Please Proofread my CS Grad School SOP

[Mochachino_60]

Please also point out any flaws in Technical Details if you see any. Thanks a lot.

Statement of Purpose

The scientific approach has always been a fundamental facet of my outlook towards life, belief system and academic pursuits. Driven by my scientific curiosity, I have arrived at a juncture where I am confident that a graduate research course with an initial M.Eng. Thesis and leading to a PhD is the most suitable step for me towards a career in research that I aspire to have.  

My principle interest is to conduct research in the broad area of Machine Learning (ML) and Signal Processing including their application in fields ranging from Robotics, Auditory Perception and Computer Vision. I am particularly fascinated by the application of Reinforcement Learning in Robotics, navigation and object recognition and identification. My goal is to learn and understand the decision making process of present day artificial agents under uncertainty and figuring out ways to improve them for novel applications in future. Thus one of my research interest is on tasks that can be framed as Markov Decision Processes in general and solved by Reinforcement Learning and dynamic programming. Furthermore, having an interest in Neuroscience and Behavioral Psychology, I am quite intrigued by how concepts of learning and intelligence from these fields can help make artificial intelligent systems more capable of human like decision making. 

I became interested to pursue a graduate degree at the School of Computer Science at XYZ University due to the world-class research going on here in the fields of my interest. At XYZ University, the works of Dr. XY are well suited to my research experiences and goals. I was particularly fascinated by his Lab's project "WWWW” as it utilizes RL techniques to make robots adaptive to changing ocean conditions without directly engineered controllers.  I would also like to work on similar projects like Dr. XY’s work, “WWWW”, as it directly aligns with my research interest of working in the field of computer vision and scene understanding. His upcoming project utilizing Reinforcement Learning to identify objects based on how they sound to interact with has attracted me to apply to the program due to my past experience of working with audio classification and future goal of working with Reinforcement Learning. Moreover, the work performed in the Reasoning and Learning Lab is also well aligned with my research interests.

My interest in Machine Learning research developed in my junior year while participating in IEEE Signal Processing Cup 2016. Working in this global competition helped me gain enormous experience and insight into the world of research. To participate in this event, I put together a team of likeminded peers. With Machine Learning being a relatively fringe topic in our university without many experts, we resorted to independent study. As I was the principle programmer and signal processing researcher in the team, I conducted literature survey to learn, understand and implement the necessary algorithm and make improvements to the classification system with the help of another team mate; while the rest of the team worked on hardware implementation and data gathering, all within a narrow window of three months. In this event, we classified audio recordings according to the location those recordings were made, exploiting the unique Electrical Network Frequency (ENF) embedded in those recordings. We extracted a wide array of statistical features ranging from mean and variance to more subtle ones, such as those derived from Multilevel Wavelet Decomposition and Auto-regressive models, for the implementation of a Supervised Multiclass Support Vector Machine (SVM). 

During this work, while implementing a classification system based on high dimensional features, I became introduced to concepts and techniques such as Principal Component Analysis (PCA), Independent Component Analysis (ICA) and Fisher’s Linear Discriminant Analysis (LDA). Moreover, I became familiarized with Bayesian Inference, Gaussian mixture models (GMM), and weighted SVM as well among other concepts from statistics and machine learning. This work taught me how to develop an end-user platform from scratch by taking a concept from existing literature and improving it. Eventually, I managed to lead the team to 11th position among 52 teams worldwide. An extension of this work has been accepted for publication in the peer-reviewed conference proceedings of IEEE R10HTC 2017. This success coupled with the theoretical understanding acquired through great performance in coursework on Data Structure and Algorithms, Random Signals and Processes, Discreet Mathematics, Numerical Analysis, Signals and Systems, and Control System Engineering convinced me that a career as a researcher in the field of Machine Learning or its application is the one I should pursue.

I was accepted into two of Bangladesh’s premier universities, Bangladesh University of Engineering and Technology (BUET) and Dhaka University, ranking above 95 percent of the test takers in their entrance exams. At my undergraduate university, Islamic University of Technology (IUT), I received the prestigious OIC scholarship, which provided a tuition waiver for three years of my stay and included a generous monthly allowance. In my senior year at IUT, I designed and instructed a MATLAB based signal processing course for IEEE IUT Student Branch, which was attended by over a 100 students from various levels of my university. 

Even though I struggled in the initial semesters at IUT due to family problems as my mother passed away from cancer, I showed grit and determination by consistently ranking around the top of my class in the core courses of the last five semesters. At IUT, I worked on many projects ranging from voice pattern recognition system to hardware and software implementation of Microcontroller based temperature sensors for electric motor control. These experiences helped me develop my programming and logical skills. Concurrently, I engrossed myself into machine learning and mathematics by learning from massive open online courses: Machine Learning (Stanford Online, Coursera), Introduction to AI (Udacity), and independent study of books such as Linear Algebra by Gilbert Strang, Reinforcement Learning: An Introduction by Sutton and Barto, and Calculus by Michael Spivak. I also kept following online streams of International Conference on Machine Learning (ICML) and Neural Information Processing Systems (NIPS) to keep in touch with the latest developments in the general field of Machine Learning. In my senior year, I worked in IUT’s nanophotonics group for my undergraduate thesis on computational nanophotonics. We computationally investigated the effects of the near-field enhancements of metal nanoparticles in GaAs, CdTe and Perovskite absorber layers and worked on finding efficient and cost effective materials. The resultant work was published in the proceedings of 2016 IEEE APACE conference.

I believe I have the research exposure, determination, and academic preparation to be a part of the world-class graduate research community of XYZ University. I am confident that XYZ University is the best place for me to have a career in the forefront of cutting edge research that I intend to have.