Artificial Intelligence Undergraduate Curriulum
Artificial Intelligence Undergraduate Curriulum
Background:
The Rawalpindi Medical University (RMU) has taken the initiative and lead by
starting a program in Artificial Intelligence at the graduate level to help establish
and grow the industry in medicine field in Pakistan.
Curriculum:
The proposed curriculum is unified for all RMU partner universities. For the sake
of uniformity and ease of transfer of courses, a national course code has also been
defined for each course. This will be treated as a reference for course
compatibility between RMU partner institutions.
Core Courses
National
Course Course Name CHs
Code
Elective Courses
National
Course Course CHs
Code
Textbook
• Reinforcement Learning: An Introduction by Sutton and Barto. Also
available at http://incompleteideas.net/book/the-book-2nd.html
• Handouts and research articles may also be used by the instructor.
Objective
To realize the dreams and impact of AI requires autonomous systems that learn
to make good decisions. Reinforcement learning is one powerful paradigm for
doing so, and it is relevant to an enormous range of tasks, including robotics, game
playing, consumer modeling and healthcare. This class will provide a solid
introduction to the field of reinforcement learning and students will learn about
the core challenges and approaches, including generalization and exploration.
Through a combination of lectures, and written and coding assignments, students
will become well versed in key ideas and techniques for RL. Assignments will
include the basics of reinforcement learning as well as deep reinforcement
learning — an extremely promising new area that combines deep learning
techniques with reinforcement learning. In addition, students will advance their
understanding and the field of RL through a final project.
Pre-Requisite
Linear Algebra, Probability, Machine Learning and Deep Learning
Course Outcome
During the course, students will learn to:
Course Outline
Allocated
Topics
Periods
• Introduction to Reinforcement Learning 45
• Markov Processes
• Policy Search and Iteration
• Value Iteration
• Policy Evaluation
• Model Free Learning: Q-Learning and SARSA
• RL with function approximation
• Imitation Learning in Large spaces
• Exploration/Exploitation
• Batch Reinforcement Learning
Textbook
• Probabilistic Graphical Models: Principles and Techniques by Daphne
Koller and Nir Friedman. MIT Press.
• Handouts and research articles may also be used by the instructor.
Objective
Probabilistic graphical models are a powerful framework for representing
complex domains using probability distributions, with numerous applications in
machine learning, computer vision, natural language processing and
computational biology. Graphical models bring together graph theory and
probability theory, and provide a flexible framework for modeling large
collections of random variables with complex interactions. This course will
provide a comprehensive survey of the topic, introducing the key formalisms and
main techniques used to construct them, make predictions, and support decision-
making under uncertainty.
Pre-Requisite
Basic probability theory, statistics, programming, algorithm design and analysis.
Course Outcome
During the course, students will learn to develop the knowledge and skills
necessary to design, implement and apply these models to solve real problems.
Be able to describe directed and undirected graphical models, exact and
approximate inference methods and estimation of parameters.
Course Outline
Allocated
Topics
Periods
• Introduction, Probability Theory, Bayesian Networks 45
• Undirected models
• Learning Bayes Nets
• Exact Inference; Message Passing
• Sampling
• MAP Inference; Structured prediction
• Parameter Learning
• Bayesian Learning; Structure Learning
• Exponential families; variational inference
Topics Allocated
Periods
Introduction to AI 45
Definitions (Acting Humanly, Cognitive, “laws of
Thought”, Rational agent
approaches
Historical perspective
Physical symbol system hypothesis
Intelligent agents
Agents and Environment
The concept of rationality
Performance measures
Omniscience, learning and autonomy
Nature of environments, Task environments and their
Properties
Search Strategies
Uninformed Searches
Breadth first
Depth first
Depth limited
Iterative deepening depth-first
Bidirectional
Comparison of Uninformed problem-solving methods
Informed searches
Greedy best-first
A*
Heuristic Functions (learning, devising)
Local search algorithms
Hill climbing
Simulating annealing
Local beam
Genetic algorithms
Logical agents
Knowledge-based agents
Propositional logic (syntax, semantics)
Conjunctive/disjunctive Normal form, Horn clauses
Reasoning in Propositional logic
Resolution
Forward and backward chaining
Reasoning algorithms
First-Order Logic (syntax, semantics)
Inference in FOL
Conjunctive Normal form
Unification and lifting, forward and backward
chaining, resolution
Comparison of two representational languages
Machine Learning
Introduction, induction, Types of machine learning
Nearest Neighbors
Decision Trees
Neural Networks
Learning Conjunctions
Linear and non-Linear
separabilityEvaluating learning
algorithms
Topics Allocated
Periods
Image Formation 45
- Sampling
- Quantization
- Geometry
- Photometry
- Color
Calibration
Filtering and Convolution
Feature Extraction
- Corners, edges, regions
Segmentation and grouping
- Parametric fitting,
- Hough transform,
- Tensor Voting
Stereo
Structure and motion
Dense motion and Optical flow
Range Image Analysis
Recognition
Textbook
1. Deep Learning. By Ian Goodfellow, Yoshua Bengio, and Aaron Courville,
MIT Press, 2016. Also available at www.deeplearningbook.org
Handouts and research articles may also be used by the instructor.
