Spark + AI Summit - The Importance of Model Fairness and Interpretability in AI Systems 

Spark + AI Summit - The Importance of Model Fairness and Interpretability in AI Systems 

• 2.
  The Importance ofModel Fairness and Interpretability in AI Systems Spark Summit 2020
• 3.
  Machine Fairness How toassess AI system’s fairness and mitigate any observed unfairness issues www.aka.ms/Fairlearn-build2020
• 4.
  Resources ➢www.aka.ms/InterpretML-github ➢www.aka.ms/InterpretML-toolkit ➢www.aka.ms/InterpretML-doc ➢www.aka.ms/FairlearnAI ➢www.aka.ms/FairlearnAIDoc ➢www.aka.ms/ResponsibleMLDoc ➢www.aka.ms/AzureResponsibleML ➢www.aka.ms/AzureMLDoc ➢www.aka.ms/AzureMLservice
• 5.
  Agenda WHY RESPONSIBLE AI? INTERPRETML FAIRLEARNTOOLKIT
• 6.
  AI will havea considerable impact on business and society as a whole
• 7.
  Report Why Responsible AI?
• 8.
  Microsoft’s AI Principles
• 9.
  Microsoft’s AI Principles
• 10.
  Transparency AI systems shouldbe explainable AI systems should have algorithmic accountability
• 11.
  Machine Learning Interpretabilityuse cases www.aka.ms/InterpretML-github
• 12.
  Interpretability Toolkit www.aka.ms/InterpretML-github
• 13.
  Explain Models -www.aka.ms/InterpretML-toolkit
• 14.
  InterpretML Repo -www.aka.ms/InterpretML-github Interpret Tabular Data Additional interpretability techniques for tabular extension www.aka.ms/InterpretML-github github/InterpretML/Interpret-Community Interpret- Community
• 15.
  Interpret Tabular Data Additional interpretability techniquesfor tabular extension github/InterpretML/Interpret github/InterpretML/Interpret-Community Interpret- Community AzureML- Interpret wrapper Azure Machine Learning SDK InterpretML Repo - www.aka.ms/InterpretML-github
• 16.
  Interpret Tabular Data Additional interpretability techniquesfor tabular extension github/InterpretML/Interpret github/InterpretML/Interpret-Community Interpret- Community Interpret- text-contrib github/InterpretML/Interpret-text-contrib Text Data (Text Classification) AzureML- Interpret wrapper Azure Machine Learning SDK InterpretML Repo - www.aka.ms/InterpretML-github
• 17.
  Types of Models supported www.aka.ms/InterpretML-github
• 18.
  Interpretability Approaches • Givenan existing model/system, extract explanations • Typically provide approximate explanations • Examples: LIME, SHAP, Permutation Feature Importance, Sensitivity Analysis, Influence Functions, etc. Tools to explain the system www.aka.ms/InterpretML-github
• 19.
  SHAP (SHapley Additive exPlanations) • Nota new concept • Concept based on game theory • Mathematically sound • Application in ML relatively new www.aka.ms/InterpretML-github
• 20.
  SHAP Modeling scenario: House pricing prediction www.aka.ms/InterpretML-github
• 21.
  How much haseach feature contributed to the prediction compared to the average prediction? House price prediction: € 300,000 Average house price prediction for all apartments is € 310,000 Delta here is - €10,000 www.aka.ms/InterpretML-github
• 22.
  Parks contributed +10k Cats banned contributed-50k 50𝑚2 contributed +10k 2nd floor had net zero contributions www.aka.ms/InterpretML-github
• 23.
  SHAP How do wecalculate Shapley values? • Take your feature of interest (e.g., cats-banned) and remove it from the feature set • Take the remaining features and generate all possible coalitions • Add and remove your feature of interested to each of the coalitions and calculate the difference it makes www.aka.ms/InterpretML-github
• 24.
  SHAP Pros SHAPCons Based on a solid theory and distributes the effects fairly Computation time: 2k possible coalitions of the feature values for k features Contrastive explanations: Instead of comparing a prediction to the average prediction of the entire dataset, you could compare it to a subset or even to a single data point. Can be misinterpreted Inclusion of unrealistic data instances when features are correlated. www.aka.ms/InterpretML-github
• 25.
  Interpretability Approaches • Modelsinherently interpretable • Typically provide exact explanations • Examples: Linear Models, Decision Rule Lists, Explainable Boosting Machines, Shallow Decision Trees, etc. Tools to explain the system www.aka.ms/InterpretML-github
• 26.
  Linear Models • (Generalized)Linear Models • Current standard for interpretable models • Learns an additive relationship between data and response: • y = β0 + β1x1 + β2x2 + ... + βn xn • βi terms are scalar – single number that multiplies to each feature value • Each feature contributes a “score” that adds up to final output Example Model : House Price = 50,000 + 0.2*sq_ft + 10000*bedrooms + 7000*bathrooms Using this model on a new house: - 5 Bedrooms - 3 Bathrooms - 3000 Sq Feet - House Price = 50000 + 0.2*3000 + 10000*5 + 7000 * 3 - House Price = $121,600 www.aka.ms/InterpretML-github
• 27.
