YoVDO

Formal Specification and Testing for Reinforcement Learning

Offered By: ACM SIGPLAN via YouTube

Tags

Reinforcement Learning Courses Software Testing Courses Functional Programming Courses Mutation Testing Courses Property-based Testing Courses Q-learning Courses Denotational Semantics Courses

Course Description

Overview

Save Big on Coursera Plus. 7,000+ courses at $160 off. Limited Time Only!
Explore a 29-minute video presentation from ICFP 2023 that introduces a formal specification and testing framework for reinforcement learning applications. Delve into the systematic approach proposed by researchers from IT University of Copenhagen and University of Oslo to address the exploratory nature of reinforcement learning development. Learn about their formal specification of reinforcement learning problems and algorithms, focusing on temporal difference methods and backup diagrams. Discover the test harness developed for temporal difference learning applications, including SARSA and Q-learning, rooted in functional programming methods. Examine the use of compositional interpreters and property-based testing to create a robust testing methodology. Gain insights into the effectiveness of their test suite through mutation testing results, demonstrating high mutant kill rates and the potential for write-once-use-everywhere tests applicable to various reinforcement learning problems.

Syllabus

[ICFP'23] Formal Specification and Testing for Reinforcement Learning


Taught by

ACM SIGPLAN

Related Courses

Functional Programming Principles in Scala
École Polytechnique Fédérale de Lausanne via Coursera
Functional Program Design in Scala
École Polytechnique Fédérale de Lausanne via Coursera
Paradigms of Computer Programming
Université catholique de Louvain via edX
Introduction to Functional Programming
Delft University of Technology via edX
Paradigms of Computer Programming – Fundamentals
Université catholique de Louvain via edX