Philosophers, computer scientists, and experimental scientists all share a concern for determining what patterns — what natural laws — are learnable from finite data, and with identifying which methods of learning are reliable. Because these disciplines speak different languages, this overlapping interest is often missed, and results from one field that address the problems of another find one another with too little frequency. This workshop is intended to address this problem by teaching philosophers how to give their theories of inductive inference purchase on the empirical world, by showing practitioners of machine learning how their concerns fit into a broader scheme of scientific inference, and by helping researchers in the experimental sciences recognize how the abstract results in formal epistemology and machine learning can provide concrete practical benefits in the lab.
Topics to be covered include:
– the problem of induction
– the logic of discovery
– statistical learning theory
– algorithmic learning theory
– causal discovery algorithms
– the problem of natural kinds
– variable choice and model parameterization
– the history of automated discovery algorithms
– experimental design
– basic elements of programming in Python
– hardware interfacing with the Raspberry Pi, SPI, and I2C
– basic elements of student cognition and effective teaching
Participants in the workshop will…
(1) …conduct a series of experimental and programming tasks designed to provide insight into the nature and implementation of both historically important and cutting-edge algorithms for automated scientific discovery. In these exercises, data will be drawn exclusively and directly from measurements of the real world using a Raspberry Pi to host a collection of sensors and actuators.
(2) …engage in philosophical discussions designed to illuminate the shared objectives and complementary methods of philosophy, machine learning, and experimental science.
(3) …share what they’ve learned by assisting in a “Robot Scientist” outreach event for secondary-school students at the Western Virginia Museum of Science. In the final two days of the program, participants will travel to Roanoke to help young students learn about the nature of scientific investigation and the use of computers to probe the physical world.
Aside from the organizer, Dr. Jantzen (Philosophy, Virginia Tech), this year’s workshop will feature presentations from:
Nicolas Fillion, Simon Fraser University
Konstantin Genin, University of Toronto
Subhradeep Roy, Virginia Tech
Alex Tolbert, Virginia Tech
Tentative reading list
Participants will make use of readings from the following (partial and tentative) list of sources:
The 2019 schedule can be found here: