Final Exam, Due by 5:55 pm on May 15

You must submit your answers electronically. Your answers should be a word processed document on no more than 5 pages (minimum 10 point font). The file should be named final_exam.doc (if produced by MS Word). To upload the code go to the link and follow the directions for uploading a file (you can do this multiple times, though it would be helpful if you would tar your files and upload one file archive).

  1. Of the 14 papers covered by the student presentations, which of the papers specifically covered issues of machine learning involved learning sequence information (learning where the information to be learned was organized as sequences of related data items)? Justify your answer (both in terms of which papers did and which did not involve sequential learning). [20 points]
  2. Of the 14 papers covered which of them covered natural language or text processing (or could be applied to natural language or text processing)? Be sure to justify your answer. [20 points]
  3. For which of the 14 papers could you imagine employing decision trees as part of a solution similar to the proposed solutions in the paper? Be sure to justify your answer. [20 points]
  4. For each of the 14 papers, describe the validation method used to demonstrate the central thrust of the paper. (For example, proofs of key concepts, empirical tests of ..., etc.). [20 points]
  5. Which algorithm or area of machine learning (that we discussed in class) do you think will be the most important over the next 20 years? Be sure to justify your answer. [20 points]