hack the derivative
play

Hack the Derivative! Erik Taubeneck Software Engineer October - PowerPoint PPT Presentation

Nov 09, 2022 •324 likes •1.39k views

The Derivative Finite Difference Floating Point Arithmetic Complex Analysis Crash Course Hack the Derivative Hack the Derivative! Erik Taubeneck Software Engineer October 20th, 2015 American University Math/Stat Dept Colloquium The

  1. The Derivative Finite Difference Floating Point Arithmetic Complex Analysis Crash Course Hack the Derivative Floating Point Numbers Multiplication is helped by Associativity and Commutativity (5 . 916829373 × 10 23 ) × (7 . 208209342 × 10 − 51 )

  2. The Derivative Finite Difference Floating Point Arithmetic Complex Analysis Crash Course Hack the Derivative Floating Point Numbers Multiplication is helped by Associativity and Commutativity (5 . 916829373 × 10 23 ) × (7 . 208209342 × 10 − 51 ) 5 . 916829373 × 10 23 × 7 . 208209342 × 10 − 51 =

  3. The Derivative Finite Difference Floating Point Arithmetic Complex Analysis Crash Course Hack the Derivative Floating Point Numbers Multiplication is helped by Associativity and Commutativity (5 . 916829373 × 10 23 ) × (7 . 208209342 × 10 − 51 ) 5 . 916829373 × 10 23 × 7 . 208209342 × 10 − 51 = 5 . 916829373 × 7 . 208209342 × 10 23 × 10 − 51 =

  4. The Derivative Finite Difference Floating Point Arithmetic Complex Analysis Crash Course Hack the Derivative Floating Point Numbers Multiplication is helped by Associativity and Commutativity (5 . 916829373 × 10 23 ) × (7 . 208209342 × 10 − 51 ) 5 . 916829373 × 10 23 × 7 . 208209342 × 10 − 51 = 5 . 916829373 × 7 . 208209342 × 10 23 × 10 − 51 = (5 . 916829373 × 7 . 208209342) × (10 23 × 10 − 51 ) =

  5. The Derivative Finite Difference Floating Point Arithmetic Complex Analysis Crash Course Hack the Derivative Floating Point Numbers Multiplication is helped by Associativity and Commutativity (5 . 916829373 × 10 23 ) × (7 . 208209342 × 10 − 51 ) 5 . 916829373 × 10 23 × 7 . 208209342 × 10 − 51 = 5 . 916829373 × 7 . 208209342 × 10 23 × 10 − 51 = (5 . 916829373 × 7 . 208209342) × (10 23 × 10 − 51 ) = but this doesn’t work for addition

  6. The Derivative Finite Difference Floating Point Arithmetic Complex Analysis Crash Course Hack the Derivative Floating Point Numbers Multiplication is helped by Associativity and Commutativity (5 . 916829373 × 10 23 ) × (7 . 208209342 × 10 − 51 ) 5 . 916829373 × 10 23 × 7 . 208209342 × 10 − 51 = 5 . 916829373 × 7 . 208209342 × 10 23 × 10 − 51 = (5 . 916829373 × 7 . 208209342) × (10 23 × 10 − 51 ) = but this doesn’t work for addition (5 . 916829373 × 10 23 ) + (7 . 208209342 × 10 − 51 )

  7. The Derivative Finite Difference Floating Point Arithmetic Complex Analysis Crash Course Hack the Derivative Floating Point Gaps In [1]: import sys In [2]: plus_epsilon_identity(x, eps): return x + eps == x In [3]: eps = sys.float_info.min

  8. The Derivative Finite Difference Floating Point Arithmetic Complex Analysis Crash Course Hack the Derivative Floating Point Gaps In [1]: import sys In [2]: plus_epsilon_identity(x, eps): return x + eps == x In [3]: eps = sys.float_info.min In [4]: plus_epsilon_identity(0.0, eps) Out [4]: False In [5]: plus_epsilon_identity(1.0, eps) Out [5]: True

  9. The Derivative Finite Difference Floating Point Arithmetic Complex Analysis Crash Course Hack the Derivative Floating Point Gaps In [1]: import sys In [2]: plus_epsilon_identity(x, eps): return x + eps == x In [3]: eps = sys.float_info.min In [4]: plus_epsilon_identity(0.0, eps) Out [4]: False In [5]: plus_epsilon_identity(1.0, eps) Out [5]: True In [6]: plus_epsilon_identity(1.0e20, 1.0) Out [6]: True

  10. The Derivative Finite Difference Floating Point Arithmetic Complex Analysis Crash Course Hack the Derivative Problems with the Derivative If we choose h that is so small such that float ( x ) + float ( h ) = float ( x )

