chinesh
Sep 30, 2018
Graduate / Applying Umass, UCI, TAMU, Gatech - MS in computer science with Machine Learning as Specialization [3]
Hello, I would be applying to Umass, UCI, TAMU, Gatech for MS in computer science. Could you please help me look for structure, content and grammatical errors and provide me with suggestions for this Statement of Purpose?
I was in my school days sitting with my elder brother watching the Google I/O. When Sundar Pichai introduced various product's advancement and new features in Google home, photos and maps, using Artificial Intelligence and Machine Learning. I was skeptical that any machine could 'learn' in a way that mirrors human learning, which made me curious about how Artificial Intelligence and Machine Learning has bettered people's lives with seemingly impossible changes. I especially enjoyed machine learning theories and their applications in natural language processing and computer vision. Being the member of Microsoft Technical Community member for 4 years I got the opportunity to explore more about ML/AI and to teach the people about the same. There, I learned more about supervised learning, unsupervised learning and reinforcement learning that made the problem-solving statements more interesting for me and deepened my interest in AI. All these experiences convinced me that my interests lie within applicable ideas that have a strong theoretical foundation.
My eagerness to build products and zeal to learn led me towards learning on MOOCs from CMU, Stanford - CNN for Visual Recognition, Natural Language Processing with Deep Learning and completed certification on Deep Learning specialization on Coursera. I started working on machine learning projects focusing on fundamentals. I was surprised after seeing the result of Deep Visual-Semantic Alignments for Generating Image Descriptions while working on Image captioning project which made me more curious for Computer Vision. Continuing with my passion, I spent my undergraduate studies working on projects involving multiple domains such as Computer Vision, Natural Language Processing, Recommender System, Forecasting. Additionally, I participated in various workshops and seminars on Neural Networks organized by Indian Institute of Technology, Indore and National Institute of Technology Rourkela where I received the award for extraordinary performance in the examination among the batch of 70 professionals. Seeing great results in these seemingly disparate domains has made me a strong believer in the limitless possibilities of Artificial Intelligence and Machine Learning.
In order to learn the possibilities of ML and its practical utilities, I spent my summer working as a Machine Learning Intern at Cuisinelinks, where I integrated Amazon Alexa with an external API facilitating online food ordering and reserving the table. I prepared use cases for the client of the company for using ML/AI in their business for forecasting their traffic, developing recommender system for the menu which helped the company to increase the profit by 20%. This internship helped me understand how companies are leveraging machine learning techniques to extract meaningful insights from their data.
In the quest to get some hands-on experience with real-world problems, I was fortunate to get an opportunity to conduct research on self-driving cars at Swaayatt Robots. During this tenure, I recognized an existing problem residing in the current system for navigation of vehicle in an unstructured and unmapped environment. The solution proposed by Google, Waymo and Tesla to this problem, although is accuracy efficient, but incurs high cost and resources(like LiDARs). Analyzing all the use cases, we were trying to come up with a much efficient solution that uses USB instead of LIDAR's that could drastically reduce the cost from $75000 to an unbelievingly smaller amount of $200. The algorithms required lots of hand labeling of data which was very time-consuming and tedious process that requires a lot of human hours. I solved this problem by creating and implementing the pipeline that will automatically label the data using Markov Random fields and CRFs. It post-processes the results of image segmentation from Segnet by reassigning the labels to each pixel thereby increasing the accuracy for pixel level image classification. My ability to challenge myself and explore my limits led me to identify the problem which is to make our car run at night time. I devised a solution for solving the problem of night perception using the Multimodel unsupervised image to image translation using GANs, the model is able to translate night images to days images and then running the car on the same algorithm as for day. This internship helped me in getting exposure to Computer Vision, Simultaneous Localization and Mapping, Artificial Intelligence which can also solve other perception problem in robotics. While working, I had faced the problem while solving the path prediction problem on Indian Highways to avoid accidents. At that time I was not able to do much work as I was diverted to another work but had developed curiosity and interest on how to work on such a challenging real world problem.
Despite recent advances, self driving car still poses problems during actual deployment. Carrying forward my experiences from Swaayatt, I am currently working at Nanyang Technological University, Singapore as Research Assistance to address the real-time deployment issues for detecting Traffic incidents. My main goal is to study the failure case of ABC and build upon EFG so as to achieve HIJ and take a step forwards actualy realization of these ideas, so that the technology becomes accessible to everyone."
I believe that the graduate program in Computer Science with a focus on Machine Learning would provide me with an opportunity to work on some of the real-world problems under proper guidance. I aim to research on designing a system for intelligent summarization of concepts in a way that is sensitive to the context of a specific user. Intuitively, the system should select the right language, but it should also select the right vocabulary and concepts and should emphasize the points which are most likely to be salient for that particular user. Another interesting problem in the field of Computer Vision which I came across while dealing with blur, noisy and dull images that result in either deletion of those images or correcting them. Working in XYZ lab my aim is to build more robust algorithms for such cases to serve police, medical sector and contribute to the field of Computer Vision. Though deep learning methods have impressive scalability, the data-driven model still needs improvement when it deals with uncertainty, unobservable information, and dynamic systems. I believe the field of Machine Learning apart from being my focus point, satisfies my ambitions perfectly since I always try to identify some real-world problems and discover means to solve them. I am excited to be a part of lab and work under xyz who specializes in ......
The thrill of discovery, connecting the dots and building innovative products that millions of users can use, is what excites me. I'm very excited to attend the NAME in the midst of a revolution in AI and ML, and I hope that I will be seen as a good fit for the MS in Computer Science, where I can extend upon my undergraduate knowledge in AI, Computer Vision, Natural Language Processing and years of practical data science experience, which will help me later build useful and pervasive products and become a contributing participant in the revolution.
