mashedpoteto / ccm-site Goto Github PK

Instructors: Brenden Lake and Todd Gureckis

Teaching Assistants: Yanli Zhou and Graham Flick

Meeting time and location:
Lecture
Mondays 1:35-3:15 PM
Silver Center for Arts & Science, 100 Washington Sq East, Room 520

Lab
Tuesdays 2:40-3:30 PM
Silver Center for Arts & Science, 100 Washington Sq East, Room 520

Course numbers:
DS-GA 1016 (Data Science)
PSYCH-GA 3405.002 (Psychology)

Contact information and Piazza:
We use Piazza for questions and class discussion. Piazza gets you help efficiently from classmates, the TA, and the instructors. Rather than emailing questions to the teaching staff, please post your questions on Piazza.

The signup link for our Piazza page is available here (piazza.com/nyu/spring2020/dsga3001005).

Once signed up, our class Piazza page is available here (piazza.com/nyu/spring2020/dsga3001005/home).

If you have a question that isn't suitable for Piazza and there is a need to email the teaching staff directly, please use the following email address: [email protected]

Office hours:
Todd Gureckis (Tuesday at 1-2pm; 6 Washington Place, Meyer, Room 859)
Brenden Lake (Tuesdays at 10-11am; 60 5th Ave., Room 610)
Yanli Zhou (Wednesdays at 1-2pm; 60 5th Ave., Room 609)
Graham Flick (Wednesdays at 3-4pm; 10 Washington Place, 6th Floor)

Summary: This course surveys the leading computational frameworks for understanding human intelligence and cognition. Both psychologists and data scientists are working with increasingly large quantities of human behavioral data. Computational cognitive modeling aims to understand behavioral data and the mind and brain, more generally, by building computational models of the cognitive processes that produce the data. This course introduces the goals, philosophy, and technical concepts behind computational cognitive modeling.

The lectures cover artificial neural networks (deep learning), reinforcement learning, Bayesian modeling, model comparison and fitting, classification, probabilistic graphical models, and program induction. Modeling examples span a broad set of psychological abilities including learning, categorization, language, memory, decision making, and reasoning. The homework assignments include examining and implementing the models surveyed in class. Students will leave the course with a richer understanding of how computational modeling advances cognitive science, how cognitive science can inform research in machine learning and AI, and how to fit and evaluate cognitive models to understand behavioral data.

Please note that this syllabus is not final and there may be further adjustments.

• Math: We will use concepts from linear algebra, calculus, and probability. If you had linear algebra and calculus as an undergrad, or if you have taken Math Tools in the psychology department, you will be in a good position for approaching the material. Familiarity with probability is also assumed. We will review some of the basic technical concepts in lab.

• Programming: Previous experience with Python is required. The assignments will use Python 3 and Jupyter Notebooks (http://jupyter.org). We will have a Python refresher in lab, and we also recommend this tutorial (http://openbookproject.net/thinkcs/python/english3e/).

The final grade is based on the homeworks (60%) and the final project (40%).

Class participation will be used to decide grades in borderline cases. As we only meet once a week, class attendance is obviously important for success in the class.

The final project proposal is due Wednesday, April 1 (one half page written). Please submit via email to [email protected] with the file name lastname1-lastname2-lastname3-ccm-proposal.pdf.

The final project is due Wednesday 5/13. Please submit via email to [email protected] with the file name lastname1-lastname2-lastname3-ccm-final.pdf.

The final project will be done in groups of 3-4 students. A short paper will be turned in describing the project (approximately 6 pages). The project will represent either an substantial extension of one of the homeworks (e.g., exploring some new aspect of one of the assignments), implementing and extending an existing cognitive modeling paper, or a cognitive modeling project related to your research. We provide a list of project ideas here, but of course you do not have to choose from this list.

Write-ups should be organized and written as a scientific paper. It must include the following sections: Introduction (with review of related work), Methods/Models, Results, and Discussion/Conclusion. A good example would be to follow the structure of this paper from the class readings:

• Peterson, J., Abbott, J., & Griffiths, T. (2016). Adapting Deep Network Features to Capture Psychological Representations. Presented at the 38th Annual Conference of the Cognitive Science Society. link here

Code submission is not required for the final project.