  11. The Derivative Finite Difference Floating Point Arithmetic Complex Analysis Crash Course Hack the Derivative Problems with the Derivative If we choose h that is so small such that float ( x ) + float ( h ) = float ( x ) then f ( x + h ) − f ( x ) = f ( x ) − f ( x ) = 0 h = 0 h h

  12. The Derivative Finite Difference Floating Point Arithmetic Complex Analysis Crash Course Hack the Derivative Problems with the Derivative Similarly, if float ( f ( x + h )) = float ( f ( x ))

  13. The Derivative Finite Difference Floating Point Arithmetic Complex Analysis Crash Course Hack the Derivative Problems with the Derivative Similarly, if float ( f ( x + h )) = float ( f ( x )) then f ( x + h ) − f ( x ) = 0 h = 0 h

  14. The Derivative Finite Difference Floating Point Arithmetic Complex Analysis Crash Course Hack the Derivative Measuring the Gaps In [1]: def eps(x): e = float(max(sys.float_info.min, abs(x))) while not plus_epsilon_identity(x, e): last = e e = e / 2. return last

  15. The Derivative Finite Difference Floating Point Arithmetic Complex Analysis Crash Course Hack the Derivative Measuring the Gaps In [1]: def eps(x): e = float(max(sys.float_info.min, abs(x))) while not plus_epsilon_identity(x, e): last = e e = e / 2. return last In [2]: eps(1.0) Out[2]: 2.220446049250313e-16

  16. The Derivative Finite Difference Floating Point Arithmetic Complex Analysis Crash Course Hack the Derivative Measuring the Gaps In [1]: def eps(x): e = float(max(sys.float_info.min, abs(x))) while not plus_epsilon_identity(x, e): last = e e = e / 2. return last In [2]: eps(1.0) Out[2]: 2.220446049250313e-16 In [3]: eps(1.0e13) Out[3]: 0.0011102230246251565

  17. The Derivative Finite Difference Floating Point Arithmetic Complex Analysis Crash Course Hack the Derivative Measuring the Gaps In [1]: def eps(x): e = float(max(sys.float_info.min, abs(x))) while not plus_epsilon_identity(x, e): last = e e = e / 2. return last In [2]: eps(1.0) Out[2]: 2.220446049250313e-16 In [3]: eps(1.0e13) Out[3]: 0.0011102230246251565 In [4]: eps(1.0e19) Out[4]: 1110.2230246251565

  18. The Derivative Finite Difference Floating Point Arithmetic Complex Analysis Crash Course Hack the Derivative Choosing Epsilon Going back to our error estimation, ∞ f ( n ) ( x ) � h ( n − 1) E ( h , x ) = n ! n =1 we can add in a term, R ( f , x , h ) which accounts for the rounding error. The total error then becomes ∞ f ( n ) ( x ) h ( n − 1) + R ( f , x , h ) � E ( h , x ) = n ! n =1

  19. The Derivative Finite Difference Floating Point Arithmetic Complex Analysis Crash Course Hack the Derivative Choosing Epsilon It turns out that a good estimate 2 is h = √ u ∗ max ( | x | , 1) where u = eps (1). 2 http://www.karenkopecky.net/Teaching/eco613614/Notes_ NumericalDifferentiation.pdf

  20. The Derivative Finite Difference Floating Point Arithmetic Complex Analysis Crash Course Hack the Derivative Choosing Epsilon It turns out that a good estimate 2 is h = √ u ∗ max ( | x | , 1) where u = eps (1). In [1]: def finite_difference(f, x, h=None): if not h: h = sqrt(eps(1.0)) * max(abs(x), 1.0) return (f(x+h) - f(x))/h 2 http://www.karenkopecky.net/Teaching/eco613614/Notes_ NumericalDifferentiation.pdf

  21. The Derivative Finite Difference Floating Point Arithmetic Complex Analysis Crash Course Hack the Derivative Testing It Out In [2]: def error(f, df, x): return abs(finite_difference(f, x) - df(x)) In [3]: def error_rate(f, df, x): return error(f, df, x) / df(x)

  22. The Derivative Finite Difference Floating Point Arithmetic Complex Analysis Crash Course Hack the Derivative Testing It Out In [2]: def error(f, df, x): return abs(finite_difference(f, x) - df(x)) In [3]: def error_rate(f, df, x): return error(f, df, x) / df(x) In [4]: f, df = lambda x:x**2, lambda x:2*x

  23. The Derivative Finite Difference Floating Point Arithmetic Complex Analysis Crash Course Hack the Derivative Testing It Out In [2]: def error(f, df, x): return abs(finite_difference(f, x) - df(x)) In [3]: def error_rate(f, df, x): return error(f, df, x) / df(x) In [4]: f, df = lambda x:x**2, lambda x:2*x In [5]: error_rate(f, df, 1.0) Out[5]: 7.450580596923828e-09