Hello, I would be applying to Umass, UCI, TAMU, Gatech for MS in computer science. Could you please help me look for structure, content and grammatical errors and provide me with suggestions for this Statement of Purpose?
MS in Computer science
I was in my school days sitting with my elder brother watching the Google I/O. When Sundar Pichai introduced various product's advancement and new features in Google home, photos and maps, using Artificial Intelligence and Machine Learning. I was skeptical that any machine could 'learn' in a way that mirrors human learning, which made me curious about how Artificial Intelligence and Machine Learning has bettered people's lives with seemingly impossible changes. I especially enjoyed machine learning theories and their applications in natural language processing and computer vision. Being the member of Microsoft Technical Community member for 4 years I got the opportunity to explore more about ML/AI and to teach the people about the same. There, I learned more about supervised learning, unsupervised learning and reinforcement learning that made the problem-solving statements more interesting for me and deepened my interest in AI. All these experiences convinced me that my interests lie within applicable ideas that have a strong theoretical foundation.
My eagerness to build products and zeal to learn led me towards learning on MOOCs from CMU, Stanford - CNN for Visual Recognition, Natural Language Processing with Deep Learning and completed certification on Deep Learning specialization on Coursera. I started working on machine learning projects focusing on fundamentals. I was surprised after seeing the result of Deep Visual-Semantic Alignments for Generating Image Descriptions while working on Image captioning project which made me more curious for Computer Vision. Continuing with my passion, I spent my undergraduate studies working on projects involving multiple domains such as Computer Vision, Natural Language Processing, Recommender System, Forecasting. Additionally, I participated in various workshops and seminars on Neural Networks organized by Indian Institute of Technology, Indore and National Institute of Technology Rourkela where I received the award for extraordinary performance in the examination among the batch of 70 professionals. Seeing great results in these seemingly disparate domains has made me a strong believer in the limitless possibilities of Artificial Intelligence and Machine Learning.
In order to learn the possibilities of ML and its practical utilities, I spent my summer working as a Machine Learning Intern at Cuisinelinks, where I integrated Amazon Alexa with an external API facilitating online food ordering and reserving the table. I prepared use cases for the client of the company for using ML/AI in their business for forecasting their traffic, developing recommender system for the menu which helped the company to increase the profit by 20%. This internship helped me understand how companies are leveraging machine learning techniques to extract meaningful insights from their data.
In the quest to get some hands-on experience with real-world problems, I was fortunate to get an opportunity to conduct research on self-driving cars at Swaayatt Robots. During this tenure, I recognized an existing problem residing in the current system for navigation of vehicle in an unstructured and unmapped environment. The solution proposed by Google, Waymo and Tesla to this problem, although is accuracy efficient, but incurs high cost and resources(like LiDARs). Analyzing all the use cases, we were trying to come up with a much efficient solution that uses USB instead of LIDAR's that could drastically reduce the cost from $75000 to an unbelievingly smaller amount of $200. The algorithms required lots of hand labeling of data which was very time-consuming and tedious process that requires a lot of human hours. I solved this problem by creating and implementing the pipeline that will automatically label the data using Markov Random fields and CRFs. It post-processes the results of image segmentation from Segnet by reassigning the labels to each pixel thereby increasing the accuracy for pixel level image classification. My ability to challenge myself and explore my limits led me to identify the problem which is to make our car run at night time. I devised a solution for solving the problem of night perception using the Multimodel unsupervised image to image translation using GANs, the model is able to translate night images to days images and then running the car on the same algorithm as for day. This internship helped me in getting exposure to Computer Vision, Simultaneous Localization and Mapping, Artificial Intelligence which can also solve other perception problem in robotics. While working, I had faced the problem while solving the path prediction problem on Indian Highways to avoid accidents. At that time I was not able to do much work as I was diverted to another work but had developed curiosity and interest on how to work on such a challenging real world problem.
Despite recent advances, self driving car still poses problems during actual deployment. Carrying forward my experiences from Swaayatt, I am currently working at Nanyang Technological University, Singapore as Research Assistance to address the real-time deployment issues for detecting Traffic incidents. My main goal is to study the failure case of ABC and build upon EFG so as to achieve HIJ and take a step forwards actualy realization of these ideas, so that the technology becomes accessible to everyone."
I believe that the graduate program in Computer Science with a focus on Machine Learning would provide me with an opportunity to work on some of the real-world problems under proper guidance. I aim to research on designing a system for intelligent summarization of concepts in a way that is sensitive to the context of a specific user. Intuitively, the system should select the right language, but it should also select the right vocabulary and concepts and should emphasize the points which are most likely to be salient for that particular user. Another interesting problem in the field of Computer Vision which I came across while dealing with blur, noisy and dull images that result in either deletion of those images or correcting them. Working in XYZ lab my aim is to build more robust algorithms for such cases to serve police, medical sector and contribute to the field of Computer Vision. Though deep learning methods have impressive scalability, the data-driven model still needs improvement when it deals with uncertainty, unobservable information, and dynamic systems. I believe the field of Machine Learning apart from being my focus point, satisfies my ambitions perfectly since I always try to identify some real-world problems and discover means to solve them. I am excited to be a part of lab and work under xyz who specializes in ......
The thrill of discovery, connecting the dots and building innovative products that millions of users can use, is what excites me. I'm very excited to attend the NAME in the midst of a revolution in AI and ML, and I hope that I will be seen as a good fit for the MS in Computer Science, where I can extend upon my undergraduate knowledge in AI, Computer Vision, Natural Language Processing and years of practical data science experience, which will help me later build useful and pervasive products and become a contributing participant in the revolution.