Mondays 1:35-3:15 PM
Silver Center for Arts & Science, 100 Washington Sq East, Room 520

• 1/27 : Introduction (slides)
• 2/3 : Neural networks / Deep learning (part 1)
  • Homework 1 assigned (Due 2/24) (instructions for accessing here)
• 2/10 : Neural networks / Deep learning (part 2)
• 2/17 : NO CLASS - President's day
• 2/24 : Reinforcement learning (part 1)
  • Homework 2 assigned (Due 3/23) (instructions for accessing here)
• 3/2 : Reinforcement learning (part 2)
• 3/9 : Reinforcement learning (part 3)
• 3/16 : NO CLASS - Spring recess
• 3/23 : Bayesian modeling (part 1)
  • Homework 3 assigned (Due 4/13) (instructions for accessing here)
• 3/30 : Bayesian modeling (part 2)
• Project proposal due (Wednesday April 1)
• 4/6 : Rational vs. mechanistic modeling
• 4/13 : Model comparison and fitting, tricks of the trade
• 4/20 : Categorization
  • Homework 4 assigned (Due 5/4) (instructions for accessing here)
• 4/27 : Probabilistic Graphical models
• 5/4 : Program induction and language of thought models
• 5/11 : Computational Cognitive Neuroscience
• Final project due (Wednesday 5/13)

Tuesdays 2:40-3:30 PM
Silver Center for Arts & Science, 100 Washington Sq East, Room 520

• 1/28 : Python and Jupyter notebooks review
• 2/4 : Introduction to PyTorch
• 2/11 : No lab
• 2/18 : HW 1 Review
• 2/25 : TBD
• 3/3 : TBD
• 3/10 : HW 2 Review
• 3/17 : SPRING RECESS
• 3/24 : Probability review
• 3/31 : TBD
• 4/7 : HW 3 Review
• 4/14 : TBD
• 4/21 : TBD
• 4/28 : HW 4 Review
• 5/5 : TBD
• 5/12 : No lab (classes end 5/11)

Papers are available for download on NYU Classes in the "Resources" folder.

Neural networks and deep learning

• McClelland, J. L., Rumelhart, D. E., & Hinton, G. E. The Appeal of Parallel Distributed Processing. Vol I, Ch 1.
• LeCun, Y., Bengio, Y. & Hinton, G. (2015). Deep learning. Nature 521:436–44.
• McClelland, J. L., & Rogers, T. T. (2003). The parallel distributed processing approach to semantic cognition. Nature Reviews Neuroscience, 4(4), 310-322.
• Elman, J. L. (1990). Finding structure in time. Cognitive Science, 14(2), 179-211.
• Peterson, J., Abbott, J., & Griffiths, T. (2016). Adapting Deep Network Features to Capture Psychological Representations. Presented at the 38th Annual Conference of the Cognitive Science Society.

Reinforcement learning and decision making

• Gureckis, T.M. and Love, B.C. (2015) Reinforcement learning: A computational perspective. Oxford Handbook of Computational and Mathematical Psychology, Edited by Busemeyer, J.R., Townsend, J., Zheng, W., and Eidels, A., Oxford University Press, New York, NY.
• Daw, N.S. (2013) "Advanced Reinforcement Learning" Chapter in Neuroeconomics: Decision making and the brain, 2nd edition
• Niv, Y. and Schoenbaum, G. (2008) “Dialogues on prediction errors” Trends in Cognitive Science, 12(7), 265-72.
• Nathaniel D. Daw, John P. O'Doherty, Peter Dayan, Ben Seymour & Raymond J. Dolan (2006). Cortical substrates for exploratory decisions in humans. Nature, 441, 876-879.

Bayesian modeling

• Russel, S. J., and Norvig, P. Artificial Intelligence: A Modern Approach. Chapter 13, Uncertainty.
• Tenenbaum, J. B., and Griffiths, T. L. (2001). Generalization, similarity, and Bayesian inference. Behavioral and Brain Sciences, 24(4), 629-640.
• Tenenbaum, J. B., Kemp, C., Griffiths, T. L., & Goodman, N. D. (2011). How to grow a mind: Statistics, structure, and abstraction. Science, 331(6022), 1279-1285.
• Ghahramani, Z. (2015). Probabilistic machine learning and artificial intelligence. Nature, 521(7553), 452.
• MacKay, D. (2003). Chapter 29: Monte Carlo Methods. In Information Theory, Inference, and Learning Algorithms.