  24. The Derivative Finite Difference Floating Point Arithmetic Complex Analysis Crash Course Hack the Derivative Testing It Out In [2]: def error(f, df, x): return abs(finite_difference(f, x) - df(x)) In [3]: def error_rate(f, df, x): return error(f, df, x) / df(x) In [4]: f, df = lambda x:x**2, lambda x:2*x In [5]: error_rate(f, df, 1.0) Out[5]: 7.450580596923828e-09 In [6]: error_rate(f, df, 1.0e5) Out[6]: 9.045761108398438e-09

  25. The Derivative Finite Difference Floating Point Arithmetic Complex Analysis Crash Course Hack the Derivative Testing It Out In [2]: def error(f, df, x): return abs(finite_difference(f, x) - df(x)) In [3]: def error_rate(f, df, x): return error(f, df, x) / df(x) In [4]: f, df = lambda x:x**2, lambda x:2*x In [5]: error_rate(f, df, 1.0) Out[5]: 7.450580596923828e-09 In [6]: error_rate(f, df, 1.0e5) Out[6]: 9.045761108398438e-09 In [7]: error_rate(f, df, 1.0e20) Out[7]: 4.5369065472e-09

  26. The Derivative Finite Difference Floating Point Arithmetic Complex Analysis Crash Course Hack the Derivative Testing It Out In [8]: import math In [9]: f, df = math.sin, math.cos

  27. The Derivative Finite Difference Floating Point Arithmetic Complex Analysis Crash Course Hack the Derivative Testing It Out In [8]: import math In [9]: f, df = math.sin, math.cos In [10]: error_rate(f, df, 1.0) Out[10]: 1.2780011808656197e-08

  28. The Derivative Finite Difference Floating Point Arithmetic Complex Analysis Crash Course Hack the Derivative Testing It Out In [8]: import math In [9]: f, df = math.sin, math.cos In [10]: error_rate(f, df, 1.0) Out[10]: 1.2780011808656197e-08 In [11]: error_rate(f, df, 1.0e10) Out[11]: 1.002014830004253

  29. The Derivative Finite Difference Floating Point Arithmetic Complex Analysis Crash Course Hack the Derivative Testing It Out In [8]: import math In [9]: f, df = math.sin, math.cos In [10]: error_rate(f, df, 1.0) Out[10]: 1.2780011808656197e-08 In [11]: error_rate(f, df, 1.0e10) Out[11]: 1.002014830004253 In [12]: error_rate(f, df, 1.0e20) Out[12]: 0.9999999999998509

  30. The Derivative Finite Difference Floating Point Arithmetic Complex Analysis Crash Course Hack the Derivative Testing It Out In [8]: import math In [9]: f, df = math.sin, math.cos In [10]: error_rate(f, df, 1.0) Out[10]: 1.2780011808656197e-08 In [11]: error_rate(f, df, 1.0e10) Out[11]: 1.002014830004253 In [12]: error_rate(f, df, 1.0e20) Out[12]: 0.9999999999998509 We can do better

  31. The Derivative Finite Difference Floating Point Arithmetic Complex Analysis Crash Course Hack the Derivative Testing It Out In [8]: import math In [9]: f, df = math.sin, math.cos In [10]: error_rate(f, df, 1.0) Out[10]: 1.2780011808656197e-08 In [11]: error_rate(f, df, 1.0e10) Out[11]: 1.002014830004253 In [12]: error_rate(f, df, 1.0e20) Out[12]: 0.9999999999998509 We can do better using Complex Analyis.

  32. The Derivative Finite Difference Floating Point Arithmetic Complex Analysis Crash Course Hack the Derivative A Crash Course in Complex Analysis • Def: i = √− 1.

  33. The Derivative Finite Difference Floating Point Arithmetic Complex Analysis Crash Course Hack the Derivative A Crash Course in Complex Analysis • Def: i = √− 1. • Why do we need i ???

  34. The Derivative Finite Difference Floating Point Arithmetic Complex Analysis Crash Course Hack the Derivative A Crash Course in Complex Analysis • Def: i = √− 1. • Why do we need i ??? • Solve x 2 + 1 = 0

  35. The Derivative Finite Difference Floating Point Arithmetic Complex Analysis Crash Course Hack the Derivative A Crash Course in Complex Analysis • Def: i = √− 1. • Why do we need i ??? • Solve x 2 + 1 = 0 • Def: C = { x + iy ∀ x , y ∈ R 2 }

  36. The Derivative Finite Difference Floating Point Arithmetic Complex Analysis Crash Course Hack the Derivative A Crash Course in Complex Analysis E x : f ( z ) = z 2

  37. The Derivative Finite Difference Floating Point Arithmetic Complex Analysis Crash Course Hack the Derivative A Crash Course in Complex Analysis E x : f ( z ) = z 2 Let z = x + iy . Then f ( z ) = f ( x + iy ) = ( x + iy ) 2 = x 2 + 2 ixy + i 2 y 2 = x 2 + 2 ixy − y 2