  Explainable Boosting Machine •Generalized Additive Models (with Pairwise Interactions) • GAMs have existed since 1980s • MSR invented new methods of training GAMs for higher accuracy • Think of it as a “more expressive” linear model • Still additive! • Linear Model: y = β0 + β1x1 + β2x2 + ... + βn xn • Additive Model: y = f1(x1) + f2(x2) + ... + fn (xn) • Additive Model with Pairwise Interactions (EBM): y = Ʃi fi (xi) + Ʃij fij (xi , xj ) • Full Complexity Model: y = f (x1, ..., xn) Linear Model can only learn a straight line! www.aka.ms/InterpretML-github
• 29.
  FAIRNESS More positive outcomes& avoiding harmful outcomes of AI systems for groups of people
• 30.
  • Allocation: extendsor withholds opportunities, resources, or information. • Quality of service: whether a system works as well for one person as it does for another • Stereotyping: reinforce existing societal stereotypes • Denigration: actively derogatory or offensive • Over or under representation: over-represent, under-represent, or even erase particular groups of people Crawford et al. 2017 Types of harm
• 31.
  What is Fairlearn? A newapproach to measuring and mitigating unfairness in systems that make predictions, serve users, or make decisions about allocating resources, opportunities, or information. www.aka.ms/FairlearnAI
• 32.
  There are manyways that an AI system can behave unfairly. Fairness in AI Avoiding negative outcomes of AI systems for different groups of people A model for screening loan or job application might be much better at picking good candidates among white men than among other groups. A voice recognition system might fail to work as well for women as it does for men. www.aka.ms/FairlearnAI
• 33.
  A toolkit thatempowers developers of artificial intelligence systems to assess their systems' fairness and mitigate any observed fairness issues. Helps users identify and mitigate unfairness in their machine learning models with a focus on group fairness. Automatically analyze a model’s predictions Provide the user with insights into (un)fairness of their model’s predictions Support (algorithmic) methods to mitigate unfairness www.aka.ms/FairlearnAI
• 34.
  • Allocation: extendsor withholds opportunities, resources, or information. • Quality of service: whether a system works as well for one person as it does for another • Stereotyping: reinforce existing societal stereotypes • Denigration: actively derogatory or offensive • Over or under representation: over-represent, under-represent, or even erase particular groups of people Crawford et al. 2017 Types of harm addressed by Fairlearn www.aka.ms/FairlearnAI
• 35.
  Harm of “allocation” ExampleScenarios: Lending The data set describes whether each individual repaid the loan or not. • [Classification] Recommends whether a given individual should get a loan. The trained model outputs: Yes/Maybe/No decision. Example Scenarios: School Admissions The data set describes what was the first-year GPA of each student. • [Regression] For a given applicant, predicts their GPA at the end of the first year. The trained model outputs a real-valued prediction that is used as a score to screen applicants. www.aka.ms/FairlearnAI
• 36.
  Harm of “qualityof service” Example Scenarios: News Recommendation The training data indicates what article was presented to which user, whether the user clicked, and how much time the user spent on the article. • [Classification] Predict which article to show to each user to optimize click- through rate. The trained model outputs: Yes/No decision. Two kinds of (group) fairness: across users (quality of service), or across publishers/topics of articles (quality of service, but also allocation). www.aka.ms/FairlearnAI
• 37.
  Assessment Evaluation Visualizations Fairness Criteria Mitigation Algorithm Mitigation •Select a protected attribute • Select a fairness metric (e.g., disparity in recall, precision, etc and disparity in selection rate) Fairlearn Toolkit • Demographic Parity • Equalized Odds • Postprocessing Method • Reduction Methodwww.aka.ms/FairlearnAI
• 38.
  Fairness assessment through disparitymetrics Disparity in Performance • How the model accuracy differs across different buckets of a sensitive feature (e.g., how accuracy of the model differs for "female" vs. "male" vs. “unspecified" data points) Disparity in Selection Rate • How the model predictions differ across different buckets of a sensitive feature (e.g., how many "female" individuals have received prediction `approved` on their loan application in contrast to "male" and “unspecified" data points?). www.aka.ms/FairlearnAI
• 39.
  Assessment Evaluation Visualizations Fairness Criteria Mitigation Algorithm Mitigation •Select a protected attribute • Select a fairness metric (e.g., disparity in recall, precision, etc and disparity in selection rate) Fairlearn Toolkit • Demographic Parity • Equalized Odds • Postprocessing Method • Reduction Methodwww.aka.ms/FairlearnAI
• 40.
