CURE: A High-Performance, Low-Power, and Reliable Network-on-Chip Design Using Reinforcement Learning

Mathematics

• Network on chip 100%
• Reinforcement Learning 75%
• High Performance 64%
• Power Consumption 60%
• Latency 49%
• Percent 45%
• Fault 43%
• Energy Efficiency 42%
• Router 37%
• Design 31%
• Hardware 30%
• Mean Time to Failure 19%
• Fault Tolerance 16%
• Error Correction 16%
• Control Policy 16%
• Chip 15%
• Life Span 15%
• Complex Dynamics 14%
• Architecture 13%
• Innovation 13%
• Optimise 13%
• Trade-offs 12%
• Benchmark 12%
• Controller 11%
• Interaction 9%
• Simulation 7%
• Performance 7%
• Framework 7%

Engineering & Materials Science

• Network-on-chip 56%
• Reinforcement learning 50%
• Electric power utilization 27%
• Routers 23%
• Energy efficiency 19%
• Time-to-failure 15%
• Computer hardware 14%
• Error correction 13%
• Fault tolerance 12%
• Controllers 6%