Runtime Verifjcation of Scientifjc Software Maxwell Shinn, Clarence - PowerPoint PPT Presentation

Runtime Verifjcation of Scientifjc Software Maxwell Shinn, Clarence Lehman, and Ruzica Piskac October 31, 2018

Motivation def graph_measure(filename): timeseries = load_from_csv(filename) corr_matrix = corr_coef(timeseries) normalized = fisher_transform(corr_matrix) G = matrix_to_graph(normalized) return graph_clustering(G) 2 of 6 Maxwell Shinn, Clarence Lehman, and Ruzica Piskac (2018)

Expected result 3 of 6 Maxwell Shinn, Clarence Lehman, and Ruzica Piskac (2018)

Actual result 4 of 6 Maxwell Shinn, Clarence Lehman, and Ruzica Piskac (2018)

Motivation def graph_measure(filename): timeseries = load_from_csv(filename) corr_matrix = corr_coef(timeseries) normalized = fisher_transform(corr_matrix) G = matrix_to_graph(normalized) return graph_clustering(G) 5 of 6 Maxwell Shinn, Clarence Lehman, and Ruzica Piskac (2018)

Motivation def graph_measure(filename): timeseries = load_from_csv(filename) normalized = fisher_transform(corr_matrix) G = matrix_to_graph(normalized) return graph_clustering(G) 5 of 6 Maxwell Shinn, Clarence Lehman, and Ruzica Piskac (2018) corr_matrix = corr_coef(timeseries) diag = 1+1e-10

Motivation def graph_measure(filename): timeseries = load_from_csv(filename) normalized = fisher_transform(corr_matrix) = NaN G = matrix_to_graph(normalized) return graph_clustering(G) 5 of 6 Maxwell Shinn, Clarence Lehman, and Ruzica Piskac (2018) corr_matrix = corr_coef(timeseries) diag = 1+1e-10

Motivation def graph_measure(filename): timeseries = load_from_csv(filename) normalized = fisher_transform(corr_matrix) = NaN return graph_clustering(G) 5 of 6 Maxwell Shinn, Clarence Lehman, and Ruzica Piskac (2018) corr_matrix = corr_coef(timeseries) diag = 1+1e-10 G = matrix_to_graph(normalized) NaN → 0

Motivation def graph_measure(filename): timeseries = load_from_csv(filename) normalized = fisher_transform(corr_matrix) = NaN return graph_clustering(G) ??? 5 of 6 Maxwell Shinn, Clarence Lehman, and Ruzica Piskac (2018) corr_matrix = corr_coef(timeseries) diag = 1+1e-10 G = matrix_to_graph(normalized) NaN → 0

Our tool In scientifjc software, we don’t know what the output should be Python library for scientifjc software Checks entry and exit conditions Conditions specifjed with refjnement types Specifjcation serves as documentation Automated offmine testing 6 of 6 Maxwell Shinn, Clarence Lehman, and Ruzica Piskac (2018)

Our tool Python library for scientifjc software Checks entry and exit conditions Conditions specifjed with refjnement types Specifjcation serves as documentation Automated offmine testing 6 of 6 Maxwell Shinn, Clarence Lehman, and Ruzica Piskac (2018) ▶ In scientifjc software, we don’t know what the output should be

Our tool Checks entry and exit conditions Conditions specifjed with refjnement types Specifjcation serves as documentation Automated offmine testing 6 of 6 Maxwell Shinn, Clarence Lehman, and Ruzica Piskac (2018) ▶ In scientifjc software, we don’t know what the output should be ▶ Python library for scientifjc software

Our tool Conditions specifjed with refjnement types Specifjcation serves as documentation Automated offmine testing 6 of 6 Maxwell Shinn, Clarence Lehman, and Ruzica Piskac (2018) ▶ In scientifjc software, we don’t know what the output should be ▶ Python library for scientifjc software ▶ Checks entry and exit conditions

Our tool Specifjcation serves as documentation Automated offmine testing 6 of 6 Maxwell Shinn, Clarence Lehman, and Ruzica Piskac (2018) ▶ In scientifjc software, we don’t know what the output should be ▶ Python library for scientifjc software ▶ Checks entry and exit conditions ▶ Conditions specifjed with refjnement types

Automated offmine testing Our tool 6 of 6 Maxwell Shinn, Clarence Lehman, and Ruzica Piskac (2018) ▶ In scientifjc software, we don’t know what the output should be ▶ Python library for scientifjc software ▶ Checks entry and exit conditions ▶ Conditions specifjed with refjnement types ▶ Specifjcation serves as documentation

Our tool 6 of 6 Maxwell Shinn, Clarence Lehman, and Ruzica Piskac (2018) ▶ In scientifjc software, we don’t know what the output should be ▶ Python library for scientifjc software ▶ Checks entry and exit conditions ▶ Conditions specifjed with refjnement types ▶ Specifjcation serves as documentation ▶ Automated offmine testing