Objective
2. Computer Vision has become ubiquitous in our society, with applications
in search, image understanding, apps, mapping, medicine, drones, and self-
driving cars. Core to many of these applications are visual recognition tasks
such as image classification, localization and detection. Recent developments in
neural network (aka “deep learning”) approaches have greatly advanced the
performance of these state-of-the-art visual recognition systems. This course is
a deep dive into details of the deep learning architectures with a focus on learning
end-to-end models for these tasks, particularly image classification.
Pre-Requisite
3. Linear Algebra, Probability, Machine Learning
Course Outcome
4. During the course, students will learn to implement, train and debug their
own neural networks and gain a detailed understanding of cutting-edge research
in computer vision and its application for robotics. The final assignment will
involve training a multi-million parameter convolutional neural network and
applying it on a large dataset.
Course Outline
4. The course covers the techniques and technology to set up the problem of image
recognition, the learning algorithms (e.g. backpropagation), practical engineering
tricks for training and fine-tuning the networks.
Allocated
Topics
Periods
• Machine Learning Basics 45
• Deep Feedforward Networks
• Regularization for Deep Learning
• Optimization for Training Deep Models
• Convolutional Neural Networks
• Sequence Modeling: Recurrent Neural Networks
• Applications of Deep Learning
Textbook
1. Fundamentals of Robotic Grasping and Fixturing. Caihua Xiong, Han
Ding, and Youlun Xiong, World Scientific Publishing Company, 2007.
ISBN-13 978-981-277-183-4, ISBN-10 981-277-183-2
Handouts and research articles may also be used by the instructor.
Objective
2. This course focuses on providing comprehensive information and
mathematic models of developing and applying grippers and fixtures in industry,
and present long term valuable essential information for the academic researchers
who are interested in robotic manipulation as a good reference.
Pre-Requisite
3. Robot Mechanics and Control
Course Outcome
4. This course will furnish the students with a comprehensive insight into
robotic grasping and fixturing. It involves study of multifigured robot hand grasp,
basic fixture design principle, and evaluating and planning of robotic
grasping/fixturing, and focuses on the modeling and applications of Robotic
Grasping and Fixturing.
Course Outline:
Allocated
Topics
Periods
Robotic Grasp and Workpiece-Fixture Systems 48
• Introduction
• Robotic Manipulation and Multifingered Robotic
Hands
• AMT and Fixtures
• Comparison between Grasping and Fixturing
Objective
2. A wearable robot is a mechatronic system that is designed around the shape
and function of the human body, with segments and joints corresponding to those
of the person it is externally coupled with. Tele-operation and power amplification
were the first applications, but after recent technological advances the range of
application fields has widened. These robots have to support the motion of human
limbs and are valuable in the process of rehabilitation as strength augmentation.
This course will enable the students to get an insight into the design and control
challenges of such robots.
Pre-Requisite
3. Linear Algebra, Robot Mechanics and Control
Course Outcome
4. By the end of this course the students will be well-versed in Biomechanics
and Human Interaction. Models of Bio-inspiration along with their actuation
requirements will also be introduced to the students and designof systems that
enable these motions will be focused.
