Book Club Presentations

Groups Presenting:

• The Creative Destruction of Medicine: How the Digital Revolution Will Create Better Health
• Data and Goliath

Book Club Presentations

Groups Presenting:

• Future Crimes
• Antisocial Media: How Facebook Disconnects Us and Undermines Democracy

Book Club Presentations

Groups Presenting:

• Hacking the Electorate
• Automating Inequality

Book Club Presentations

Groups Presenting:

• Weapons of Math Destruction
• Technically Wrong: Sexist Apps, Biased Algorithms, and Other Threats of Toxic Tech

Required Readings:

• Methodological Gaps in Predicting Mental Health States from Social Media
• “Participant” Perceptions of Twitter Research Ethics

Required Readings:

• Data Statements for Natural Language Processing
• Model Cards for Model Reporting

Differential Privacy

Optional Readings:

• Apple's Differential Privacy Overview
• Privacy Loss in Apple’s Implementation of Differential Privacy on MacOS 10.12
• Can a set of equations keep U.S. census data private?
• To Reduce Privacy Risks, the Census Plans to Report Less Accurate Data
• Implications of Differential Privacy for Census Bureau Data Dissemination

Differential Privacy

Required Readings:

• None

Differential Privacy

Required Readings:

• None

Differential Privacy

Required Readings:

• None

Differential Privacy

Required Readings:

• None

Word Embeddings

Required Readings:

• Screening for Pancreatic Adenocarcinoma Using Signals From Web Search Logs
• Privacy Implications of Health Information Seeking on the Web

* There will be no quiz due Tuesday March 19.

Racial and gender variation in system performance

Required Readings:

• A Framework for Understanding Unintended Consequences of Machine Learning
• Gender Shades: Intersectional Accuracy Disparities in Commercial Gender Classification

Google autocomplete and image search

Required Readings:

• How Google autocomplete works in Search
• Google Search Help
• Google autocomplete still makes vile suggestions

Bias encoded in word representation

Required Readings:

• From the Google intro to machine learning (Section: Embeddings)
• Better, less-stereotyped word vectors

Deriving word representations from corpora

Required Readings:

• Man is to Computer Programmer as Woman is to Homemaker? Debiasing Word Embeddings, Tolga Bolukbasi, Kai-Wei Chang, James Zou, Venkatesh Saligrama, and Adam Kalai
• Semantics derived automatically from language corpora contain human-like biases, A Caliskan, J Bryson, A Narayanan

* Please note that both of these readings have already been assigned for Lecture 6, so there will be no quiz on them assigned.

Fairness in Machine Learning V

Required Readings:

• None

Fairness in Machine Learning IV

Required Readings:

• Attacking discrimination with smarter machine learning
• How We Analyzed the COMPAS Recidivism Algorithm

Fairness in Machine Learning III

Required Readings:

• How We Analyzed the COMPAS Recidivism Algorithm

Fairness in Machine Learning II

Required Readings:

• The Frontiers of Fairness in Machine Learning, Alexandra Chouldechova, and Aaron Roth
• Inherent Trade-Offs in the Fair Determination of Risk Scores (First 8 pages only), Jon Kleinberg, Sendhil Mullainathan, and Manish Raghavan

Fairness in Machine Learning I

Required Readings:

• 21 Definitions of Fairness and Their Politics, Arvind Narayanan

Required Readings:

• Man is to Computer Programmer as Woman is to Homemaker? Debiasing Word Embeddings, Tolga Bolukbasi, Kai-Wei Chang, James Zou, Venkatesh Saligrama, and Adam Kalai
• Semantics derived automatically from language corpora contain human-like biases, A Caliskan, J Bryson, A Narayanan

Machine Learning Basics IV

Required Readings:

• From the Google intro to machine learning, sections:
  • Regularization
  • Simplicity
  • Logistic regression
  • Classification
• Deep neural networks are more accurate than humans at detecting sexual orientation from facial images, Yilun Wang, Michal Kosinski

Machine Learning Basics III

Required Readings:

• From the Google intro to machine learning, sections:
  • Training and test sets
  • Validation set
  • Representation
  • Feature crosses
• Vox article
• Response by Google researcher

Machine Learning Basics II

Required Readings:

• From the Google intro to machine learning, sections:
  • Framing (video lecture + ML terminology)
  • Descending into ML (video lecture + linear regression + training and loss)
  • Reducing loss (video lecture + all readings)
  • Generalization

Machine Learning Basics I

Required Readings:

• None

Introduction, class motivation and overview, machine learning basics

Required Readings:

• None