Rational versus mechanistic modeling approaches

• Jones, M. & Love, B.C. (2011). Bayesian Fundamentalism or Enlightenment? On the Explanatory Status and Theoretical Contributions of Bayesian Models of Cognition. Behavioral and Brain Sciences (target article).
• Griffiths, T.L., Lieder, F., & Goodman, N.D. (2015). Rational use of cognitive resources: Levels of analysis between the computational and the algorithmic. Topics in Cognitive Science, 7(2), 217-229.

Model comparison and fitting, tricks of trade

• Wilson, R.C. and Collins, A.G.E. (2019). Ten simple rules for the computational modeling of behavioral data. eLife 2019;8:e49547
• Pitt, M.A. and Myung, J (2002) When a good fit can be bad. Trends in Cognitive Science, 6, 10, 421-425.
• Roberts, S. & Pashler, H. (2000) How persuasive is a good fit? A comment on theory testing. Psychological Review, 107, 358-367.
• [optional] Myung, I.J. (2003). Tutorial on maximum likelihood estimation. Journal of Mathematical Psychology, 47, 90-100.

Probabilistic graphical models

• Charniak (1991). Bayesian networks without tears. AI Magazine, 50-63.
• Kemp, C., & Tenenbaum, J. B. (2008). The discovery of structural form. Proceedings of the National Academy of Sciences, 105(31), 10687-10692.
• [optional] Russel, S. J., and Norvig, P. Artificial Intelligence: A Modern Approach. Chapter 14, Probabilistic reasoning systems.

Program induction and language of thought models

• Ghahramani, Z. (2015). Probabilistic machine learning and artificial intelligence. Nature, 521(7553), 452.
• Goodman, N. D., Tenenbaum, J. B., & Gerstenberg, T. (2014). Concepts in a probabilistic language of thought. Center for Brains, Minds and Machines (CBMM).
• Lake, B. M., Salakhutdinov, R., & Tenenbaum, J. B. (2015). Human-level concept learning through probabilistic program induction. Science, 350(6266), 1332-1338.

Computational Cognitive Neuroscience

• Kreigeskorte, N. and Douglas, P.K. (2018) Cognitive computational neuroscience. Nature Neuroscience. 21(9): 1148-1160. doi:10.1038/s41593-018-0210-5
• Turner, B.M., Forstmann, B.U., Love, B.C., Palmeri, T.J., Van Maanen, L. (2017). Approaches to analysis in model-based cognitive neuroscience. Journal of Mathematical Psychology. 76(B), 65-79.

Auditing:
We don't have room for any additional auditors at this point, if you haven't already emailed us. If you are auditing and you would like access to NYU classes and the readings, please add your email address to this speadsheet

Collaboration and honor code:
We take the collaboration policy and academic integrity very seriously. Violations of the policy will result in zero points and possible disciplinary referral.

You may discuss the homework assignments with your classmates, but you must run the simulations and complete the write-ups for the homeworks on your own. Under no circumstance should students look at each other’s code or write ups, or code/write-ups from previous years of this course. Do not share your write up or code with any of your classmates under any circumstances.

Late work:
We will take off 10% for each day a homework or final project is late. Assignments should be turned in all-at-once and not in pieces. If an assignment is incomplete and later completed, the late penalty is applied to the entire assignment.

Extra credit:
No extra credit will be given, out of interest of fairness.

Laptops in class:
Laptops in class are strongly discouraged. We know many try to take notes on their laptops, but it’s easy to get distracted (social media, etc.). It also distracts everyone behind you if you are not engaged! We encourage you to engage with the class and material, and engage with us as the instructors. Ask questions! All slides are posted so there is no need to copy everything down, and paper notes are great too.

Students registered for the course have the option of completing homework assignments on their personal computers, or in a cloud Jupyter environment with all required packages pre-installed. Students can log onto the environment using their nyu net ids here.