While deep learning models have achieved state-of-the-art accuracies for many prediction tasks, understanding these models remains a challenge. Despite the recent interest in developing visual tools to help users interpret deep learning models, the complexity and wide variety of models deployed in industry, and the large-scale datasets that they used, pose unique design challenges that are inadequately addressed by existing work. Through participatory design sessions with over 15 researchers and engineers at Facebook, we have developed, deployed, and iteratively improved ACTIVIS, an interactive visualization system for interpreting large-scale deep learning models and results. By tightly integrating multiple coordinated views, such as a computation graph overview of the model architecture, and a neuron activation view for pattern discovery and comparison, users can explore complex deep neural network models at both the instance- and subset-level. ACTIVIS has been deployed on Facebook's machine learning platform. We present case studies with Facebook researchers and engineers, and usage scenarios of how ACTIVIS may work with different models.

The human-computer dialogue research field has been studying interaction with computers since the early stage of Artificial Intelligence, however, research has often focused on very practical tasks to be completed with the dialogues. A new trend in the field tries to implement persuasive techniques with automated interactive agents; unlike booking a train ticket, for example, such dialogues require the system to show more anthropomorphic qualities. The influences of such qualities in the effectiveness of persuasive dialogue is only starting to be studied. In this article we focus on one important perceived trait of the system: personality, and explore how it influences the persuasiveness of a dialogue system. We introduce a new persuasive dialogue system and combine it with a state of the art personality utterance generator. By doing so, we can control the system’s extraversion personality trait and observe its influence on the user’s perception of the dialogue and its output. In particular, we observe that the user’s extraversion influences their perception of the dialogue and its persuasiveness, and that the perceived personality of the system can affect its trustworthiness and persuasiveness. We believe that theses observations will help to set up guidelines to tailor dialogue systems to the user’s interaction expectations and improve the persuasive interventions.