HOMEWORK 2 (CS 2731 / ISSP 2230)

Assigned: October 4, 2013

Due: October 22, 2013 (midnight)

There are both written and programming parts to this homework. The written part will be turned in via CourseWeb to facilitate automatic grading.

1. Multiple Choice (30 points)

In CourseWeb, there is now a Course Documents link under the Syllabus link in the top/lefthand box. When you click on Course Documents, the folder's contents will appear on the main part of the screen - currently "HW2: Written Questions." Click on that.

This multiple choice part of the homework will be automatically graded. You can only submit your answers once, which will trigger the grading. I will go over the answers in class after the assignment deadline has passed.

2. Programming (70 points)

2.1 HMM Decoding (Viterbi) (35 points)

Implement the Viterbi algorithm (Fig. 5.17 on page 147 in Jurafsky and Martin).
- (25 points) Demonstrate the correctness of your implementation by running it with the HMM in Figure 6.3 on page 178 to compute the most likely weather sequence for the following observation sequence: 331122313
- (10 points) We will also test your implementation on other blind tests (including same HMM but new observation sequence, and a new HMM).

Your program should accept two command line parameters. The first parameter is the filename of the HMM model; the second one is the input observation sequence you want to decode. Your program should output to standard output the decoded results (sequence of the hidden states). An example is shown below.
- ./hmm_decode model.txt 331122313

An example of the model file for Figure 6.3 (Comments shown here are NOT part of the file), download it from here.

Alphabet
3				#number of alphabets
1	2	3		#alphabets, split by '\t'

States
2				#number of states
HOT	COLD			#states, split by '\t'

StartProbability		#prior probabilities of the states, the order is the same as before
0.8	0.2

TransitionProbability		#transition probability matrix(T), N*N, N is the number of states, T[i][j] = p(s_j|s_i)
0.7	0.3
0.4	0.6

EmissionProbability		#emission probability matrix(E), N*M, N is the number of states, M is the number of alphabets. E[i][j] = p(a_j|s_i)
0.2	0.4	0.4
0.5	0.4	0.1

2.2 Parsing (Earley) (35 points)

Implement a basic version of the Earley algorithm (Fig 13.13 on page 444 in Jurafsky and Martin).
- (15 points) Demonstrate the correctness of your implementation by parsing "book that flight" using the L1 grammar from the textbook (Fig. 13.1 on page 428).
- (10 points) We will also test your implementation using L1 to parse some other blind test sentence, and using a new grammar as well.
(5 points) Augmentation 1: Modify your basic algorithm so that parse trees are retrieved from the chart. To do this, modify the psuedocode for COMPLETER as described on page 448 and reparse "book that flight." We will also rerun a blind test.
(5 points) Augmentation 2: Modify your basic algorithm so that it makes better use of bottom-up information to reduce the number of useless predictions. Parse an example with and without Augmentation 2 that illustrates the advantage of your approach (and explain why!).
Your 3 programs (which you should call earley_parse_basic, earley_parse_a1, earley_parse_a2) should each accept two command line parameters. An example is shown below. The first parameter is the filename of the grammar and lexicon; the second one is the input sentence that you want to parse. Your program should output to standard output the appropriate results for each version of the parser. For earley_parse_basic, the output is just whether the string is recognized or not. For earley_parse_a1, the output should in addition include all parse tree(s) for the input sentence. You can define how to print your parse tree output yourself. For early_parse_a3, you need to output some processing information to show how you eliminate some predictions from the basic algorithm.
- ./earley_parse_basic grammar.txt "book that flight"

An example of the grammar file for Figure 13.1, download it from here.

Grammar
S->NP VP
S->Aux NP VP
S->VP
NP->Pronoun
NP->Proper-Noun
NP->Det Nominal
Nominal->Noun
Nominal->Nominal Noun
Nominal->Nominal PP
VP->Verb
VP->Verb NP
VP->Verb NP PP
VP->Verb PP
VP->VP PP
PP->Preposition NP

Lexicon					#grammar and lexicon are split by a blank line
Det->that|this|the|a
Noun->book|flight|meal|money
Verb->book|include|prefer
Pronoun->I|she|me
Proper-Noun->Houston|NWA
Aux->does
Preposition->from|to|on|near|through

Submission Procedure

The submission should be done using the Assignment Tool in CourseWeb/ Blackboard. The file should have the following naming convention: pittID_hw2.zip (ex: dlitman_hw2.zip). The required outputs, the code, and a readme file to run your code should be submitted inside the archived folder. Clarification 10/22: Please include in your README some explanation of your code to demonstrate that you wrote this code (e.g. you could specify how specific parts of your actual code correspond to each specific part of the psuedocode, what you did to flesh it out etc.).

The date in CourseWeb will be used to determine when your assignment was submitted (to implement the late policy).