Kindly reveiw my SOP

I am in the final year of my B.Sc. Computer Science at National University of Computer and Emerging Sciences, Pakistan’s Top Tier Computer Science university. I have a CGPA of 3.19 and strong academic background in computer science and mathematics courses and i wish to pursue my studies along with research in computer science.

For the first 4 semesters in university, i focused on learning tools like android, web development and then in the 5th semester i started research. In the beginning, i took many online courses covering the fields of Machine learning, Artificial Intelligence, Deep Learning, Networks and Data Science and then I joined LUMS as a research assistant with Dr. Fareed Zaffar (PhD from Duke) and begin applying the knowledge I had gained, in different research based projects.

PROJECT EXPERIENCE

I worked on a Reverse Image Search Project at LUMS with Dr. Fareed Zafar (phD from Duke). I used a deep convolutional network for object recognition, pre-trained on Imagenet Dataset [2] to extract the features of the image. Since more granular details of image were required so i used the the less refined feature layer, of length 25088, from the last block of architecture of this model, to represent the image. Aside from this i extracted a description of image using state of the art google’s im2txt model [3] pre-trained on COCO dataset [4]. Then i found similarities between between image features using cosine similarity and similarities between descriptions using spaCy library [5]. Finally i found the image similarity by adding the similarity found out by above mentioned two methods.

I also worked on analyzing the impact of transfer players in football leagues with Dr. Fareed Zafar (PhD from Duke). The particular dataset we worked on captures all events happening during a football match on the field including distance travelled by players, their respective coordinates on the football field after every second, fouls committed, red cards given, throw ins, corners e.t.c. We analyzed statistical impact of player in terms of fouls/misconducts on field and found within team and cross team contagion impact of the player using a set of predefined features involving foul events of match. An impact analysis before and after a player is transferred to another team helps in predicting the impact a new player can cause on team’s performance.

For my Undergraduate Honor’s Thesis, i am currently working with Dr. Mubasher Baig (phD from LUMS) on Audio Classification and Adaptive Learning. Currently we are in the Audio Classification phase. We are classifying arabic alphabets’ sounds using CNN. Firstly we tried classifying by feeding the spectrograms of these audios to the CNN but the accuracy was stuck at 88%. Then instead of feeding the spectrograms directly, we extracted the features of spectrogram image using the VGG19 model [1] pre-trained on Imagenet Dataset [2]. This helped our accuracy boost to 95%. To achieve more accuracy we reduced the number of classes i.e 24 to 14 on the basis of similar sounds, and then we classified the audios on 2 levels.

AREAS OF INTEREST

My research experience at LUMS and my undergrad thesis opened me to many interesting fields within Computer Science. The ones that really interest me include Machine Learning and Deep learning. The reason of my fascination is the power that Machine Learning and Deep Learning Algorithms have. One particular class of deep, feed-forward artificial neural networks that i have keen interest in and in which i want to carry on my research are CNNs. I also have this crave for working on IOT and 3D printing because of the tremendous ease they are bringing in a common person’s life.

MY MOTIVATION FOR PHD

During my Research experience at LUMS, I found out that my capability increased tremendously with the toughness of situation i was put in. So my motivation for phD is based on two reasons. Firstly, I know that my research experience has made me ready for the rigorous research environment of a phD program. Secondly, I want to work hard and reach the position from where i can help those who were not presented with the opportunities . I am really excited to work with some of the best researchers in the world and learn from the pioneers of Computer Science.