National Computing Education Accreditation

Council
NCEAC
NCEAC.FORM 001-D

COURSE DESCRIPTION FORM: DS-5007: Natural Language Processing

INSTITUTION FAST School of Computing, National University of Computer

and Emerging Sciences, Islamabad Campus

PROGRAM TO BE MS-DS: Spring-2021

EVALUATED

Course Description

1
NCEAC.FORM.001.D
 National Computing Education Accreditation
Council
NCEAC
NCEAC.FORM 001-D

Course Code DS-5007

Course Title Natural Language Processing
Credit Hours 3
Course Instructors Dr Mehreen Alam

Grading Policy Absolute grading

Policy about missed Retake of missed assessment items (other than sessional/ final exam) will not be held.
assessment items in Student who misses an assessment item (other than sessional / final exam) is awarded
the course zero marks in that assessment item i.e. late submission will not be accepted.
For missed sessional/ final exam, exam retake/ pretake application along with necessary
evidence are required to be submitted to the department secretary. The examination
assessment and retake committee decides the exam retake/ pretake cases.
Course Plagiarism Plagiarism in project or sessional/ final exam will result in F grade in the course.
Policy Plagiarism in an assignment will result in zero marks in the whole assignments category.
Prerequisites by Preferred but not necessary: Data Mining, Machine Learning
Course(s) or Topics
Assessment Assessment with the weight.
Instruments with Assessment Type Weight
Weights (homeworks,
quizzes, sessional Quizzes (4) 10
exams, final exam, Sessional Exams (1) 30
assignments, etc.)
Project 20
Final Exam 40
Course Coordinator Dr Mehreen Alam
URL (if any)
Course Catalog Introduction; Role of text mining in Computing and Artificial Intelligence, Bag of words
Description model, Parsing algorithms, CFGs, Representing meaning /Semantics, Semantic roles,
Temporal representations, Corpus-based methods, N-grams and HMMs, Smoothing and
backoff, POS tagging and morphology, Information retrieval, Vector space model,
Precision and recall, Information extraction, Language translation, Text classification,
categorization, Q&A. Word Embeddings, word2vec and GloVe models, Logistic
Regression, Neural Networks for text clasification, RNNs, LSTMs, Attention networks,
Transformers and other Deep Learning methods, Lexicons and Semi-Supervised
Learning.
Textbook Daniel Jurafsky and James H. Martin. 2018. Speech and Language Processing: An
Introduction to Natural Language Processing. Third Edition (or the latest). Prentice
Hall

Reference Material Steven Bird, Ewan Klein and Edward Loper.2019. Analyzing Text with the Natural
Language Toolkit. Natural Language Processing with Python. O'Reilly.
Proceedings of ACL, EMNLP, NAACL, TACL

2
NCEAC.FORM.001.D
 National Computing Education Accreditation
Council
NCEAC
NCEAC.FORM 001-D

Course Goals
A. Course Learning Outcomes (CLOs)

After course completion, the students shall be able to:

1. Identify techniques for information retrieval, language translation, and text

classification.
2. List the advantages of using standard corpora. Identify examples of current corpora
for a variety of NLP tasks.
3. Define and contrast Language models using traditional and Deep Learning Methods.
4. Understand and interpret recent research work in Natural Language Processing.

B. Program Learning Outcomes (PLOs)

1. To equip students to transform data into actionable insights to make complex
business decisions.
2. To develop understanding of the fundamental problems of NLP and the challenges
of NLP at large.
3. To enable students to understand and analyze a problem and arrive at computable
solutions.
4. To develop an awareness of the wide spectrum of NLP applications, and
technologies used to implement them.
5. Select and implement appropriate pre-processing techniques and derive a vector
representation from a given corpus of text.
6. Implement text classification algorithms on an annotated corpus of text.
7. To design and implement an end-to-end NLP system and evaluate its performance
using suitable metrics.

C. Mapping of CLOs to PLOs

(CLO: Course Learning Outcome, PLOs: Program Learning
Outcomes)
PLOs

1 2 3 4 5 6

1  
2  
CLOs

3  
4  
Topics covered in
the course Topics to be covered:
(assume 15-week
instruction and 3 No. of Contact
List of Topics
contact hours per Weeks Hours CLO(s)
week) Introduction to NLP, Data
1 3 1,2
Preprocessing Techniques
3
NCEAC.FORM.001.D
 National Computing Education Accreditation
Council
NCEAC
NCEAC.FORM 001-D

Data Curation 1 3 2

Language Modeling 1 3 2
Evaluation Metrics: bleu, glue, rouge,
1 3 1,2
wer, glue tasks, leaderboard
Loss Functions: cross entropy, kl
1 3 2
divergence, etc
Backpropagation 1 3 1
NLP Problems: text summarization,
chatbots, natural language inference, 1 3 2,3,4
question answering
Hands-on session with PyTorch and
1 3 1,4
Tensorflow
Transformers: self-Attention, multi-
headed attention, positional
1 3 1,3,4
encoding, embedding, residuals,
limitations
BERT: architecture, bert base vs bert
large, masked language modeling,
1 3 3,4
next sentence prediction, bert
variants
Hands-on building of a Machine
1 3 2
Translation System
RNNs, bi-RNNs, 1 3 3

LSTMs, Attention mechanism 1 3 3

Text to Vector Techniques:
Word2vec, architecture, 1 3 3
backpropagation
Text to Vector Techniques: TFIDF 1 3 3
Project Demos / Discussions /
1 3 1,2,3,4
Review
Total 16 48
Programming Preferred language is Python
Language for
Assignments
Class Time Spent Theory Problem Analysis Solution Design Social and Ethical
(in percentage) Issues
55 20 20 5
Oral and Written Every student is required to submit at least __5___ written reports of typically ___5____
Communications pages each and to make __1___ oral presentation of typically ____10___ minutes’
duration.

4
NCEAC.FORM.001.D