XMA: A Crossbar-aware Multi-task Adaption Framework via Shift-based Mask Learning Method

Mathematics

• Mask 100%
• Learning 71%
• Programming 55%
• High Energy 48%
• Framework 44%
• Cell 43%
• Neural Networks 36%
• Energy Efficiency 33%
• Energy Efficient 32%
• Energy Consumption 30%
• Accelerator 27%
• Masking 27%
• Backbone 25%
• Leverage 25%
• Engine 24%
• Hardware 23%
• Learning Algorithm 23%
• High Efficiency 23%
• Innovation 20%
• High Accuracy 20%
• Experimental Results 19%
• Convolution 16%
• Binary 14%
• Design 12%
• Model 6%

Engineering & Materials Science

• RRAM 82%
• Masks 74%
• Deep neural networks 38%
• Data storage equipment 23%
• Computer programming 20%
• Particle accelerators 19%
• Convolution 19%
• Learning algorithms 16%
• Durability 15%
• Energy efficiency 15%
• Energy utilization 13%
• Engines 12%
• Computer hardware 11%