Iliad Intensive Curriculum
The Iliad Intensive is a month-long, full-time AI alignment course for students with strong mathematics, physics, or theoretical-CS backgrounds. The materials are self-contained lecture notes and worksheets on various topics, and pointers for further study. About 20 contributors developed them. We welcome feedback via issues on GitHub.
A — Alignment
B — Learning
- B.3Singular Learning Theory
Singular learning theory (SLT) places degeneracy as a core part of understanding how neural networks learn. We cover the parameter-function map, the meaning of degeneracy through the local learning coefficient, to Watanabe's free energy formula and Bayesian phase transitions.
- B.3Singular Learning Theory
- B.4Training Dynamics
Exact learning dynamics of deep linear networks -- loss-landscape geometry, balanced gradient flow and the NTK, the rich (saddle-to-saddle) and lazy regimes, their mixed unification, and the implicit bias of SGD noise.
- B.4Training Dynamics
C — Abstractions, Representations, and Interpretability
D — Agency
D.1Decision Theory and Reinforcement Learning
- D.1.1Preferences to Rewards
Building from preferences and a minimal set of axioms to a utility function expressible as a sum of discounted rewards: the familiar framing in reinforcement learning.
- D.1.2Reinforcement Learning
The Bellman equations and what follows from them: the existence of optimal policies, the policy improvement theorem, the rate of convergence of Bellman updates, and the convergence of Q-learning.
- D.1.1Preferences to Rewards
- D.2Policy Gradients and Misgeneralization
Recorded lecture for the day. The coding-day worksheet is not yet ported.
- D.2Policy Gradients and Misgeneralization
D.3AIXI
- D.3.1Solomonoff Induction
How an idealized agent should predict. A Bayesian mixture over a countable class of computable hypotheses learns to predict any sequence, with total error bounded by the description length of the truth -- a formal Occam's razor.
- D.3.2AIXI
Exploring the Bayesian optimal policy for history based reinforcement learning.
- D.3.1Solomonoff Induction
- D.6Instrumental Convergence
We discuss a simple mathematical formalization of what it means to "seek power" in a Markov decision process (MDP), and conditions under which such behavior emerges.
- D.6Instrumental Convergence