An End‐to‐End Conversational Second‐Screen Application for TV Program Discovery

Abstract

Articles FALL 2015 73T he recent explosion of content (such as movies, TV shows, sports, and others) available on television coupled with an increase use in mobile devices (that is, smartphones and tablets) has created significant interest in second-screen applications from both end users and content providers. Second screen applications are designed to run from mobile devices and to enhance the television viewing experience in numerous ways, one of which is helping end users effectively find and control content on television through spoken natural language (that is, conversational TV program discovery).Conversational TV program discovery applications have recently become available in the marketplace from select cable/satellite providers. However, these applications are limited. They support a predefined set of utterance types (for example, switch to <channel>, find a <genre>movie, and find a movie with <actor >). Hence, end users must conform to

Dialog state tracking, a machine reading approach using Memory Network

“…Various approaches have been proposed to define dialog state trackers. The traditional methods used in most commercial implementations use hand-crafted rules that typically rely on the most likely result from an NLU module (Yeh et al, 2014) and hardly models uncertainty. However, these rule-based systems are prone to frequent errors as the most likely result is not always the correct one (Williams, 2014).…”