GTC ON-DEMAND

 
SEARCH SESSIONS
SEARCH SESSIONS

Search All
 
Refine Results:
 
Year(s)

SOCIAL MEDIA

EMAIL SUBSCRIPTION

 
 

GTC ON-DEMAND

Presentation
Media
Abstract:
The meta-data provided by content creators is not sufficient to create the exciting and customer-focused discovery and navigational experiences that premium video customers demand from their services these days. In our talk we will describe how Comcast uses machine learning and AI such as computer vision and natural language processing technologies to better understand the content distributed on our platform. We will conclude with examples of how this extracted information can then be used to create novel and compelling offerings which lead to a better customer experience and higher engagement with our products.
The meta-data provided by content creators is not sufficient to create the exciting and customer-focused discovery and navigational experiences that premium video customers demand from their services these days. In our talk we will describe how Comcast uses machine learning and AI such as computer vision and natural language processing technologies to better understand the content distributed on our platform. We will conclude with examples of how this extracted information can then be used to create novel and compelling offerings which lead to a better customer experience and higher engagement with our products.  Back
 
Topics:
Intelligent Video Analytics, Intelligent Machines, IoT & Robotics, AI & Deep Learning Research
Type:
Talk
Event:
GTC Silicon Valley
Year:
2019
Session ID:
S9875
Streaming:
Share:
 
Abstract:

We'll describe how Comcast uses GPU-powered machine learning to build intelligent products specifically focusing on the areas of smart home video analytics and virtual assistants for customer service. We'll explain how we use deep learning to alert our customers of noteworthy events being observed by their smart home cameras, and how it helps us to accurately understand the intent of our customers when they contact us via our virtual assistants and how we use reinforcement learning to identify how to best resolve their concerns. We'll also talk about how our distributed multi-GPU clusters speed up training the models and enable inference at Comcast scale.

We'll describe how Comcast uses GPU-powered machine learning to build intelligent products specifically focusing on the areas of smart home video analytics and virtual assistants for customer service. We'll explain how we use deep learning to alert our customers of noteworthy events being observed by their smart home cameras, and how it helps us to accurately understand the intent of our customers when they contact us via our virtual assistants and how we use reinforcement learning to identify how to best resolve their concerns. We'll also talk about how our distributed multi-GPU clusters speed up training the models and enable inference at Comcast scale.

  Back
 
Topics:
Consumer Engagement & Personalization, 5G & Edge, Speech & Language Processing
Type:
Talk
Event:
GTC Silicon Valley
Year:
2018
Session ID:
S8259
Streaming:
Share:
 
Abstract:
We'll describe the deep learning models behind Comcast's X1 Voice Remote and Smart Video Analytics and how we use GPUs to train and run these models. We'll explain how we can accurately parse the millions of voice queries we receive every day, how we automatically determine the domain of a query (TV, sports, billing, etc.), and how deep learning helps us understand what is happening on TV at any given moment. We'll also go into detail about how our distributed multi-GPU clusters speed up training the models and enable inference on millions of voice commands and hundreds of thousands video clips every day.
We'll describe the deep learning models behind Comcast's X1 Voice Remote and Smart Video Analytics and how we use GPUs to train and run these models. We'll explain how we can accurately parse the millions of voice queries we receive every day, how we automatically determine the domain of a query (TV, sports, billing, etc.), and how deep learning helps us understand what is happening on TV at any given moment. We'll also go into detail about how our distributed multi-GPU clusters speed up training the models and enable inference on millions of voice commands and hundreds of thousands video clips every day.  Back
 
Topics:
Artificial Intelligence and Deep Learning, Intelligent Video Analytics, Media and Entertainment
Type:
Talk
Event:
GTC Silicon Valley
Year:
2017
Session ID:
S7618
Download:
Share:
 
 
Previous
  • Amazon Web Services
  • IBM
  • Cisco
  • Dell EMC
  • Hewlett Packard Enterprise
  • Inspur
  • Lenovo
  • SenseTime
  • Supermicro Computers
  • Synnex
  • Autodesk
  • HP
  • Linear Technology
  • MSI Computer Corp.
  • OPTIS
  • PNY
  • SK Hynix
  • vmware
  • Abaco Systems
  • Acceleware Ltd.
  • ASUSTeK COMPUTER INC
  • Cray Inc.
  • Exxact Corporation
  • Flanders - Belgium
  • Google Cloud
  • HTC VIVE
  • Liqid
  • MapD
  • Penguin Computing
  • SAP
  • Sugon
  • Twitter
Next