Admission requirements
Admission to this course is restricted to students enrolled in MA Philosophy 120 EC, specialisation Philosophy of Natural Sciences.
Description
Semi-supervised or unsupervised machine learning algorithms, or “deep learning”, have the capacity to revolutionise scientific practice as much as they have implications for communication, health care, commerce, finance, and other domains of modern life. In this course, we consider the use of deep learning in scientific discovery and address two general questions: how should our image of scientific practice change in response, and to what extent are established concepts and theories in philosophy of knowledge adequate to analyse these changes? The course involves studying recent applications of deep learning in scientific discovery, for example in protein folding prediction competitions; assessing to what extent the outcomes satisfy standards of justifiable or reliable knowledge in epistemology and philosophy of science; and clarifying the relation between the virtues of empiricism, surveyability, and understanding in this new form of science.
Course objectives
Students who successfully complete the course will be able to:
Present and discuss the intellectual principles underlying deep learning and explain the differences between deep learning and earlier algorithms for data handing;
Discuss connections and similarities between uses of deep learning for pattern recognition and decision making;
Give a critical account of recent applications of deep learning in scientific practice;
Apply concepts and theories from philosophy of knowledge to such cases, and use such cases to evaluate the adequacy of those philosophical concepts and theories;
Assess the implications of these developments for the changing nature of science;
Give a class presentation and write a paper on the above topics;
Design and present a PhD-level research proposal on the topics of the course.
Timetable
The timetable is available on the following website:
Mode of instruction
- Seminars, or tutorials if the enrolment is small.
Class attendance is required.
Course load
Total course load (10 EC x 28 hours): 280 hours
Class attendance: 13 × 3 hours = 39 hours
Reworking of class notes: 13 hours
Literature study (approx. 500 pages): 74 hours
Preparation for class presentation: 24 hours
Practical assignment: 10 hours
Writing of research proposal/final essay: 120 hours
Assessment method
Assessment
Two compulsory presentations during the semester;
Two shorter papers;
Term paper;
Oral class participation.
Weighting
The final mark for the course is established by determination of the weighted average of several subtests.
Resit
One resit will be offered, covering the entire course content and consisting of a paper. The grade will replace previously earned grades for subtests. Class participation and practical assignments (presentations) are mandatory requirements for taking the tests and resit. Students who have obtained a satisfactory grade for the first examination(s) cannot take the resit.
Inspection and feedback
How and when an exam review will take place will be disclosed together with the publication of the exam results at the latest. If a student requests a review within 30 days after publication of the exam results, an exam review will have to be organized.
Blackboard
Blackboard will be used for:
- Posting course material.
Reading list
Research articles and other recent scholarly literature. A list will be circulated to students later. All material available online: no purchase required.
Registration
Enrolment through uSis is mandatory.
General information about uSis is available on the website.
Students are strongly advised to register in uSis through the activity number, which can be found in the timetables for courses and exams.
Registration Studeren à la carte and Contractonderwijs
Not applicable.
Contact information
Remarks
None.