Course Outline

Spring 2026

CPSY 1950 — Course Overview (Spring 2026)

Core format

  • T/Th, 80 minutes
  • Tue = lecture introducing the week’s theme (conceptual, figure-first)
  • Thu = mini-conference (lightning talks) on the same theme
    • ~15 lightning talks (2:00 talk + 0:30 transition)
    • ~20–30 min synthesis discussion/activity

Group & presentation plan

  • Lightning talks are prepared in small groups.
  • Groups are not fixed: students will form new groups for each lightning week (rotating teams over the semester).

Schedule

Week 1 — Course kickoff

Thu 1/22 — Course Kickoff

NeuroAI goals, course structure, and how we will simulate scientific conferences (lightning talks and posters).

No readings or pre-class activities for Week 1—this is the introductory session.

Week 2 — Bootcamp (async; replaces Tue/Thu lectures)

Tue 1/27 — Bootcamp I (async; completed during normal Tue class time)

Assigned reading (all students):

Deep learning intuition (conditional):

Optional for everyone, but MANDATORY for students who have not taken a deep learning course.

Linear algebra bootcamp (conditional):

Optional for everyone, but MANDATORY for students who have not taken linear algebra and have not taken any ML/AI course that used vectors/matrices seriously.

Neuroscience intro video (conditional):

Optional for everyone, but MANDATORY for students who have not taken an intro course in neuroscience and/or cognitive science and/or cognitive neuroscience.

Textbook-style foundations reading (conditional, skim):

Optional for everyone, but MANDATORY for the same students who are required to watch the neuroscience intro video.

Thu 1/29 — Bootcamp II (async; completed during normal Thu class time)

Assigned reading (all students):

Note: Incorporate world models in reinforcement learning.

Week 3 — The three levers of deep learning

Tue 2/3 — Lecture: The Three Levers of Deep Learning

How architecture, learning objectives, and experience (data/scale) shape representations, behavior, and generalization across modalities.

Note: Connect back to Nancy Kanwisher's examples of color and fruit detection in primates as the objective functions to be optimized / goal of the system/agent.

Note: Incorporate world models in reinforcement learning. See Ha & Schmidhuber (2018), World Models.

Thu 2/5 — Lightning Mini-Conf 1: Three Levers of DL

Details TBD

Week 4 — Scaling and emerging capabilities

Tue 2/10 — Lecture: Scaling and Emerging Capabilities

Pretraining and fine-tuning/transfer; in-context learning and reasoning; what 'emergence' claims mean and how to evaluate them critically.

📑 Slides

Thu 2/12 — Lightning Mini-Conf 2: Scaling & Emergence

Details TBD

Week 5 — Prediction vs Understanding

Tue 2/17 — No lecture (university holiday)

No class.

Thu 2/19 — Background reading and required viewing

Background paper: Serre, T. & Pavlick, E. (2025). From Prediction to Understanding: Will AI Foundation Models Transform Brain Science? Neuron. Brown Library proxy.

Required viewing (watch both):

  • Marcus, G. (2024). Keynote at AGI-24. Machine Learning Street Talk. Watch from ~5:00 to ~35:00.
  • LeCun, Y. (2024). Objective-Driven AI. Ding Shum Lecture, Harvard CMSA. Watch the first ~36 minutes.

Paper response due Thu 2/19, 2:00pm: 📝 Submit on Canvas

Week 6 — Representation-level interpretability

Tue 2/24 — Lecture: Representation-Level Interpretability

Feature visualization, concept-based methods, sparse/dictionary approaches (incl. SAEs); what we can and can't reliably name in representations.

Thu 2/26 — Lightning Mini-Conf 4: Representation Interpretability

Details TBD

Week 7 — Mechanistic interpretability

Tue 3/3 — Lecture: Mechanistic Interpretability

Circuits, causal interventions, and standards of evidence for mechanistic claims.

Thu 3/5 — Lightning Mini-Conf 5: Mechanistic Interpretability

Details TBD

Week 8 — Neural alignment

Tue 3/10 — Lecture: Neural Alignment and Model-to-Brain Mapping

Predicting neural data across measurement modalities; encoding/decoding and representational similarity; what alignment can and cannot justify.

Thu 3/12 — Lightning Mini-Conf 6: Neural Alignment

Details TBD

Week 9 — Behavioral and cognitive alignment

Tue 3/17 — Lecture: Behavioral and Cognitive Alignment

Treating models as participants in cognitive tasks; behavioral signatures beyond accuracy (generalization, planning, decision making, cognitive control); confounds and best practices.

Thu 3/19 — Lightning Mini-Conf 7: Behavioral Alignment

Details TBD

Week 10 — Spring Break

Tue 3/24 — Spring Break

No class

Thu 3/26 — Spring Break

No class

Week 11 — Project studio

Tue 3/31 — Project Studio I

Project launch and evaluation design; in-class time for groups to plan, run pilot tests, and produce first results/figures.

Thu 4/2 — Project Studio II

Continue project work: complete runs and draft poster.

Week 12 — Project poster presentations

Tue 4/7 — Project Poster Mini-Conf A

Students present project findings in posters (17 posters); structured peer feedback and synthesis discussion.

Thu 4/9 — Project Poster Mini-Conf B

Students present project findings in posters (17 posters); structured peer feedback and synthesis discussion.

Week 13 — Guest lectures

Tue 4/14 — Guest lecture (TBD)

Details TBD

Thu 4/16 — Guest lecture: Rufin VanRullen

Frontier topics in NeuroAI: global workspace / consciousness & deep learning.

Week 14 — Guest lectures

Tue 4/21 — Guest lecture (TBD)

Details TBD

Thu 4/23 — Guest lecture: Victor Boutin

Frontier topics in NeuroAI: generative models, EBMs, cognitive science. Plus course wrap-up and final exam briefing.

Final Exam

Tuesday, May 12, 2026, 9:00am