  38. The Derivative Finite Difference Floating Point Arithmetic Complex Analysis Crash Course Hack the Derivative A Crash Course in Complex Analysis E x : f ( z ) = z 2 Let z = x + iy . Then f ( z ) = f ( x + iy ) = ( x + iy ) 2 = x 2 + 2 ixy + i 2 y 2 = x 2 + 2 ixy − y 2 We define u ( x , y ) = R ( f ( x + iy )) and v ( x , y ) = I ( f ( x + iy ))

  39. The Derivative Finite Difference Floating Point Arithmetic Complex Analysis Crash Course Hack the Derivative A Crash Course in Complex Analysis E x : f ( z ) = z 2 Let z = x + iy . Then f ( z ) = f ( x + iy ) = ( x + iy ) 2 = x 2 + 2 ixy + i 2 y 2 = x 2 + 2 ixy − y 2 We define u ( x , y ) = R ( f ( x + iy )) and v ( x , y ) = I ( f ( x + iy )) In our example u ( x , y ) = x 2 − y 2 and v ( x , y ) = 2 xy

  40. The Derivative Finite Difference Floating Point Arithmetic Complex Analysis Crash Course Hack the Derivative A Crash Course in Complex Analysis u ( x , y ) = x 2 − y 2 and v ( x , y ) = 2 xy We can now take 4 different derivatives. ∂ u ∂ v ∂ x = 2 x ∂ x = 2 y ∂ u ∂ v ∂ y = − 2 y ∂ y = 2 x

  41. The Derivative Finite Difference Floating Point Arithmetic Complex Analysis Crash Course Hack the Derivative A Crash Course in Complex Analysis u ( x , y ) = x 2 − y 2 and v ( x , y ) = 2 xy We can now take 4 different derivatives. ∂ u ∂ v ∂ x = 2 x ∂ x = 2 y ∂ u ∂ v ∂ y = − 2 y ∂ y = 2 x Note that: ∂ u ∂ x = ∂ v ∂ u ∂ y = − ∂ v ∂ y ∂ x

  42. The Derivative Finite Difference Floating Point Arithmetic Complex Analysis Crash Course Hack the Derivative Cauchy-Riemann Equations These are the Cauchy-Riemann equations and hold for all analytic functions on C ! 3 ∂ u ∂ x = ∂ v ∂ u ∂ y = − ∂ v ∂ y ∂ x 3 See your favorite Complex Analysis textbook or Dr. Casey for a proof.

  43. The Derivative Finite Difference Floating Point Arithmetic Complex Analysis Crash Course Hack the Derivative Time for the Hack! Given f : C → C such that f : R → X ⊆ R , we can rewrite f ( z ) = f ( x + iy ) = u ( x , y ) + iv ( x , y )

  44. The Derivative Finite Difference Floating Point Arithmetic Complex Analysis Crash Course Hack the Derivative Time for the Hack! Given f : C → C such that f : R → X ⊆ R , we can rewrite f ( z ) = f ( x + iy ) = u ( x , y ) + iv ( x , y ) Now, for all ¯ z ∈ R , we have y = 0, and so f (¯ z ) = f ( x , 0) = u ( x , 0) + iv ( x , 0)

  45. The Derivative Finite Difference Floating Point Arithmetic Complex Analysis Crash Course Hack the Derivative Time for the Hack! Since f : R → X ⊆ R , f (¯ z ) ∈ R for all ¯ z ∈ R . Therefore v ( x , 0) = 0

  46. The Derivative Finite Difference Floating Point Arithmetic Complex Analysis Crash Course Hack the Derivative Time for the Hack! Since f : R → X ⊆ R , f (¯ z ) ∈ R for all ¯ z ∈ R . Therefore v ( x , 0) = 0 and so f (¯ z ) = f ( x , 0) = u ( x , 0) + iv ( x , 0) = u ( x , 0)

  47. The Derivative Finite Difference Floating Point Arithmetic Complex Analysis Crash Course Hack the Derivative Time for the Hack! Since f : R → X ⊆ R , f (¯ z ) ∈ R for all ¯ z ∈ R . Therefore v ( x , 0) = 0 and so f (¯ z ) = f ( x , 0) = u ( x , 0) + iv ( x , 0) = u ( x , 0) We want to estimate df ∂ x . Since for all z ∈ R , f = u , ∂ x = ∂ u ∂ x = ∂ v df ∂ y by the Cauchy-Riemann equations.

