CS288 Natural Language Processing Spring 2011 Assignments

[email protected]

a1: A fast, efficient Kneser-Ney trigram language model.

a2: Phrase-Based Decoding using 4 different models.

• monotonic beam-search decoder with no language model
• monotonic beam search with an integrated trigram language model
• beam search that permits limited distortion
• last model unknown yet

a3: Word-based alignment.

• heuristic aligner
• IBM model 1 aligner with hard EM and soft EM implementations
• HMM aligner

a4: semantic parser and classifier (skipped)