Computer Science Exam

Just so y’all know that you did­n’t do ALL of my exam study work for me…

CS1002/01 — Final Exam­i­na­tion — Sum­mer, 2003

1. What was the most impor­tant thing you learned in this class?
That I had absolute­ly no idea what “com­put­er sci­ence” tru­ly meant. I have a fair amount of IT expe­ri­ence, but no CS expe­ri­ence at all.

2. Who was John von Neu­mann and why was he impor­tant in the devel­op­ment of computing?
He was a math­e­mati­cian and physi­cist who was born in Hun­gary and emi­grat­ed to the Unit­ed States. He is cred­it­ed with the stored pro­gram con­cept and the log­i­cal com­put­er design (von Neu­mann machine mod­el) that is used in all com­put­ers today.

3. Give a brief def­i­n­i­tion of the type of AI known as an “expert system”.
Expert sys­tems are designed to repro­duce the prob­lem-solv­ing behav­ior of an expert in a par­tic­u­lar, nar­row­ly-defined field. They store infor­ma­tion orga­nized by rules gleaned from those experts.

5. How should you eval­u­ate infor­ma­tion found via a search engine? Why?
The results should be care­ful­ly eval­u­at­ed based on the cred­i­bil­i­ty of sources, dates, and rel­e­vance to your top­ic. Check­ing mul­ti­ple sources is high­ly rec­om­mend­ed, and it is often wise to check snopes.com and sim­i­lar sites if the infor­ma­tion seems out­ra­geous in any way.

6. List and define the three lev­els of for­mal meth­ods in soft­ware engineering.
Infor­mal — no con­straint on the mod­el you develop
Semi-for­mal — well-defined syn­tax (pseudocode)
For­mal — rig­or­ous­ly defined (with math) syn­tax and semantics

8. Give two exam­ples of dif­fer­ences between “big AI” and “small AI”.
Big AI — large scale, gen­er­al, hard to do, advances gen­er­al knowledge—for instance, rec­og­niz­ing spo­ken Eng­lish from any­one say­ing anything
Small AI — small scale, spe­cif­ic, eas­i­er to do, may not advance gen­er­al knowledge—for instance, rec­og­niz­ing one per­son say­ing a few words

9. Does Michie’s match­book com­put­er exhib­it intel­li­gent behav­ior? Does it learn? How can you do one with­out the other?
It does not tru­ly “learn” or exhib­it intel­li­gent behav­ior, but it does improve its algo­rithms over time. It can­not exhib­it spon­ta­neous behav­ior to solve a prob­lem, nor does it have any cog­ni­tive capa­bil­i­ty. It can­not cre­ate any­thing new. It does not per­ceive any rela­tion­ships but is pro­grammed to always select the move (box) con­tain­ing the most beads. That is a sim­ple cal­cu­la­tion. Stor­ing infor­ma­tion is not the same as learn­ing, as the infor­ma­tion does not nec­es­sar­i­ly have any relevance—witness how many chil­dren learn to rat­tle off mul­ti­pli­ca­tion tables by rote but are not able to actu­al­ly DO mul­ti­pli­ca­tion correctly.

10. Why should you learn how to ana­lyze algorithms?
Fail­ure to ana­lyze algo­rithms and use the best method for the task at hand results in wast­ed resources. The most mem­o­rable exam­ple giv­en involved a process that would take eight months using the most obvi­ous algo­rithm and about one hour using anoth­er one.

13. What are the three kinds of things you can say to a com­put­er in a program?
‑sequences of direct instruc­tions (assign­ment, I/O)
‑selec­tion rules to decide what to do next (if case)
‑iter­a­tion of a group of instruc­tions (for, while)

14. Describe and illus­trate with a dia­gram the “fetch-inter­pret-exe­cute” cycle of a stan­dard computer.
Stored pro­grams are run by fetch­ing the instruc­tion, inter­pret­ing or decod­ing it, then exe­cut­ing it, and repeat­ing that cycle.

(I can’t real­ly put in what I drew, but it was very simple.)