Alara Dirik, Hilal Dönmez, and Pınar Yanardağ received the Best Paper Award at the Controllable Generative Modeling in Language and Vision Workshop of the NeurIPS 2021 conference with their paper titled “Controlled Cue Generation for Play Scripts” which proposes a novel task of cue generation given a theater playscript dialogue. They approach the cue generation problem as a controlled text generation task with large-scale language models to steer the generation of the desired attributes without retraining.