  48. The Derivative Finite Difference Floating Point Arithmetic Complex Analysis Crash Course Hack the Derivative Time for the Hack! Again, using the definition of the derivative ∂ v v ( x , y + h ) − v ( x , y ) ∂ y = lim h h → 0

  49. The Derivative Finite Difference Floating Point Arithmetic Complex Analysis Crash Course Hack the Derivative Time for the Hack! Again, using the definition of the derivative ∂ v v ( x , y + h ) − v ( x , y ) ∂ y = lim h h → 0 Since y = 0 for all ¯ z ∈ R , v ( x , h ) − v ( x , 0) ∂ v ∂ y = lim h h → 0

  50. The Derivative Finite Difference Floating Point Arithmetic Complex Analysis Crash Course Hack the Derivative Time for the Hack! Recall that v ( x , 0) = 0, so ∂ v v ( x , h ) ∂ y = lim h h → 0

  51. The Derivative Finite Difference Floating Point Arithmetic Complex Analysis Crash Course Hack the Derivative Time for the Hack! Recall that v ( x , 0) = 0, so ∂ v v ( x , h ) ∂ y = lim h h → 0 Now, to estimate df ∂ x = ∂ v ∂ y , for a very small h df ∂ x ≈ v ( x , h ) = I ( f ( x + ih )) h h

  52. The Derivative Finite Difference Floating Point Arithmetic Complex Analysis Crash Course Hack the Derivative Time for the Hack! We’ve now replaced f ( x + h ) − f ( x ) h with I ( f ( x + ih )) . h Note that we have eliminated the floating point addition!

  53. The Derivative Finite Difference Floating Point Arithmetic Complex Analysis Crash Course Hack the Derivative Complex Step Finite Difference Implemented in Python In [1]: import sys In [2]: def complex_step_finite_diff(f, x): h = sys.float_info.min return (f(x+h*1.0j)).imag / h

  54. The Derivative Finite Difference Floating Point Arithmetic Complex Analysis Crash Course Hack the Derivative Complex Step Finite Difference Implemented in Python In [3]: f, df = lambda x:x**2, lambda x:2*x

  55. The Derivative Finite Difference Floating Point Arithmetic Complex Analysis Crash Course Hack the Derivative Complex Step Finite Difference Implemented in Python In [3]: f, df = lambda x:x**2, lambda x:2*x In [4]: complex_step_finite_diff(f, 1.0) Out[4]: 2.0

  56. The Derivative Finite Difference Floating Point Arithmetic Complex Analysis Crash Course Hack the Derivative Complex Step Finite Difference Implemented in Python In [3]: f, df = lambda x:x**2, lambda x:2*x In [4]: complex_step_finite_diff(f, 1.0) Out[4]: 2.0 In [5]: complex_step_finite_diff(f, 1.0e10) Out[5]: 20000000000.0

  57. The Derivative Finite Difference Floating Point Arithmetic Complex Analysis Crash Course Hack the Derivative Complex Step Finite Difference Implemented in Python In [3]: f, df = lambda x:x**2, lambda x:2*x In [4]: complex_step_finite_diff(f, 1.0) Out[4]: 2.0 In [5]: complex_step_finite_diff(f, 1.0e10) Out[5]: 20000000000.0 In [6]: complex_step_finite_diff(f, 1.0e20) Out[6]: 2e+20

  58. The Derivative Finite Difference Floating Point Arithmetic Complex Analysis Crash Course Hack the Derivative Testing It Out In [7]: def cerror(f, df, x): return abs(complex_step_finite_diff(f, x) - df(x)) In [8]: def cerror_rate(f, df, x): return cerror(f, df, x) / x

  59. The Derivative Finite Difference Floating Point Arithmetic Complex Analysis Crash Course Hack the Derivative Testing It Out In [7]: def cerror(f, df, x): return abs(complex_step_finite_diff(f, x) - df(x)) In [8]: def cerror_rate(f, df, x): return cerror(f, df, x) / x In [9]: cerror_rate(f, df, 1.0) Out[9]: 0.0

  60. The Derivative Finite Difference Floating Point Arithmetic Complex Analysis Crash Course Hack the Derivative Testing It Out In [7]: def cerror(f, df, x): return abs(complex_step_finite_diff(f, x) - df(x)) In [8]: def cerror_rate(f, df, x): return cerror(f, df, x) / x In [9]: cerror_rate(f, df, 1.0) Out[9]: 0.0 In [10]: cerror_rate(f, df, 1.0e10) Out[10]: 0.0

  61. The Derivative Finite Difference Floating Point Arithmetic Complex Analysis Crash Course Hack the Derivative Testing It Out In [7]: def cerror(f, df, x): return abs(complex_step_finite_diff(f, x) - df(x)) In [8]: def cerror_rate(f, df, x): return cerror(f, df, x) / x In [9]: cerror_rate(f, df, 1.0) Out[9]: 0.0 In [10]: cerror_rate(f, df, 1.0e10) Out[10]: 0.0 In [11]: cerror_rate(f, df, 1.0e20) Out[11]: 0.0

