Update two lectures on affine asset pricing models (#834)

* update lecture

* updates

* updates

* updates

* Tom's light edits of affine lecture

* updates

* updates

* updates

* update

* update

* updates

* update lrp lecture

* updates

* updates

* minor fix

* updates

* Add optimistic initialization to Q-learning lecture (#830)

* Add optimistic initialization to Q-learning lecture

Initialize Q-table to 20 (above true value range of 13-18) instead of
zeros, which drives broader exploration via "optimism in the face of
uncertainty". This speeds convergence enough to reduce the training run
from 20M to 5M steps. Added a new subsection explaining the technique.

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

* Add optimistic initialization to risk-sensitive Q-learning lecture

Initialize Q-table to 1e-5 (below true Q-value range of ~1e-5 to 1e-4)
instead of ones. Since the optimal policy minimizes q, optimistic means
initializing below the truth — the reverse of the risk-neutral case.
This speeds convergence enough to reduce training from 20M to 5M steps.
Added a subsection explaining the reversed optimistic init logic.

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

* Add \EE macro to MathJax config

* Restore profit to observation list; fix risk-sensitive optimistic init value and narrative

---------

Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com>
Co-authored-by: Matt McKay <mmcky@users.noreply.github.com>

* Tom's March 16 edits of Backus lecture lecture

---------

Co-authored-by: thomassargent30 <ts43@nyu.edu>
Co-authored-by: John Stachurski <john.stachurski@gmail.com>
Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com>
Co-authored-by: Matt McKay <mmcky@users.noreply.github.com>

Author: 5 people