Course Outline
5. The mechanics of the wearable robot and its biomechanical interaction with
the user, including state-of-the-art technologies will be covered. The basis for
bioinspiration and biomimetism, general rules for the development of
biologically-inspired designs, and how these could serve recursively as biological
models to explain biological systems and finally the incorporation of this
knowledge in the design of exoskeletons and prosthetic limbs will be taught.
Topics Allocated Periods
Textbook
1. Designing Sociable Robots (Intelligent Robotics and Autonomous Agents
series) by Cynthia Breazeal ISBN: 9780262025102. The course will also be
research paper based. Depending upon the module, the instructor(s) may also
assign a reference book.
Reference Books
2. • Human-Robot Interaction in Social Robotics by Hiroshi Ishiguro and
Takayuki Kanda
• Human-Robot Interactions: Principles, Technologies and Challenges.
Editor: Diana Coleman.
Pre-Requisite
3. CSE 860 Artificial Intelligence (or equivalent), Robotics, Strong
programming background.
Objective
4. This is a highly research-oriented course with the basic objective of
bringing the students face to face with the latest cutting-edge research in the
world. This will play a key role in helping them select suitable topics for their
research that would lead to publications in leading international journals and
conferences
Course Outcome
5. Students completing this course are expected to have developed a firm
knowledge base of critical understanding of the main theories, principles and
concepts relating to the domain of Human-Robot Interaction.
Course Outline
6. This course will cover a variety of topics related to social intelligence and
socially intelligent robots. Readings/lectures will cover (1) what’s known about
how this ability arises in human intelligence, and (2) state-of-the-art approaches
to building computational systems with this type of social ability.
Textbook
1. Probabilistic Robotics. By Sebastian Thrun, Wolfram Burgard, and Dieter
Fox, MIT Press (2006), ISBN: 978-0-262-20162-9.
Handouts and research articles may also be used by the instructor.
Objective
2. This course focuses on robot perception and control in the face of
uncertainty. Building on the field of mathematical statistics, probabilistic robotics
endows robots with a new level of robustness in real-world situations.
Course Outcome
3. This course will furnish the students with a practical experience in robot
perception in partially known environments through implementations in pseudo
code, detailed mathematical derivations, discussions from a practitioner's
perspective, and extensive lists of exercises and class projects
Course Outline:
Topics Allocated Periods
Introduction 45
• Uncertainty in Robotics
• Probabilistic Robotics
Bayes Filter
• Gaussian Filters
• Kalman Filter
• Extended Kalman Filter (EKF)
Nonparametric Filters
• Histogram Filter
• Particle Filter
Localization
• Markov Localization
• EKF Localization
• Multi-Hypothesis Tracking
• Monte Carlo Localization
Occupancy Grid Mapping
Simultaneous Localization and Mapping
• EKF SLAM
• Graph SLAM
10. RIME 843: Sensors and Sensing
Textbook
1. Handbook of Modern Sensors. By Jacob Fraden, Springer, 2010.
ISBN-13: 978-1-4419-6465-6
Reference Books
2. Introduction to Autonomous Mobile Robots. By Roland Siegwart and Illah
R. Nourbakhsh, The MIT Press, 2004.ISBN-10: 0-262-19502-X, ISBN-13:978-
0-262-19502-7
Probabilistic Robotics. By Sebastian Thrun, Wolfram Burgard, and Dieter
Fox, MIT Press (2006)
ISBN: 978-0-262-20162-9.
Objective
3. This course focuses on various sensors used for Robot Navigation and
Control. Its objective is to teach students about sensing and modeling surrounding
environment for motion planning and navigation.
Course Outcome
4. This course will furnish the students with an understanding of the design
and working of sensors for Robotics while also teaching them about the
techniques needed for signal processing of the sensor data.
Course Outline:
Topics Allocated Periods
• Sensor Characteristics 45
• Calibration
• Accuracy
• Repeatability
• Data Acquisition from Sensors
• Odometers
• Heading Sensors
• Accelerometer
• Inertial Measurement Unit
• Vision Sensing
• Range Sensing
• Sonar Sensing
• Flow Sensing
• Touch Sensing
• Sensing for Surveillance
• People Sensing
• Multi-Sensor Fusion
11. RIME 837: Simultaneous Localization and Mapping
Textbook
1. Probabilistic Robotics. By Sebastian Thrun, Wolfram Burgard, and Dieter
Fox, MIT Press (2006)
ISBN: 978-0-262-20162-9.
Reference Book
2. Introduction to Autonomous Mobile Robots. By Roland Siegwart and Illah
R. Nourbakhsh, The MIT Press, 2004. ISBN-10: 0-262-19502-X, ISBN-13:978-
0-262-19502-7.
Handouts and research articles may also be used by the instructor.
Objective
3. This course focuses on Robot Localization and Mapping in unknown or
partially known environments.
Course Outcome
4. This course will furnish the students with practical knowledge of SLAM
algorithms required for mapping and navigation in unknown of partially known
environments.
Course Outline
Textbook
1. Biorobotics–Methods and Applications, edited by Barbara
Webb and Thomas R. Consi.
Objective
3. The objective of this course is to develop expertise of the multidisciplinary
field of BioRobotics. The course involves detailed study of Biosignals and the
principals of Bio Mechanical Designs.
Course Outcome
4. After studying this course students will be able to use HumanNeurological
Signals for operations of various kinds of Robots. The course will also enable
students to develop Biomatic Robotic systems.
Course Outline
Course Outline:
This course will present an introduction to algorithms for machine learning
and data mining. These algorithms lie at the heart of many leading-edge
computer applications including optical character recognition, speech
recognition, text mining, document classification, pattern recognition,
computer intrusion detection, and information extraction from web pages.
Every machine learning algorithm has both a computational aspect (how to
compute the answer) and a statistical aspect (how to ensure that future
predictions are accurate). Algorithms covered include linear classifiers
(Gaussian maximum likelihood, Naive Bayes, and logistic regression) and
non-linear classifiers (neural networks, decision trees, support-vector
machines, nearest neighbor methods). The class will also be introduced to
techniques for learning from sequential data and advanced ensemble
methods such as bagging and boosting.
Topics Allocated
Periods
Overview/Introduction to machine learning 45
Hypothesis spaces
Space of Algorithms, Linear Threshold Classifiers
Project details
Perceptrons
Logistic Regression
Linear Discriminant Analysis
Off-The-Shelf Learning Algorithms
Decision Trees
Nearest Neighbor
Neural networks
Bayesian Learning
Support Vector Machines
Learning Theory
Learning Theory finished
Bias/Variance Theory & Ensemble Methods
Preventing Over-fitting: Penalty and Hold-out
methods
Hold-Out and Cross-validation Methods
Penalty methods: decision trees, neural nets, SVMs
Evaluating and Comparing Classifiers
Unsupervised Learning
Textbook
• Shoham, Y. & Leyton-Brown, K. (2009). Multiagent Systems: Algorithmic,
Game-Theoretic, and Logical Foundations. The book is available at
http://www.masfoundations.org/
• Handouts and research articles may also be used by the instructor.
Objective
Multiagent systems (MAS) can be defined as loosely coupled networks of
problem solvers that interact to solve problems that are beyond the individual
capabilities or knowledge of each problem solver. These problem solvers, often
called agents, are autonomous and can be heterogeneous in nature.
Research and development in MAS are concerned with the study and
construction of a collection of autonomous agents that interact with each other
and their environments. The study of such systems goes beyond the study of
individual intelligence in its consideration of problem solving with social
components.
This course introduces the students to the main topics in the theory and practice
of MAS, currently one of the most important and rapidly expanding areas of
computer science, having emerged from the study of distributed artificial
intelligence (DAI). Multiagent systems have been used as an important means
with which to address the development of large and complex information systems
(IS) and decision support systems (DSS). Finally, this course introduces
coalitional game theory and its potential applications.
Pre-Requisite
The student should be comfortable with mathematical notation and basic
computer algorithms
Course Outcome
During the course, students will learn to:
Allocated
Topics
Periods
• Foundations 45
• Intelligent Agents and Multiagent Systems
• Multiagent Learning
• Social Choice
• Mechanism Design
• Multiagent Resource Allocation
• Coalition Game
Many modern systems where NLP is used for exploring patterns in the data, or
in the cases where multimedia including text is considered, it becomes important
to see relationships between different objects and patterns. Thus,ontologies need
to be built for reasoning and knowledge representation becomes the key to it.