  62. The Derivative Finite Difference Floating Point Arithmetic Complex Analysis Crash Course Hack the Derivative Testing It Out In [7]: def cerror(f, df, x): return abs(complex_step_finite_diff(f, x) - df(x)) In [8]: def cerror_rate(f, df, x): return cerror(f, df, x) / x In [9]: cerror_rate(f, df, 1.0) Out[9]: 0.0 In [10]: cerror_rate(f, df, 1.0e10) Out[10]: 0.0 In [11]: cerror_rate(f, df, 1.0e20) Out[11]: 0.0

  63. The Derivative Finite Difference Floating Point Arithmetic Complex Analysis Crash Course Hack the Derivative Testing It Out In [11]: import cmath In [11]: f, df = cmath.sin, cmath.cos

  64. The Derivative Finite Difference Floating Point Arithmetic Complex Analysis Crash Course Hack the Derivative Testing It Out In [11]: import cmath In [11]: f, df = cmath.sin, cmath.cos In [12]: cerror_rate(f, df, 1.0) Out[12]: 0.0

  65. The Derivative Finite Difference Floating Point Arithmetic Complex Analysis Crash Course Hack the Derivative Testing It Out In [11]: import cmath In [11]: f, df = cmath.sin, cmath.cos In [12]: cerror_rate(f, df, 1.0) Out[12]: 0.0 In [13]: cerror_rate(f, df, 1.0e10) Out[13]: 0.0

  66. The Derivative Finite Difference Floating Point Arithmetic Complex Analysis Crash Course Hack the Derivative Testing It Out In [11]: import cmath In [11]: f, df = cmath.sin, cmath.cos In [12]: cerror_rate(f, df, 1.0) Out[12]: 0.0 In [13]: cerror_rate(f, df, 1.0e10) Out[13]: 0.0 In [14]: cerror_rate(f, df, 1.0e20) Out[14]: 0.0

  67. The Derivative Finite Difference Floating Point Arithmetic Complex Analysis Crash Course Hack the Derivative It’s Not Perfect, But It’s Pretty Good In [15]: cerror_rate(f, df, 1.0e5) Out[15]: .110933124815182e-16

  68. The Derivative Finite Difference Floating Point Arithmetic Complex Analysis Crash Course Hack the Derivative It’s Not Perfect, But It’s Pretty Good In [15]: cerror_rate(f, df, 1.0e5) Out[15]: .110933124815182e-16 In [16]: f = lambda x: cmath.exp(x) / cmath.sqrt(x) In [17]: df = lambda x: (cmath.exp(x)* (2*x - 1))/(2*x**(1.5))

  69. The Derivative Finite Difference Floating Point Arithmetic Complex Analysis Crash Course Hack the Derivative It’s Not Perfect, But It’s Pretty Good In [15]: cerror_rate(f, df, 1.0e5) Out[15]: .110933124815182e-16 In [16]: f = lambda x: cmath.exp(x) / cmath.sqrt(x) In [17]: df = lambda x: (cmath.exp(x)* (2*x - 1))/(2*x**(1.5)) In [18]: cerror_rate(f, df, 1.0) Out[18]: 0.0

Recommend

Hack the Derivative! Erik Taubeneck GameChanger Media - Software Engineer and Data Maven Were

Hack the Derivative! Erik Taubeneck GameChanger Media - Software Engineer and Data Maven Were

Calculus Crash Course Finite Difference Floating Point Arithmetic Complex Analysis Crash Course Hack the Derivative Hack the Derivative! Erik Taubeneck GameChanger Media - Software Engineer and Data Maven Were hiring!

1.21k views • 89 slides
SLIPPERY SLIDES HACK.|100% WORKING!|NEW METHOD|HACK TOOL. Free No Ads Hard-Boiled Egg Hack Goes

SLIPPERY SLIDES HACK.|100% WORKING!|NEW METHOD|HACK TOOL. Free No Ads Hard-Boiled Egg Hack Goes

SLIPPERY SLIDES HACK.|100% WORKING!|NEW METHOD|HACK TOOL. Free No Ads Hard-Boiled Egg Hack Goes Viral on Twitter By using our site, you acknowledge that you have read and understand our Cookie PolicyPrivacy Policyand our Terms of Service.

