Tools and applications of machine learning

Teacher
Nicolò Navarin
Department of Mathematics
nicolo.navarin[at]unipd.it
INF/01

Merlyin Monaro
Department of General Psychology
merylin.monaro[at]unipd.it
M-PSI/06

Aim

The course is divided in two parts. The first part provides a practical understanding of basic machine learning tools and their use. Specifically, the student will learn when it is reasonable to use machine learning tools, what are the basic constituents of a machine learning tool, which type of tasks they can achieve, how to evaluate their performance. Examples of applications of the tools will be given. Finally, cognitive services based on machine learning tools will be presented and their application demonstrated. The second part of the course is focused on understanding how machine learning techniques can be applied to answer research questions in psychology. In particular, the goal is i) to provide to students with a computer science/engineering background some basic knowledge from psychology and neuroscience domains where machine learning is now largely applied; ii) to provide to students with a psychological background an overview of the usefulness and potential of applying machine learning to different problems in psychological research, with examples of successful applications. Teachers will present several case studies, covering a broad range of applications from cognitive science (e.g. emotion recognition), cognitive neuroscience (e.g. fMRI analysis), social psychology (e.g. social network analysis, marketing applications), personality psychology (e.g. personality prediction). Students will be encouraged to discuss the limitations of current studies and brainstorm future research developments. Collaboration among students with different backgrounds will be encouraged through research-oriented practical group projects and brainstorming future research developments.

Syllabus
- Introduction to machine learning tools and their correct use.
- What can machine learning tools achieve: supervised vs unsupervised techniques.
- Examples of application and how to assess performances.
- Introduction to cognitive services.
- Case study: “Reading minds” with machine learning and EEG. Case study on brain-computer interfaces.
- Case study: “Reading minds deeper” with fMRI data. Case study on image reconstruction from brain activity.
- Case study: Personality prediction from social data. Case studies on Facebook and Twitter.
- Case study: Predicting sexual orientation from faces. Considerations on reproducibility, bias in data, the burden of proof, and why it is important to explain what machine learning models learn.
- Case study: Emotion recognition from images and multimodal data.
- Case study: Machine Learning for lie detection (e.g., fake identities, fake reviews).
- Case study: Machine Learning for neuromarketing, with guest intervention (Dr. Mirko Polato) on recommender systems.
- Guest intervention: Dr. Andrea Zangrossi “Machine Learning in brain imaging”

Course requirements
The student is expected to have basic knowledge of probability. Basic programming skills are a plus.

Introductory readings
Tom Mitchell, Machine Learning, McGraw-Hill, ISBN:0070428077 9780070428072

Examination modality
The students will be evaluated on practical group projects concerning a machine learning application to a problem of personal research interest. Four hours will be assigned to project discussion, in which students will be encouraged to critical thinking making questions after each presentation.

Course material, enrollment and last minute notifications
Made available by the teacher at this Moodle address

Schedule
25 Jan 2021, 14:00-17:00
27 Jan 2021, 14:00-17:00
29 Jan 2021, 14:00-16:00
01 Feb 2021, 14:00-17:00
03 Feb 2021, 14:00-17:00
05 Feb 2021, 14:00-16:00
examination on Feb 17 14:00-18:00

Location
Zoom; the link is in the Moodle page of the course.

<< Courses in 2020-2021