;;; Dribble file "full-adder.drib" started T > (load "full-adder-data") ;;; Loading source file "full-adder-data.lisp" 8 training patterns successfully loaded. #P"/tmp_mnt/u2/csfaculty/rjw/teach/com1410/learn/full-adder-data.lisp" > (show-training-data) ((0 0 0) (0 0)) ((0 0 1) (0 1)) ((0 1 0) (0 1)) ((0 1 1) (1 0)) ((1 0 0) (0 1)) ((1 0 1) (1 0)) ((1 1 0) (1 0)) ((1 1 1) (1 1)) > (init-backprop-net 3) Building net with 3 input units (plus bias), 3 hidden units (plus bias), and 2 output units. Weights initialized to small random values. DONE > (show-backprop-weights) Input-to-hidden weights: -0.01 -0.43 0.04 -0.31 0.46 -0.26 0.44 0.20 0.33 0.27 0.19 0.07 Hidden-to-output weights: -0.11 -0.22 0.32 0.46 0.27 0.44 -0.12 0.27 > (train-backprop-net) 0 passes through data: RMS error = 0.5149 25 passes through data: RMS error = 0.4875 50 passes through data: RMS error = 0.4540 75 passes through data: RMS error = 0.4119 100 passes through data: RMS error = 0.3845 125 passes through data: RMS error = 0.3694 150 passes through data: RMS error = 0.3605 175 passes through data: RMS error = 0.3545 200 passes through data: RMS error = 0.3499 225 passes through data: RMS error = 0.3472 250 passes through data: RMS error = 0.3444 275 passes through data: RMS error = 0.3412 300 passes through data: RMS error = 0.3375 325 passes through data: RMS error = 0.3331 350 passes through data: RMS error = 0.3278 375 passes through data: RMS error = 0.3212 400 passes through data: RMS error = 0.3129 425 passes through data: RMS error = 0.3027 450 passes through data: RMS error = 0.2914 475 passes through data: RMS error = 0.2802 500 passes through data: RMS error = 0.2702 525 passes through data: RMS error = 0.2618 550 passes through data: RMS error = 0.2551 575 passes through data: RMS error = 0.2498 600 passes through data: RMS error = 0.2455 625 passes through data: RMS error = 0.2419 650 passes through data: RMS error = 0.2388 675 passes through data: RMS error = 0.2357 700 passes through data: RMS error = 0.2324 725 passes through data: RMS error = 0.2283 750 passes through data: RMS error = 0.2232 775 passes through data: RMS error = 0.2166 800 passes through data: RMS error = 0.2088 825 passes through data: RMS error = 0.2007 850 passes through data: RMS error = 0.1920 875 passes through data: RMS error = 0.1839 900 passes through data: RMS error = 0.1752 925 passes through data: RMS error = 0.1667 950 passes through data: RMS error = 0.1589 975 passes through data: RMS error = 0.1512 1000 passes through data: RMS error = 0.1440 1025 passes through data: RMS error = 0.1375 1050 passes through data: RMS error = 0.1310 1075 passes through data: RMS error = 0.1256 1100 passes through data: RMS error = 0.1209 1125 passes through data: RMS error = 0.1165 1150 passes through data: RMS error = 0.1122 1175 passes through data: RMS error = 0.1087 1200 passes through data: RMS error = 0.1052 1225 passes through data: RMS error = 0.1024 1250 passes through data: RMS error = 0.0995 1275 passes through data: RMS error = 0.0974 1300 passes through data: RMS error = 0.0955 1325 passes through data: RMS error = 0.0939 1350 passes through data: RMS error = 0.0919 1375 passes through data: RMS error = 0.0907 1400 passes through data: RMS error = 0.0891 1425 passes through data: RMS error = 0.0882 1450 passes through data: RMS error = 0.0873 1475 passes through data: RMS error = 0.0864 1500 passes through data: RMS error = 0.0859 1525 passes through data: RMS error = 0.0855 1550 passes through data: RMS error = 0.0849 1575 passes through data: RMS error = 0.0846 1600 passes through data: RMS error = 0.0842 1625 passes through data: RMS error = 0.0841 1650 passes through data: RMS error = 0.0830 Patterns all learned to within tolerance 0.10 after 1663 passes. DONE > (show-backprop-weights) Input-to-hidden weights: -3.27 -3.27 -3.27 8.03 -4.62 -4.63 -4.62 0.86 5.93 5.93 5.93 -8.30 Hidden-to-output weights: -3.95 -1.42 5.33 0.27 -6.83 -7.26 -6.52 9.72 > (test-backprop-net '(0 0 0)) Input: 0.00 0.00 0.00 1.00 Hidden: 1.00 0.70 0.00 1.00 Output: 0.01 0.10 Output attribute 0: 0.01 Output attribute 1: 0.10 > (test-backprop-net '(0 0 1)) Input: 0.00 0.00 1.00 1.00 Hidden: 0.99 0.02 0.09 1.00 Output: 0.04 0.90 Output attribute 0: 0.04 Output attribute 1: 0.90 > (test-backprop-net '(0 1 0)) Input: 0.00 1.00 0.00 1.00 Hidden: 0.99 0.02 0.09 1.00 Output: 0.04 0.90 Output attribute 0: 0.04 Output attribute 1: 0.90 > (test-backprop-net '(0 1 1)) Input: 0.00 1.00 1.00 1.00 Hidden: 0.82 0.00 0.97 1.00 Output: 0.90 0.10 Output attribute 0: 0.90 Output attribute 1: 0.10 > (test-backprop-net '(1 0 0)) Input: 1.00 0.00 0.00 1.00 Hidden: 0.99 0.02 0.09 1.00 Output: 0.04 0.90 Output attribute 0: 0.04 Output attribute 1: 0.90 > (test-backprop-net '(1 0 1)) Input: 1.00 0.00 1.00 1.00 Hidden: 0.82 0.00 0.97 1.00 Output: 0.90 0.10 Output attribute 0: 0.90 Output attribute 1: 0.10 > (test-backprop-net '(1 1 0)) Input: 1.00 1.00 0.00 1.00 Hidden: 0.82 0.00 0.97 1.00 Output: 0.90 0.10 Output attribute 0: 0.90 Output attribute 1: 0.10 > (test-backprop-net '(1 1 1)) Input: 1.00 1.00 1.00 1.00 Hidden: 0.15 0.00 1.00 1.00 Output: 0.99 0.90 Output attribute 0: 0.99 Output attribute 1: 0.90 > (dribble) ;;; Dribble file "full-adder.drib" finished