GTC ON-DEMAND

 
SEARCH SESSIONS
SEARCH SESSIONS

Search All
 
Refine Results:
 
Year(s)

SOCIAL MEDIA

EMAIL SUBSCRIPTION

 
 

GTC ON-DEMAND

Presentation
Media
Abstract:
Learn about recent achievements in deep reinforcement learning (RL) with a focus on using large-scale compute resources. We'll cover basic algorithms, discuss development of RL agents for playing Atari games, and provide a chronology of implementations leveraging increasing amounts of hardware to achieve better results faster. We will describe large-scale RL projects in which learned agents surpassed human-level performance in the challenging games of Go, Quake III, and Dota2. For each project, we'll discuss the distinct hardware used and the techniques we developed to scale up the algorithm. These projects demonstrate a range of strategies for harnessing many CPUs and GPUs. We'll outline continued research in this area and explore the potential for more exciting results around the corner.
Learn about recent achievements in deep reinforcement learning (RL) with a focus on using large-scale compute resources. We'll cover basic algorithms, discuss development of RL agents for playing Atari games, and provide a chronology of implementations leveraging increasing amounts of hardware to achieve better results faster. We will describe large-scale RL projects in which learned agents surpassed human-level performance in the challenging games of Go, Quake III, and Dota2. For each project, we'll discuss the distinct hardware used and the techniques we developed to scale up the algorithm. These projects demonstrate a range of strategies for harnessing many CPUs and GPUs. We'll outline continued research in this area and explore the potential for more exciting results around the corner.  Back
 
Topics:
Deep Learning & AI Frameworks, AI & Deep Learning Research
Type:
Talk
Event:
GTC Silicon Valley
Year:
2019
Session ID:
S9786
Streaming:
Download:
Share:
 
Abstract:
Deep reinforcement learning (RL) has achieved many recent successes, yet experiment turn-around time remains a key bottleneck in research and in practice. We investigate how to optimize existing deep RL algorithms for modern computers, specifically for a combination of CPUs and GPUs. We confirm that both policy gradient and Q-value learning algorithms can be adapted to learn using many parallel simulator instances. We further find it possible to train using batch sizes considerably larger than are standard, without negatively affecting sample complexity or final performance. We leverage these facts to build a unified framework for parallelization that dramatically hastens experiments in both classes of algorithm. All neural network computations use GPUs, accelerating both data collection and training. Our results include using an entire NVIDIA DGX-1 to learn successful strategies in Atari games in single-digit minutes, using both synchronous and asynchronous algorithms.
Deep reinforcement learning (RL) has achieved many recent successes, yet experiment turn-around time remains a key bottleneck in research and in practice. We investigate how to optimize existing deep RL algorithms for modern computers, specifically for a combination of CPUs and GPUs. We confirm that both policy gradient and Q-value learning algorithms can be adapted to learn using many parallel simulator instances. We further find it possible to train using batch sizes considerably larger than are standard, without negatively affecting sample complexity or final performance. We leverage these facts to build a unified framework for parallelization that dramatically hastens experiments in both classes of algorithm. All neural network computations use GPUs, accelerating both data collection and training. Our results include using an entire NVIDIA DGX-1 to learn successful strategies in Atari games in single-digit minutes, using both synchronous and asynchronous algorithms.  Back
 
Topics:
Deep Learning & AI Frameworks, Tools & Libraries, Performance Optimization
Type:
Talk
Event:
GTC Silicon Valley
Year:
2018
Session ID:
S8272
Streaming:
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