Homework 0
- Download non-case sensitive 2-grams.
- Build a bigram model by measuring P(wi+1|wi) for each bigram.
- Given the first word being "dear", generate 10 sentences of length 20 where each bigram probability P(wi+1|wi) is greater than certain thresholds.
- Try out at least 5 different thresholds and compare the results.
- Write a report including your approach, results, and analysis. Use the ACL latex template.
- Compress your code and report into
hw0.zipand submit it to: https://canvas.emory.edu/courses/29596/assignments/28821
