Families and Former Friends; Unsupervised Learning for Dynamic Fictional Relationships
Mohit Iyyer, Anupam Guha, Snigdha Chaturvedi, Jordan Boyd-Graber and Hal Daumé III
This paper addresses a novel problem of modeling interpersonal states between characters and how they evolve overtime. It brings new insights into unsupervised relationship modeling which involves joint learning a set of relationship descriptors and their trajectories. It was selected to receive a best paper award for its originality.
Learning to Compose Neural Networks for Question Answering
Jacob Andreas, Marcus Rohrbach, Trevor Darrell and Dan Klein
This paper is noteworthy for the innovative algorithmic ideas which bring together the strengths of neural network models and compositional logical semantics, and it demonstrates this approach in two different experiments, text-based question answering as well as image-based question answering, showing impact in multiple communities. The paper is pointing in a good direction for people to continue to expand, challenge and enhance this work.
Multi-way, Multilingual Neural Machine Translation with a Shared Attention Mechanism
Orhan Firat, Kyunghyun Cho and Yoshua Bengio
Black Holes and White Rabbits; Metaphor Identification with Visual Features
Ekaterina Shutova, Douwe Kiela and Jean Maillard