Reliable and Interpretable Artificial Intelligence
Creating reliable and explainable probabilistic models is a major challenge to solving the artificial intelligence problem. This course covers some of the latest advances that bring us closer to constructing such models. These advances span the areas of program synthesis/induction, programming languages, machine learning, and probabilistic programming. The material consists of three interconnected parts:
Part I: Program Synthesis/Induction
Synthesis is a new frontier in AI where the computer programs itself from user provided examples. Synthesis has significant applications for nonprogrammers as well as for programmers where it can provide massive productivity increase (e.g., wrangling for data scientists). Modern synthesis techniques excel at learning functions over discrete spaces from (partial) intent. There have been a number of recent, exciting breakthroughs in techniques that discover complex, interpretable/explainable functions from few examples, partial sketches and other forms of supervision.
Topics covered:
 Theory of program synthesis: version spaces, counterexample guided inductive synthesis (CEGIS) with SAT/SMT, synthesis from noisy examples, learning with few examples, compositional synthesis, lower bounds on learning.
 Applications of techniques: synthesis for end users (e.g., spreadsheets), data analytics and financial computing, interpretable machine learning models for structured data.
 Combining neural networks and synthesis
Part II: Robustness of Deep Learning
Deep learning methods based on neural networks have made impressive advances in recent years. A fundamental challenge with these models is that of understanding what the trained neural network has actually learned, for example, how stable / robust the network is to slight variations of the input (e.g., an image or a video), how easy it is to fool the network into misclassifying obvious inputs, etc.
Topics covered:
 Basics of neural networks: fully connected, convolutional networks, residual networks, activation functions
 Finding adversarial examples in deep learning with SMT
 Methods and tools to guarantee robustness of deep nets (e.g., via affine arithmetic, SMT solvers); synthesis of robustness specs
Part III: Probabilistic Programming
Probabilistic programming is an emerging direction whose goal is democratize the construction of probabilistic models. In probabilistic programming, the user specifies a model while inference is left to the underlying solver. The idea is that the higher level of abstraction makes it easier to express, understand and reason about probabilistic models.
Topics covered:
 Inference: MCMC samplers and tactics (approximate), symbolic inference (exact).
 Semantics: basic measure theoretic semantics of probability; bridging measure theory and symbolic inference.
 Frameworks and languages: WebPPL (MIT/Stanford), PSI (ETH), Picture/Venture (MIT), Anglican (Oxford).
 Synthesis for probabilistic programs: this connects to Part I
 Applications of probabilistic programming: using the above solvers for reasoning about bias in machine learning models (connects to Part II), reasoning about computer networks, security protocols, approximate computing, cognitive models, rational agents.
Lectures (Tentative Schedule):
No.  Date  Content  Slides  Exercises  Solutions 

1  Sept 19  Motivation, topics, organization  
Part I: Program Synthesis/Induction  
2  Sept 26  Sketching, Counterexample inductive synthesis (CEGIS)  , blackboard:  
3  Oct 3  Exact Learning  
4  Oct 10  ML for Synthesis, Synthesis for Interpretable ML 
,
RNNs on blackboard: 

5  Oct 17  Neural Turing Machines, Synthesis for Interpretability 
RNNs summary: 

Part I Project 


Part II: Robustness of Deep Learning  
6  Oct 24  Deep Learning  
7  Oct 31  Adversarial Examples in Learning Models  
8  Nov 7  Basics of Abstract Interpretation  
9  Nov 14  Robustness Analysis of Neural Networks (Continued)  
10  Nov 21  Deep Learning for Program Synthesis (Guest lecture)  
Part III: Probabilistic Programming  
11  Nov 28  Introduction, Analysis with Discrete Distributions  
12  Dec 5  Synthesis for Probabilistic Programs: Security Application  exercise12.txt  solutions12.txt  
13  Dec 12  PSI: Exact Inference for Probabilistic Programs  exercise13.txt  solution13.txt  
14  Dec 19  Exam Topics, Potential Thesis Topics 
Readings (Optional):
Lecture 2
 Combinatorial Sketching for Finite Programs, ASPLOS 2006
 OracleGuided ComponentBased Program Synthesis, ICSE 2010
 Automating String Processing in Spreadsheets using InputOutput Examples, POPL 2011
 Learning Programs from Noisy Data, POPL 2016
Lecture 3
Lecture 4
Lecture 5
Lecture 6
Lecture 7
 Explaining and Harnessing Adversarial Examples
 Towards Evaluating the Robustness of Neural Networks
 Robust PhysicalWorld Attack on Deep Learning Models
 DeepXplore: Automated Whitebox Testing of Deep Learning Systems