W Y W X W Y WX W Y WX W Y WX O O X X O O O - PowerPoint PPT Presentation

• •

  W  Y W X

W Y  WX

W Y  WX

W Y  WX

O O X X O O O X O O X ↓ ↑ ↑ ↓ O O

→

M M

O O E E E O O E O E O E E O O O E vs. E O O O O E X O E E F O O O O O O E E E E E O O E E

Computational Costs: SVM During Learning F SVM Learning (W) Fastest Feedforward Learning (W) Illuminated Learning (M) Out of Computing Time (s) 7,000 Memory! 6,000 5,000 Out of Memory! 4,000 3,000 2,000 1,000 … 120 ! 0 Million 0 5 10 15 20 Matrix Size

Computational Costs: During Recognition SVM Feedforward AI (W) F Illuminated AI (M) Best Alternate AI (KNN) Without Optimization Computational Cost 10 8 per Test (s) Out of Memory! Even Faster with 6 Converted Networks 4 2 0 0 50 100 20 120 Millions Matrix Size

– –

14 25 12 12 10 20 10 8 15 8 6 6 10 4 4 5 2 2 0 0 0 -0.25 -0.2 -0.15 -0.1 -0.05 0 0.05 0.1 0.15 -0.5 0 0.5 1 1.5 -0.2 -0.1 0 0.1 0.2 0.3 0.4 0.5 0.6

W 11 I 1 O 1 W 11 * I 1 = + W 12 * I 2 + W 13 * I 3 … …

• Ideal "Expected" Patterns 1 2 3 4 5 6 7 8 9 0 • • •

• • • • • • • •