Large language models are typically trained densely: all parameters are
...
We present Branch-Train-Merge (BTM), a communication-efficient algorithm...
Self-supervised pre-training of large-scale transformer models on text
c...
Models trained on large unlabeled corpora of human interactions will lea...
We seek to create agents that both act and communicate with other agents...
We present our view of what is necessary to build an engaging open-domai...
Dialogue research tends to distinguish between chit-chat and goal-orient...
Dialogue research tends to distinguish between chit-chat and goal-orient...
Generative dialogue models currently suffer from a number of problems wh...
While dialogue remains an important end-goal of natural language researc...
Beyond understanding what is being discussed, human communication requir...