Part 8: Appendices

Lectures 8 ( part ) , 9 and 10 ( part )

The Ellipsoid algorithm was developed by (formerly) Soviet mathematicians (Shor (1970), Yudin and Nemirovskii (1975)). Khachian (1979) proved that it provides a polynomial time algorithm for linear programming. The average behavior of the Ellipsoid algorithm is too slow, making it not competitive with the simplex algorithm. However, the theoretical implications of the algorithm are very importa...

Part 8. Cognitive illusions.

Evolution’s view Many optical illusions have evolutionary survival value, and thus many are hard-wired. We can train ourselves to be less influenced by such illusions, but we can’t eliminate them entirely. The analogous errors of the cognitive system are what I call cognitive illusions (or distortions in orthodox cognitive behavioural therapy literature). Overgeneralization is a common cognitiv...

Sequential Bayesian Search Appendices

Assume that at the beginning of game t, the system's belief in the user's preference is P t. Then, the certainty-equivalet user preference during game t is π * t (i) = E π∼Pt [π(i)] ∀i ∈ I. Recall we define π * min = min i∈I π * (i), Lemma A-1 formalizes the result that if π * t is " close " to π * , then for any decision tree T , E i∼π * t [N (T, i)] is " close " to E i∼π * [N (T, i)]: Lemma A...

PART 8 Robustness , Robust Control

Verbalizing Business Rules: Part 8

Business rules should be validated by business domain experts, and hence specified using concepts and languages easily understood by business people. This is the eighth in a series of articles on expressing business rules formally in a high-level, textual language. The first article [4] discussed criteria for a business rules language, and verbalization of simple uniqueness and mandatory constr...