ASSIGNMENT-I : Part-of-Speech (PoS) tagging using HMM

CS563 - NLP
ASSIGNMENT-I : Part-of-Speech (PoS) tagging using HMM and Recurrent
Neural Network
(Read all the instructions carefully and adhere to them)

Instructions:
1. Markings will be based on the correctness and soundness of the outputs.
2. Marks will be deducted in case of plagiarism.
3. Proper indentation and appropriate comments (if necessary) are mandatory.
4. You should zip all the required files and name the zip file as:
<roll_no>_assignment_<#>.zip , eg. 1501cs11_assignment_01.zip.
5. Upload your assignment ( the zip file ) in the following link:
https://www.dropbox.com/request/w5PQuhRtWIMSFvFtFVVh

__________________________________________________________________
Problem Statement: Part-of-Speech (PoS) tagging assigns grammatical categories
to every token in a sentence. In this assignment, you have to develop a PoS tagger
using Hidden Markov Model (HMM) and Recurrent Neural Network.
Dataset: WSJ (Wall Street Journal)
Number of PoS tags: 46
List of tags : 'MD', 'TO', 'WP', 'WP$', '.', 'PRP', 'PDT', '#', 'POS', 'VBN', '-RRB-',
'DT', "''", ':', 'EX', 'RP', 'RBR', '-NONE-', 'UH', 'VBZ', 'VBG', '$', 'RBS', 'JJR', 'IN',
',', 'VBD', 'LS', 'JJS', 'WRB', 'VBP', '-LRB-', 'NNP', 'NNS', 'PRP$', 'JJ', 'CC', 'FW',
'CD', 'VB', 'NN', 'NNPS', 'SYM', 'WDT', '``', 'RB'
Link to download the dataset:
https://drive.google.com/file/d/1GnH_RD087pyyMwJr4JoQSsDv-q9TrUB0/view?
usp=sharing
● Hidden Markov Model (HMM)
You have to implement HMM on your own. Do not use any existing
libraries. Consider a bigram HMM model. Calculate the Emission and
Transition Probability matrices. Use Viterbi decoding to obtain the best PoS
sequence.
● Recurrent Neural Network:
○ You may consider the following details for the implementation.
● Input Vec(Wi ): The word embeddings will be the input to the
model. You can use the Word2Vec or GLOVE embedding.
● Link → Word2vec: http://vectors.nlpl.eu/repository/20/5.zip or
https://drive.google.com/file/d/0B7XkCwpI5KDYNlNUTTlSS
21pQmM/edit?usp=sharing
● Link→ Glove: http://nlp.stanford.edu/data/glove.840B.300d.zip
● Output (Ti): Sequence of POS tags.
○ You may use any deep learning libraries such as TensorFlow, PyTorch,
Keras etc. for the implementation. Use 300 dimensions for word
embeddings.
Evaluation:
1. Perform 3-fold cross-validation.
2. Compute the overall accuracy for each of HMM and RNN models for each
of the 3 folds.
3. Show the class-wise accuracy of the best-performing fold (i.e. out of the
3-folds, you have to show for the fold that shows the best performance).