1 / 42

Neural Network Verification Part 4: Incomplete Methods

Explore techniques like Interval Bound Propagation & Linear Programming for robust neural network verification. Understand non-convex problem-solving & optimize erroneous outputs. Achieve error-safe deep learning models.

jelmer
Download Presentation

Neural Network Verification Part 4: Incomplete Methods

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript

  1. Neural Network VerificationPart 4: Incomplete Methods Some true properties can be proved true

  2. Condition on inputs Robust Deep Learning Post Is there an erroneous output? Error Safe Non-convexity makes the problem NP-hard

  3. Condition on inputs Robust Deep Learning Post Is there an erroneous output? Error Safe Replace by a convex superset

  4. Condition on inputs Robust Deep Learning Post Is there an erroneous output? Error Safe Say, non-convex set has no erroneous output

  5. Condition on inputs Robust Deep Learning Post Is there an erroneous output? Error Safe Convex superset might give incorrect answer

  6. Outline • Interval Bound Propagation • Linear Programming Relaxation • Results Mirman et al., 2018; Gowal et al., 2018

  7. Condition on inputs Robust Deep Learning Post Is there an erroneous output? Error Safe

  8. Condition on inputs Robust Deep Learning Post Is there an erroneous output? Error Safe Axis aligned convex superset

  9. Example Post -2 ≤ x1≤ 2 a [-4, 4] [0, 4] 1 -2 ≤ x2 ≤ 2 -1 x1 1 [-2, 2] ain = x1 + x2 z x2 [-2, 2] 1 -1 aout = max{ain,0} -1 b Minimum value of ain? -4 Minimum value of aout? 0 Maximum value of aout? 4 Maximum value of ain? 4

  10. Example Post -2 ≤ x1≤ 2 a [-4, 4] [0, 4] 1 -2 ≤ x2 ≤ 2 -1 x1 1 [-2, 2] bin = x1 - x2 z x2 [-2, 2] 1 -1 bout = max{bin,0} -1 b [-4, 4] [0, 4] Minimum value of bin? -4 Minimum value of bout? 0 Maximum value of bout? 4 Maximum value of bin? 4

  11. Example Post -2 ≤ x1≤ 2 a [-4, 4] [0, 4] 1 -2 ≤ x2 ≤ 2 -1 x1 1 [-2, 2] [-8, 0] bin = x1 - x2 z x2 [-2, 2] 1 -1 bout = max{bin,0} -1 b [-4, 4] [0, 4] z = -aout-bout Minimum value of z? -8 Maximum value of z? 0

  12. Outline • Interval Bound Propagation • Linear Programming Relaxation • Results Wong and Kolter, 2018

  13. Example Post min z s.t. -2 ≤ x1≤ 2 a 1 -2 ≤ x2 ≤ 2 -1 x1 1 [-2, 2] ain = x1 + x2 z x2 [-2, 2] 1 -1 bin = x1 - x2 -1 aout = max{ain,0} b bout = max{bin,0} z = - aout - bout

  14. Example Post min z Linear constraints s.t. -2 ≤ x1≤ 2 -2 ≤ x2 ≤ 2 Easy to handle ain = x1 + x2 bin = x1 - x2 aout = max{ain,0} bout = max{bin,0} z = - aout - bout

  15. Example Post min z s.t. -2 ≤ x1≤ 2 -2 ≤ x2 ≤ 2 ain = x1 + x2 bin = x1 - x2 aout = max{ain,0} Non-linear constraints bout = max{bin,0} NP-hard problem z = - aout - bout

  16. Relaxation Post ain ∈ [l,u] aout = max{ain,0} aout ain l u Ehlers 2017 Replace with convex superset

  17. Example Post min z s.t. -2 ≤ x1≤ 2 -2 ≤ x2 ≤ 2 ain = x1 + x2 bin = x1 - x2 aout = max{ain,0} bout = max{bin,0} z = - aout - bout

  18. Example Post min z Linear Program s.t. -2 ≤ x1≤ 2 -2 ≤ x2 ≤ 2 Several “efficient” solvers ain = x1 + x2 bin = x1 - x2 aout ≥ 0, aout≥ ain, aout ≤ 0.5ain + 2 bout ≥ 0, bout ≥ bin, bout ≤ 0.5bin + 2 z = - aout - bout

  19. Outline • Interval Bound Propagation • Linear Programming Relaxation • LP Duality • Results

  20. Example minx -3x1 - x2 - 2x3 7 x 2 x -x1≤ 0, -x2≤ 0, -x3≤ 0 s.t. x1 + x2 + 3x3≤ 30 3 x 2x1 + 2x2 + 5x3 ≤ 24 4x1 + x2 + 2x3 ≤ 36 Scale the constraints, add them up 3x1 + x2 + 2x3 ≤ 90

  21. Example minx -3x1 - x2 - 2x3 7 x 2 x -x1≤ 0, -x2≤ 0, -x3≤ 0 s.t. x1 + x2 + 3x3≤ 30 3 x 2x1 + 2x2 + 5x3 ≤ 24 4x1 + x2 + 2x3 ≤ 36 Scale the constraints, add them up Lower bound on solution -3x1 - x2 - 2x3 ≥ -90

