SEARCH SESSIONS

Search All
 
Refine Results:
 
Year(s)

SOCIAL MEDIA

EMAIL SUBSCRIPTION

 
 

GTC ON-DEMAND

Acoustics and Audio Processing
Presentation
Media
Efficient Automatic Speech Recognition on the GPU
Speakers:
Jike Chong
Abstract:

Automatic speech recognition (ASR) technology is emerging as a critical component in data analytics for a wealth of media data being generated everyday. ASR-based applications contain fine-grained concurrency that has great potential to be exploited on the GPU. However, the state-of-art ASR algorithm involves a highly parallel graph traversal on an irregular graph with millions of states and arcs, making efficient parallel implementations highly challenging. We present four generalizable techniques including: dynamic data-gather buffer, find-unique, lock-free data structures using atomics, and hybrid global/local task queues. When used together, these techniques can effectively resolve ASR implementation challenges on an NVIDIA GPU.

 
Topics:
Acoustics and Audio Processing
Type:
Poster
Event:
GTC Silicon Valley
Year:
2010
Session ID:
P10C02
Download:
Share:
Artificial Intelligence and Deep Learning
Presentation
Media
Efficient Automatic Speech Recognition on the GPU
Speakers:
Jike Chong
Abstract:

Learn about how the GPU is able to meet the challenges of implementing automatic speech recognition (ASR), gain insights into the data-parallel implementation techniques that can provide 10x faster performance compared to sequentially processing ASR on a CPU. The state-of-art algorithm for ASR performs a graph traversal on a large, irregular graph with millions of states and arcs, guided by speech input only known at runtime. We present four generalizable techniques including: dynamic data-gather buffer, find-unique, lock-free data structures using atomics, and hybrid global/local task queues. When used together, these techniques can effectively resolve ASR implementation challenges on a GPU.

 
Topics:
Artificial Intelligence and Deep Learning, Acoustics and Audio Processing, Developer - Algorithms
Type:
Talk
Event:
GTC Silicon Valley
Year:
2010
Session ID:
P10C02
Streaming:
Download:
Share: