Wednesday, October 2, 2013
11:25 AM - 12:10 PM
|Level: ||Technology - Intermediate|
|Location: ||Grand Ballroom|
With the growing proliferation of television and video content, it is becoming increasingly difficult for consumers to find what they want to watch.
Conversational interfaces, built on knowledge graph-based semantic technologies, create intuitive ways for discovery and access to huge libraries of programming, bringing a next-generation user experience to television and video applications. The difference between traditional speech interfaces and conversational interfaces is substantial. Speech-driven interfaces typically use speech-to-text conversion, which maps to a menu or keywords for navigation. Conversational interfaces dramatically advance that experience by allowing users to converse with devices/applications using naturally spoken language, dispensing with menus and lexical matches for more semantic correlations using knowledge graph-based technologies. Intelligent conversational interfaces dynamically adapt to users' context, making user intent and behavior relevant. These interfaces simulate human-like intelligence, so users can talk to their TVs and connected devices instead of pushing buttons to find the content they want.
- With the growing proliferation of content, including thousands of video on demand titles, hundreds of TV channels and millions of online videos, it is becoming increasingly difficult for consumers to find what they want to watch.
- There are three areas where knowledge graph-based semantic technologies become an essential foundation for conversational interface technologies: speed, context and natural language processing.
- Speed - Conversational interfaces work at the speed of speech; they understand what a user has said and respond in a natural timeframe -- including responding for clarifications if the user has been ambiguous in the request.
- Context - Conversational interfaces have the ability to understand ongoing context - and change of context - through natural dialog process. For example, a user is not required to say, "I am done talking about movies, now I want to talk about sports" or issue other instructions or commands. The conversational system is context aware to allow the user to maintain natural dialog.
- Natural Language Processing- Conversational interfaces process natural language more effectively than simple lexical systems using the Knowledge Graph's semantic capabilities, thereby providing users more freedom in how they state their requests. In other words, users are not being trained or limited to using specific spoken command sets.
Sam Vasisht is a high-tech product/marketing executive with over 15 years' experience leading marketing/product management for Motorola, Bose, Real Networks, On2 Technologies (Google), and various startups. Sam has expertise in the digital media industry and a background of cable, Internet and wireless media. Sam developed and launched various industry-first products for emerging industry categories. At Motorola, Sam launched the first DOCSIS-certified residential gateway and wireless digital audio receiver. At Real Networks, Sam launched an early TVEverywhere platform and one of the first OTT video platforms. Before Veveo, Sam was president of marketing advisory firm 21TechMedia LLC, for venture-backed digital media technology companies. During five years with 21TechMedia, he helped venture-backed digital media technology startups with market entry, growth, turnarounds, and exits. As On2 Technologies' VP of Marketing, Sam positioned the company for subsequent acquisition by Google.