1.1k views • 3 slides
2. Theory of the Derivative 2.1 Tangent Lines 2.2 Definition of Derivative 2.3 Rates of Change

2. Theory of the Derivative 2.1 Tangent Lines 2.2 Definition of Derivative 2.3 Rates of Change

2. Theory of the Derivative 2.1 Tangent Lines 2.2 Definition of Derivative 2.3 Rates of Change 2.4 Derivative Rules 2.5 Higher Order Derivatives 2.6 Implicit Differentiation 2.7 LHpitals Rule 2.8 Some Classic Theoretical Results

1.08k views • 105 slides
Hack The CPython Batuhan Taskaya @isidentical What is hacking? Why do we hack? Yes, we want

Hack The CPython Batuhan Taskaya @isidentical What is hacking? Why do we hack? Yes, we want

Hack The CPython Batuhan Taskaya @isidentical What is hacking? Why do we hack? Yes, we want FREEDOM! We want to use PEP313! Before we hack, Learn the internals Lexing - Tokenization - Read #define NEWLINE 4 #define INDENT

670 views • 36 slides
u t = u t u t d u = u u + u u d t = u A Vectors Derivative is

u t = u t u t d u = u u + u u d t = u A Vectors Derivative is

Vector Derivative expressed in terms of its smoothly varying (time dependent) magnitude & direction (unit vector) functions. u t = u t u t d u = u u + u u d t = u A Vectors Derivative is determined by its current

662 views • 11 slides
Derivative Function Math 132 Stewart 2.2 In Notes 2.1, we defined the derivative of a

Derivative Function Math 132 Stewart 2.2 In Notes 2.1, we defined the derivative of a

Derivative Function Math 132 Stewart 2.2 In Notes 2.1, we defined the derivative of a function f ( x ) at x = a , namely the number f ( a ). Since this gives an output f ( a ) for any input a , the derivative defines a function.

176 views • 5 slides
Geoffrey Vaughan Lets Hack NFC How does NFC work? How could we hack it? Where

Geoffrey Vaughan Lets Hack NFC How does NFC work? How could we hack it? Where

Geoffrey Vaughan Lets Hack NFC How does NFC work? How could we hack it? Where are the weaknesses? What are the security implications? Security Compass and NFC Currently we are devoting a lot of energy towards NFC

704 views • 35 slides
Securities & Securities & Derivative Derivative Litigation Repor t t Litigation Repor

Securities & Securities & Derivative Derivative Litigation Repor t t Litigation Repor

Securities & Securities & Derivative Derivative Litigation Repor t t Litigation Repor 2004 Second Quarter SEC Appellate Decisions To keep our clients abreast of securities law developments, Carlton Fields' Inside This Issue:

196 views • 8 slides
Securities & Securities & Derivative Derivative Litigation Report Litigation Report

Securities & Securities & Derivative Derivative Litigation Report Litigation Report

Securities & Securities & Derivative Derivative Litigation Report Litigation Report 2004 First Quarter Eleventh Circuit Securities Law Update To keep our clients abreast of securities law developments in the Inside This Issue:

133 views • 12 slides
Securities & Securities & Derivative Derivative Litigation Report Litigation Report

Securities & Securities & Derivative Derivative Litigation Report Litigation Report

Securities & Securities & Derivative Derivative Litigation Report Litigation Report 2004 Second Quarter Eleventh Circuit Securities Law Update To keep our clients abreast of securities law developments in the Southeast, Inside This

378 views • 4 slides
Geometric Interpretation of the Derivative (Review) Geometric Interpretation of the Derivative

Geometric Interpretation of the Derivative (Review) Geometric Interpretation of the Derivative

Geometric Interpretation of the Derivative (Review) Geometric Interpretation of the Derivative (Review) The derivative of a function f ( x ) at point x 0 is the slope of the tangent line at that point. Geometric Interpretation of the Derivative

550 views • 32 slides
ART SLIDES: VAN GOGH EDITION HACK.|100% WORKING!|NEW METHOD|HACK TOOL. Free Buy 30 coins Nearly

ART SLIDES: VAN GOGH EDITION HACK.|100% WORKING!|NEW METHOD|HACK TOOL. Free Buy 30 coins Nearly

ART SLIDES: VAN GOGH EDITION HACK.|100% WORKING!|NEW METHOD|HACK TOOL. Free Buy 30 coins Nearly 1, Paintings and Drawings by Vincent van Gogh Digitized and Put Online | Hacker News For people who don't live close to huge cities with big

269 views • 3 slides
Derivative Applications MAC 2233 Instantaneous Rates of Change of a Function The derivative

Derivative Applications MAC 2233 Instantaneous Rates of Change of a Function The derivative

3/16/2010 Derivative Applications MAC 2233 Instantaneous Rates of Change of a Function The derivative is: The slope of the tangent line at a point The instantaneous rate of change of the function Marginal Analysis The study

561 views • 18 slides
SSA Introduction Sebastian Hack hack@cs.uni-saarland.de Compiler Construction 2013 saarland

SSA Introduction Sebastian Hack hack@cs.uni-saarland.de Compiler Construction 2013 saarland

SSA Introduction Sebastian Hack hack@cs.uni-saarland.de Compiler Construction 2013 saarland university computer science 1 Another kind of CFGs p D ( ) x e x e D ( ) q Effects on edges. Nodes called Nodes are

310 views • 19 slides
FC2015 RSDYK Dyke quality assessment by remote sensing Robert Hack Robert Hack 14-Apr-09 1