  Demographic parity: Applicants ofeach race (gender, ...) have the same odds of getting approval on their loan applications Loan approval decision is independent of protected attribute Equalized odds: Qualified applicants have the same odds of getting approval on their loan applications regardless of race (gender, …) Unqualified applicants have the same odds of getting approval on their loan applications regardless of race (gender, …) Fairness Criteria www.aka.ms/FairlearnAI
• 41.
  Reductions approach: Wrapper around standard ML algorithms Input: • anystandard ML training algorithm (as a black box) • data set including sensitive feature Output: • a trained model that minimizes error subject to fairness constraints Advantages: • doesn’t need to access the sensitive feature at test time • works for a wide range of disparity metrics • allows extracting the full disparity-accuracy frontier Disadvantages: • requires re-training: the black box is called 10-20 times www.aka.ms/FairlearnAI
• 42.
  Post processing: Picking a fair thresholdrule Input: • an existing (already trained) scoring model • data set including sensitive feature Output: • the most accurate among all fair threshold rules (a separate threshold for each subpopulation) Advantages: • simplicity • no need to re-train the model Disadvantages: • requires sensitive feature at test-time • doesn’t allow trade-offs between disparity and accuracy www.aka.ms/FairlearnAI
• 43.
  Demo of Interpretability for TabularData in Azure Databricks Hardware Performance Dataset - The Regression goal is to predict the performance of certain combinations of hardware parts. www.aka.ms/InterpretML-github www.aka.ms/DatabricksRuntimeML
• 44.
  Azure ML ServeStore Prepand trainIngest Power BIAzure Data Factory Azure Data Lake Storage Azure Kubernetes Service Model Serving Cosmos DB, SQL DB Operational Databases Azure Synapse Analytics Streaming data Azure Databricks Azure Machine Learning Apps Batch data Ad-hoc analysis Azure Event Hubs Reference architecture
• 45.
  Provisioning Azure DatabricksWORKSPACE
• 46.
  © Microsoft Corporation Fast,easy, and collaborative Apache Spark-based analytics platform Azure Databricks for machine learning modeling Tools InfrastructureFrameworks Leverage powerful GPU-enabled VMs pre-configured for deep neural network training Use HorovodEstimator via a native runtime to enable build deep learning models with a few lines of code Full Python and Scala support for transfer learning on images Automatically store metadata in Azure Database with geo-replication for fault tolerance Use built-in hyperparameter tuning via Spark MLLib to quickly drive model progress Simultaneously collaborate within notebooks environments to streamline model development Load images natively in Spark DataFrames to automatically decode them for manipulation at scale Improve performance 10x-100x over traditional Spark deployments with an optimized environment Seamlessly use TensorFlow, Microsoft Cognitive Toolkit, Caffe2, Keras, and more
• 47.
  © Microsoft Corporation www.aka.ms/DatabricksRuntimeML
• 48.
  Demo dataset Breast CancerWisconsin (Diagnostic) Data Set Features are computed from a digitized image of a fine needle aspirate (FNA) of a breast mass. They describe characteristics of the cell nuclei present in the image. A few of the images can be found at [Web Link] Attribute Information: 1) ID number 2) Diagnosis (M = malignant, B = benign) 3-32) Ten real-valued features are computed for each cell nucleus: a) radius (mean of distances from center to points on the perimeter) b) texture (standard deviation of gray-scale values) c) perimeter d) area e) smoothness (local variation in radius lengths) f) compactness (perimeter^2 / area - 1.0) g) concavity (severity of concave portions of the contour) h) concave points (number of concave portions of the contour) i) symmetry j) fractal dimension ("coastline approximation" - 1)
• 49.
  © Microsoft Corporation Installazureml-interpret and azureml-contrib-interpret packages.
• 50.
  © Microsoft Corporation Traina sample model in a local Jupyter notebook.
• 51.
  © Microsoft Corporation Callthe explainer locally.
• 52.
  © Microsoft Corporation Explainthe entire model behavior (global explanation)
• 53.
  © Microsoft Corporation
• 54.
  © Microsoft Corporation Explainan individual prediction (local explanation)
• 55.
  © Microsoft Corporation
• 56.
  © Microsoft Corporation
• 57.
  © Microsoft Corporation
• 58.
  Mehrnoosh Sameki Senior ProductManager Minsoo Thigpen Product Manager Ehi Nosakhare Machine Learning Scientist
• 59.
  Article: Machine Fairness How toassess AI system’s fairness and mitigate any observed unfairness issues www.aka.ms/Fairlearn-build2020
• 60.
  Resources ➢www.aka.ms/InterpretML-github ➢www.aka.ms/InterpretML-toolkit ➢www.aka.ms/InterpretML-doc ➢www.aka.ms/FairlearnAI ➢www.aka.ms/FairlearnAIDoc ➢www.aka.ms/ResponsibleMLDoc ➢www.aka.ms/AzureResponsibleML ➢www.aka.ms/AzureMLDoc ➢www.aka.ms/AzureMLservice
• 61.
  Thank You!