Recommended Books:
1. Knowledge Representation and Reasoning, 3rd edition, Gerardus Blokdyk,
Emereo Pty Limited, 2018.
2. Knowledge Representation and Reasoning with Deep Neural Networks,
Arvind Ramanathan Neelakantan, University of Massachusetts Libraries,
2017.
3. Knowledge Representation and Reasoning, Ronald Brachman, Hector
Levesque, Elsevier, 2004.
Credit Hours: 3 (3, 0)
Course Objectives:
• To understand what representation is for knowledge.
• To use logic (both propositional and first order) for representation.
• To understand how reasoning can be done in logic, and what entailment
is.
• To understand the theory and principles of automated theorem proving.
• To get acquainted to tools that can be used for knowledge representation
and reasoning.
Allocated
Topics / Contents
Periods
Topics include propositional logic, its syntax and semantics.
Proof by resolution and automated theorem proving. First-
order logic, its syntax and semantics. Forward and backward
chaining. Semantic nets. Inheritance in taxonomies. Other 45
topics include other approaches and issues (e.g. predicate
logic, fuzzy logic, week and strong slot and filler
structures), knowledge acquisition, the frame problem,
symbolic reasoning under uncertainty (nonmonotonic
reasoning, augmenting a problem Solver), statistical
reasoning (e.g. probability and Bays Theorem, Bayesian
networks, Dumpster-Shafer theory), building
knowledge-based systems.
Mathematical modeling and its simulation is the key to understanding where the
system will be successful and where it will fail to perform as per the requirement.
Not every scenario can be generated in the lab environment and tested for its
efficacy. In most of the cases it is much cheaper to first simulate it and only
prototype it once it has been deemed to be workable.
Recommended Books:
1. Theory of Modeling and Simulation: Discrete Event & Iterative System
Computational Foundations, Bernard P. Zeigler, Alexandre Muzy,Ernesto
Kofman, Elsevier Science, 2018.
2. Modeling and Simulation: An Application-Oriented Introduction, Hans-
Joachim Bungartz, Stefan Zimmer, Martin Buchholz, Springer, 2014.
3. Modeling and Simulation of Complex Systems: A Framework for Efficient
Agent-Based Modeling and Simulation, Robert Siegfried, Springer
Vieweg, 2014.
Credit Hours: 3 (3, 0)
Course Objectives:
• To understand what problems may be most suitable to be modeled and
simulated as opposed to being physically implemented.
• To understand parameters and constraints that may define the scope of the
problem to be modeled and simulated.
• To be able to demonstrate coding and implementation skills in the available
tools to simulate some real-world problem, which is not easily expressed
using mathematics.
Allocated
Topics / Contents
Periods
Course contents include topics on system analysis,
classification of systems, and system theory basics and its
relevance to simulation. For modeling the course includes
model classification which refers to conceptualization, 45
abstraction, and simulation of the models. It also includes
topics related to simulation systems and languages. It also
covers topics such as Petri nets and finite automata and
their use in representing models. Both discrete and
continuous systems are covered. For continuous systems,
numerical methods are included. For testing, validity and
verification of models is discussed. Optimization is alsomade
part of it so that the developed systems are optimized
as per the set criteria.
NLP is one of the main tasks where AI is the key underlying technology.
Processing text in main cases is the main problem to be solved specially when
data is available in that form, and also when robots have to communicate with
humans they may resort to it. Also, its coupling with the speech processing unit
would give extra leverage to the robots to do their tasks in a more effective
manner.
Recommended Books:
1. Natural Language Processing Recipes: Unlocking Text Data withMachine
Learning and Deep Learning using Python, Akshay Kulkarni, Adarsha
Shivananda, Apress, 2019.
2. Deep Learning in Natural Language Processing, Li Deng, Yang Liu
editors, Springer 2018.
3. Python Natural Language Processing, Jalaj Thanaki, Packt Publishing,
2017.
Intelligent system relates to planning, execution, and learning and is essential for
us due to the nature of robotics that we are going to do where all the three aspects
above would be covered.
Recommended Books:
1. Learning and Execution of Object Manipulation Tasks on Humanoid
Robots, Waechter, Mirko, Scientific Publishing, 2018.