FC2015 RSDYK Dyke quality assessment by remote sensing Robert Hack Robert Hack 14-Apr-09 1

FC2015 RSDYK Dyke quality assessment by remote sensing Robert Hack Robert Hack 14-Apr-09 1 FC2015-RSDYK - Hack Pilot project: RSDYK2008 Trial to establish whether remote sensing in combination with geological knowledge can be used for

826 views • 10 slides
Code Generation: Intro Sebastian Hack hack@cs.uni-saarland.de Compiler Construction 2017

Code Generation: Intro Sebastian Hack hack@cs.uni-saarland.de Compiler Construction 2017

Code Generation: Intro Sebastian Hack hack@cs.uni-saarland.de Compiler Construction 2017 Saarland University, Computer Science 1 Code Generation Consists (roughly) of three parts: 1. Instruction Selection Select processor instructions for

282 views • 9 slides
SLIPPERY SLIDES HACK and CHEATS.|100% WORKING!|NEW METHOD|HACK TOOL. Free No Ads 7 Hacks To Make

SLIPPERY SLIDES HACK and CHEATS.|100% WORKING!|NEW METHOD|HACK TOOL. Free No Ads 7 Hacks To Make

SLIPPERY SLIDES HACK and CHEATS.|100% WORKING!|NEW METHOD|HACK TOOL. Free No Ads 7 Hacks To Make Your Boots Slip-Proof May 18, Attach your Water hose. Wet your slide, this just helps for the first few slides. Once the kids get going, their

308 views • 3 slides
Global Value Numbering Sebastian Hack hack@cs.uni-saarland.de 5. Dezember 2017 Saarland

Global Value Numbering Sebastian Hack hack@cs.uni-saarland.de 5. Dezember 2017 Saarland

Global Value Numbering Sebastian Hack hack@cs.uni-saarland.de 5. Dezember 2017 Saarland University, Computer Science 1 Value Numbering a := 2 a := 3 x := a + 1 x := a + 1 y := a + 1 Replace second computation of a + 1 with a copy

394 views • 22 slides
Global Value Numbering Sebastian Hack hack@cs.uni-saarland.de 17. Dezember 2013 saarland

Global Value Numbering Sebastian Hack hack@cs.uni-saarland.de 17. Dezember 2013 saarland

Global Value Numbering Sebastian Hack hack@cs.uni-saarland.de 17. Dezember 2013 saarland university computer science 1 Value Numbering a := 2 a := 3 x := a + 1 x := a + 1 y := a + 1 Replace second computation of a + 1 with a copy from x 2

223 views • 20 slides
Securities Board of India Guest Lecture Convergence of Derivative and Cash Markets Andrew Sheng

Securities Board of India Guest Lecture Convergence of Derivative and Cash Markets Andrew Sheng

Securities Board of India Guest Lecture Convergence of Derivative and Cash Markets Andrew Sheng Chairman Securities and Futures Commission Hong Kong 26 May 2005 Table of Contents 1. Global Trend in Derivative Markets 2. Derivative Markets

462 views • 32 slides
Last time: the second derivative test Consider the function f ( x , y ) = x 2 2 xy + 2 y . Find

Last time: the second derivative test Consider the function f ( x , y ) = x 2 2 xy + 2 y . Find

Last time: the second derivative test Consider the function f ( x , y ) = x 2 2 xy + 2 y . Find all of its critical points, and use the second derivative test to determine whether each is a local maximum, local minimum, or saddle point. (a)

326 views • 6 slides
MAT 166 Calculus for Bus/Soc Chapter 4 Notes Techniques for Finding the Derivative

MAT 166 Calculus for Bus/Soc Chapter 4 Notes Techniques for Finding the Derivative

6/10/2013 MAT 166 Calculus for Bus/Soc Chapter 4 Notes Techniques for Finding the Derivative Calculating the Derivative David J. Gisch The Derivative The Power Rule 1 6/10/2013 Power Rule Power Rule Example: Find when

180 views • 14 slides
Posting Independent Amounts under Derivative Transactions: Industry Recommendations for End User

Posting Independent Amounts under Derivative Transactions: Industry Recommendations for End User

Posting Independent Amounts under Derivative Transactions: Industry Recommendations for End User Protection 16 April 2010 Introduction Credit risk associated with derivative transactions can be managed through a variety of methods such as

362 views • 4 slides
Adjoint Derivative Computation Moritz Diehl and Carlo Savorgnan Adjoint Derivative Computation

Adjoint Derivative Computation Moritz Diehl and Carlo Savorgnan Adjoint Derivative Computation

Adjoint Derivative Computation Moritz Diehl and Carlo Savorgnan Adjoint Derivative Computation Moritz Diehl and Carlo Savorgnan 1/16 There are several methods for calculating derivatives: 1 By hand 2 Symbolic differentiation 3 Numerical

260 views • 24 slides

More recommend