Zhanpeng Jin Allen C Cheng zhj6pittedu acc33pittedu ASPLOS 2010 The Wild and Crazy Session VIII Artificial Neural Network Source Anatomy and Physiology ID: 915590
Download Presentation The PPT/PDF document "Engineering Next-generation Self-healing..." is the property of its rightful owner. Permission is granted to download and print the materials on this web site for personal, non-commercial use only, and to display it on your personal computer provided you do not modify the materials and that you retain all copyright notices contained in the materials. By downloading content from our website, you accept the terms of this agreement.
Slide1
Engineering Next-generation Self-healing And Self-optimizing Neural Network Based Medical Platforms
Zhanpeng Jin Allen C. Chengzhj6@pitt.edu acc33@pitt.edu
ASPLOS 2010, The Wild and Crazy Session VIII
Slide2Artificial Neural Network
(Source: "
Anatomy and Physiology
" by the US National Cancer Institute's Surveillance, Epidemiology and End Results (SEER) Program.)
(Source:
CapitalISM
BI)
Slide3Emerging Neural Hardware
Neural chip (384 neurons + 100,000 synapses) @ FACET Project
Microchip of a neural network
@ RIEC, Tohoku University
The China-Brain Project (10,000 – 15,000 neural nets) @ Hugo de
Garis
3-D neural chip for better visualization @
CalTech
Slide4Bio-inspired Autonomously Reconfigurable Mechanism
Autonomous
Reconfigurability
Topology Adaptation
Redundancy
(Picture Source: “
Brain Injury: Recovery
” in Psychology Wiki)
Slide5Autonomously Reconfigurable ANN
Autonomously Reconfigurable ANN (ARANN) Based Medical Processing Platforms
Error Detection
µController
Run-Time Reconfiguration Controller
Mask-Based
Topology Adaption
Coarse-/Fine-Grained Hybrid Reconfiguration
N
ew topology
Flash Memory
Coarse-grained Reconfiguration
Bitstream
Database
Current ANN Configuration
Fine-grained
Reconfiguration
Bitstream
New Configuration
Select/Merge
Bitstreams
Coarse-grained
Reconfiguration
Bitstream
Error Scale
and Location
Outputs
Data Traffic
Request
Request
Inputs
Sensors/
Database
Diagnosis/
Controllers
Slide6Virtual-physical Neuron Mapping
Neural Topology
Physical Neurons
Virtual-to-Physical
(V2P)
Neuron Mapping
Topology
Adaptation
(Coarse-grained)
Connectionism Evolvement
(Fine-grained)
Autonomous ANN
+
=
Slide7Mask-based Topology Adaptation
Slide8WACI Conclusion
Systems are increasingly vulnerable to unexpected faults and defects, especially for emerging biomedical systems.Non-invasive autonomous reconfigurability is promising, particularly for ANN-based biomedical platforms. Autonomously adapting ANN’s behaviors and structures, both algorithmically
and
microarchitecturally
.
Neuron Virtualization
helps to decouple the fault scale and reduce the reconfiguration latency (idle time).
Mask-based Topology Adaptation
can achieve significant reduction of design complexity and spatial overhead.
Slide9Thanks for Listening
Questions?(This work is supported by NSF No. 0832990)