2. Machine Learning Methods for Planning, Steven Minton, Elsevier Science,
2014.
3. Using Neural Networks and Dyna Algorithm for Integrated Planning,
Reacting and Learning in Systems, National Aeronautics and Space Adm
Nasa, Independently Published, 2018.
Credit Hours: 3 (3, 0)
Course Objectives:
• To comprehend the challenges in developing autonomous systems oragents
that plan, execute, and learn from the consequences of their actions.
• To understand what planning is, how it is represented, how it is evaluated,
and how it is learned.
• To implement a simple complete system which shows autonomous
behavior.
Allocated
Topics / Contents
Periods
The topics include closed-world assumption, linear
programming, GPS algorithm, prodigy algorithm, S-Space
transformation, constructing graphs of the planning
problem, Planning graphs for planning and heuristic search, 45
A* and weighted A* search, multi-heuristic A*, probabilistic
path planning, planning under uncertainty, reinforcement
learning, experience graphs, learning in planning, deep
reinforcement learning, learning cost functions, multi-agent
decision making, and short-sighted
probabilistic planning.
Course Objectives:
• To be able to use Fast Fourier Transform in a variety of applications such
as signal analysis, fast convolution, spectral and temporal interpolation, and
filtering.
• To understand how to design digital filters.
• To be able to construct a simple digital communication system.
Allocated
Topics / Contents
Periods
The topics include digital systems and signals, impulse
response, frequency response, discrete-time Fourier
transform, z-transform, interpolation and decimation,
sample rate changes, polyphase decomposition, random
signals and vectors, joint random signals and their statistical 45
description, Stationarity and ergodicity, Linear systems with
stationary random inputs, Power spectral density, Metric
spaces, vector spaces, normed vector spaces, inner projects
and inner-product spaces, Hilbert spaces, Projections,
Orthogonality, Orthogonalization of vectors, Approximation
in Hilbert spaces, The orthogonality principle, Least squares
filtering, Linear prediction and AR spectrum estimation,
MMSE estimation and MMSE filtering, Weiner filtering,
Signal transformation and generalized Fourier series, Scaling
functions and wavelets, Wavelet transform and its
implementation, Eigenvalues and linear systems/operators,
Diagnolization of a matrix, Karhunen-Loeve expansion,
Principal component analysis, Pisarenko harmonic
decomposition, MUSIC, SVD and applications, Applications
of adaptive filtering, LMS
algorithm, RLS algorithm.
Credit Hours: 3
Course Objectives:
• To understand the basic principles of speech perception.
• To understand basic principles of speech recognition, synthesis, and
dialogue.
• To be able to generate speech datasets and build applications to recognize
speeches and speakers.
Allocated
Topics / Contents
Periods
The topics include physical and physiological acoustics,
perception of speech, phonetics and phonology, signal
processing, speech synthesis, prosody and emotions, hidden 45
markov models, language modelling, human-computer
communication, and dialogue systems. Other topics include
linear-predictive model, cepstral analysis, speech coding,and
parameterization of DTW.
Allocated
Topics / Contents
Periods
Computational creativity is a multidisciplinary field thatlies
at the intersection of artificial intelligence, cognitive
psychology, philosophy, and the arts. The field is concerned 45
with the theoretical and practical issues in the study of
creativity. This course is about Computational Creativity
with a focus on modeling/discovery & design/invention.
Contents of the course include, human creativity,information
processing theories of human creativity, interactive tools for
augmenting and amplifying human creativity, design of
autonomous creative systems for
creative tasks.
Allocated
Topics / Contents
Periods
The field of modeling and simulation is large and diverse;
modeling and simulation is applied in every discipline to
answer questions in research and development and for 45
education and training. Recently there has been a push in the
use of specific simulations which have an internal goal. Such
simulations are known as games (when applied toeducation
and training, they are known as serious games) and provide
an experience that is fun and engaging making them
particularly relevant to the current generation of learners.
This course is an introduction to simulation with an emphasis
on serious games. Primary contents which would be covered
in this course will include, Ethical considerations of using
games to change behavior, Game interfaces, Universal
design, Research methodologies for game interactions,
Game telemetry, Game metrics, AI
evaluation techniques for "big data" from game telemetry,
Evaluation techniques for games research.