From Conversation to Comprehension: How to design Interactive, Intelligent Tools for Understanding our Data
Abstract:
As we continue to produce data in ever-increasing quantities, the battle to organize, understand and use it rages on. Tools such as search and recommendation systems help us to find relevant information quickly, but with such complexity in the underlying data, traditional one-shot approaches are not quite enough. Conversational, interactive systems have been proposed over the years to engage the user in a dialog, aimed to guide them towards the right information in a more nuanced way. Despite these largely academic efforts, interactive search and recommendation systems that go beyond simple likes or thumbs-up remain relatively uncommon in mainstream applications. In this talk, I will discuss the state of the art in interactive search, recommendation and data analytics, including some of my own research from academia and from my work in the health industry.
Interactive Browsing Through Movies According
to Their Emotional Signature
Abstract:
I will present and demo a system that visualizes the emotional signature of films and enables users to browse through these emotional signatures. Two design trade-offs will be discussed. The first is the complexity of the interactions vs. ease-of-use and simplicity. The second is the complexity of the presented information vs. readability. I will also talk about analysis and browsing decisions, such as computing and displaying distances between emotional signatures and the choice to present emotional anchors.