  22. Example minx -3x1 - x2 - 2x3 1 x -x1≤ 0, -x2≤ 0, -x3≤ 0 s.t. x1 + x2 + 3x3≤ 30 2x1 + 2x2 + 5x3 ≤ 24 1 x 4x1 + x2 + 2x3 ≤ 36 Scale the constraints, add them up 3x1 + x2 + 2x3 ≤ 36

  23. Example minx -3x1 - x2 - 2x3 1 x -x1≤ 0, -x2≤ 0, -x3≤ 0 s.t. x1 + x2 + 3x3≤ 30 2x1 + 2x2 + 5x3 ≤ 24 1 x 4x1 + x2 + 2x3 ≤ 36 Scale the constraints, add them up Lower bound on solution -3x1 - x2 - 2x3 ≥ -36

  24. Example minx -3x1 - x2 - 2x3 1 x -x1≤ 0, -x2≤ 0, -x3≤ 0 s.t. x1 + x2 + 3x3≤ 30 2x1 + 2x2 + 5x3 ≤ 24 1 x 4x1 + x2 + 2x3 ≤ 36 Scale the constraints, add them up Tightest lower bound? -3x1 - x2 - 2x3 ≥ -36

  25. Example minx -3x1 - x2 - 2x3 y1 y2 y3 -x1≤ 0, -x2≤ 0, -x3≤ 0 s.t. y4 x1 + x2 + 3x3≤ 30 y5 2x1 + 2x2 + 5x3 ≤ 24 y6 4x1 + x2 + 2x3 ≤ 36 We should be able to add up the inequalities y1, y2, y3, y4, y5, y6 ≥ 0

  26. Example minx -3x1 - x2 - 2x3 y1 y2 y3 -x1≤ 0, -x2≤ 0, -x3≤ 0 s.t. y4 x1 + x2 + 3x3≤ 30 y5 2x1 + 2x2 + 5x3 ≤ 24 y6 4x1 + x2 + 2x3 ≤ 36 Coefficient of x1 should be 3 -y1+ y4 + 2y5 + 4y6 = 3

  27. Example minx -3x1 - x2 - 2x3 y1 y2 y3 -x1≤ 0, -x2≤ 0, -x3≤ 0 s.t. y4 x1 + x2 + 3x3≤ 30 y5 2x1 + 2x2 + 5x3 ≤ 24 y6 4x1 + x2 + 2x3 ≤ 36 Coefficient of x2 should be 1 -y2 + y4 + 2y5 + y6 = 1

  28. Example minx -3x1 - x2 - 2x3 y1 y2 y3 -x1≤ 0, -x2≤ 0, -x3≤ 0 s.t. y4 x1 + x2 + 3x3≤ 30 y5 2x1 + 2x2 + 5x3 ≤ 24 y6 4x1 + x2 + 2x3 ≤ 36 Coefficient of x3 should be 2 -y3 + 3y4 + 5y5 + 2y6 = 2

  29. Example minx -3x1 - x2 - 2x3 y1 y2 y3 -x1≤ 0, -x2≤ 0, -x3≤ 0 s.t. y4 x1 + x2 + 3x3≤ 30 y5 2x1 + 2x2 + 5x3 ≤ 24 y6 4x1 + x2 + 2x3 ≤ 36 Lower bound should be tightest maxy -30y4 - 24y5 - 36y6

  30. Dual maxy -30y4 - 24y5 - 36y6 s.t. y1, y2, y3, y4, y5, y6 ≥ 0 -y1+ y4 + 2y5 + 4y6 = 3 -y2 + y4 + 2y5 + y6 = 1 -y3 + 3y4 + 5y5 + 2y6 = 2 Original problem is called primal Dual of dual is primal

  31. minxcTx Primal s.t. A x ≤ b maxy≥0 bTy Dual s.t.ATy = c

  32. Weak Duality Post Any feasible primal solution x Any feasible dual solution y Primal value at x ≥ Dual value at y Dual provides a lower bound

  33. Strong Duality Post Optimal primal solution x* Optimal dual solution y* Primal value at x* = Dual value at y* Some mild conditions have to be satisfied

  34. Example Post min z Linear Program s.t. -2 ≤ x1≤ 2 -2 ≤ x2 ≤ 2 Dual feasible solution ain = x1 + x2 Evaluated using dual network bin = x1 - x2 aout ≥ 0, aout≥ ain, aout ≤ 0.5ain + 2 bout ≥ 0, bout ≥ bin, bout ≤ 0.5bin + 2 z = - aout - bout

  35. Outline • Interval Bound Propagation • Linear Programming Relaxation • Results

  36. Experimental Setup Post Input model is ε-perturbation of an image Robust deep learning

  37. MNIST with ε = 0.1 Post Nominal: Test error = 0.65%, PGD = 27.72%

  38. MNIST with ε = 0.3 Post Nominal: Test error = 0.65%, PGD = 99.63%

  39. CIFAR-10 with ε = 8/255 Post * Nominal: Test error = 16.66%, PGD = 100% * PGD error not available (replaced by Incomplete)

  40. Results Post Interval bounds are useful in practice Significantly faster than LP relaxation methods Requires careful tuning of more hyperparameters Still a long way to go to reach “real” networks

  41. Robustness Metric for MNIST Post

  42. Questions? References for other incomplete methods on tutorial webpage

More Related