Pharmacy Meets AI - アンサング・シンデレラ 病院薬剤師の処方箋
Lecture 1 class notes
Please answer the following quiz questions when you finish lecture 1:
- Q1 what is the end of shoelaces called?
- Q2 once you give a name to something you get power over it, what is this principle called?
- Q3 how many possible representations can we have for the farmer goose grain and fox problem?
- Q4 how many countries does the equator cross in Africa?
Lecture 2 class notes
- Q1: Who is the phd advisor of James Slagle?
- Q2: Is James Slagle completely blind when he wrote SAINT (Symbolic Automatic INTegrator)?
- Q3: How much knowledge is needed to do integration at MIT freshmen calculus final?
- Q4: Should a program that is able to do symbolic integration of calculus be considered intelligent?
Lecture 3 class notes
Please answer the following quiz questions when you finish lecture 3:
- Q1: Which university developed the rule-based system to identify bacterial infection in the mid 1980s?
- Q2: Who is the AI expert that wrote "The Sciences of the Artificial" and has a chinese name 司马贺?
- Q3: When we tell a story, it is mostly a matter of controlled hallucination. Give one example of this statement in your life.
Lecture 4 class notes
Please answer the following quiz questions when you finish lecture 4:
- Q1: Give one example where you use your eyes (visual system) to find a pretty good path on a map, but not the optimal path.
- Q2: Is backtracking applicable to British Museum search?
- Q3: Hill climbing is to Depth First Search as what to Breadth First Search?
- Q4: Why do many people get frozen to death and fail to climb uphill to range stations on Mount Washington?
Lecture 5 class notes
- Q1: What is the name of the highway that is very popular among people who want to drive across USA because it is a short path?
- Q2: When we have an oracle giving us the answer, do we check if the oracle is trying to delude us?
- Q3: What is the Dead-horse Principle?
- Q4: When is a heuristic admissble?
- Q5: Is A-star search always the best way to find the shortest (optimal) path?
Lecture 6 class notes
- Q1: Hubert Dreyfus, a MIT philosopher, said "computers cannot play chess" in 1963. When he sat down and play aginst Greenblatt chess machine from MIT AI lab, what was the game result?
- Q2: In 1968, chess master David Levy bet on founding figures of AI, John McCarthy that no computers would beat chess world champion in 10 years. John McCarthy admit defeat in 5 years. Why did McCarthy give up?
- Q3: How would you construct a linear scoring polynomial function to evaluate the chess board situations? What features of the chess board would you consider, other than piece count for example?
- Q4: How many possible chess states are there as estimated by Claude Shannon? Explain why the entire universe since Big Bang would not be able to evaluate all the possible chess states.
- Q5: Will alpha-beta algorithm give different or approximate answers than that by min-max algorithm?
Lecture 7 class notes
- Q1: Adolfo Guzman went to purchase a lot of children's blocks to study the computer vision of detecting objects from lines, what is the name of the store he went to? (Hint: it is a precursor to Toys R Us)
- Q2: David Huffman is a mathematician, and explain why that might make him wanting to reject Guzman's PhD thesis?
- Q3: Give one example of a vertex formed by more than three planes/faces.
- Q4: Do you think we have similar constraint propagation apparatus that we use in our vision system?
Lecture 8 class notes
- Q1: Give one example of a map coloring situation where the color choice on the 4th state will get yourself into a dead end in your 48th state?
- Q2: When we run out of a suitable color choice for Texas, what should we do?
- Q3: For a 48-state USA map, what is the maximum and minimum number of colors you need to color the map?
- Q4: Besides the constraint of no single physical aircraft can fly two flights at the same time (no same time constraint), what other constraints can you think of for allocating resources to schedule airline flights?
Lecture 9 class notes
- Q1: Why is David Marr's theories treated like Gospels in computer vision research?
- Q2: Shimon Ullman, Marr's phd student, proposed to use alignment theory to recognize objects. What kind of projection does he use?
- Q3: How does Ullman's correlation theory use Goldilock's principle?
- Q4: Brothers Grimm collected fairy tales, and which fairy tale gives the Rumpelstiltskin principle?
- Q5: Why do you think we can recognize faces even from the side-view?
Lecture 10 class notes
- Q1:...