Research Areas

Biography

My name is Than Lwin Aung. I am a full-stack software engineer and independent AI researcher, without any affliations. I graduated from Elizabethtown College, PA, USA with Computer Engineering Degree. Later, I attended Keller School of Management to pursue Masters in Management (Project Management). However, I dropped out after attending for one year. After that, I worked as a Software Engineer in S2T in Singapore. In S2T, I had a chance to work with Lexalytics to visualize and analyze Sentiment Analysis of public opinions on public web forums and portals. Actually, working with lexalytics sparked my interest in Natural Language Processing, and I co-founded 2KHA with my friends. 2KHA was aimed to be a search portal, which includes text and image search functions, along with E-Commerce Capabilities. However, due to many circumstances, after 2019, the project was scrapped, and I have been only focused on doing independent research on English and Myanmar (Bilingual) NLP, since then.

Expertise

Actually, although I am, by profession, a full-stack software developer and AI Engineer, I have a chance to work on variety of Independent Projects. I had some experience with Labview Machine Vision. In addition, I also have some experience with 3D Game Development and Computer Graphics.

Also, I have some expertise on Embedded Systems and Communcation Protocols, such as TCP/IP, Websocket, NFC and Bluetooth.

As I am working on Bilingual NLP (English and Myanmar) with Machine Learning, I am also a bit familar with many AI Frameworks and Tools.

Special Interest

My Special Interest in AI Research is actually NLP (Natural Language Processing) and Sentiment Analysis. However, classification of sentiment only into "Postive, Neutral, Negative, Ambiguous" is not quite satisfactory in my humble opinion. In reality, emotional perceptions and responses of humans, or even animals, are more diversed and more complex. Therefore, I would like to do more research on different emotional aspects in NLP as well as in other areas of AI Development. Emotions are not just greys; they are, in fact, fifty shades of greys. In my humble opinion, sentiments (emotions) should be a distribution (like a color spectrum) in Vector Space, instead of point-wise classifcation. If "Emotions" could be represented as a Vector, what could be the major basis for that Vector Space?

Language Skills

Actually, for a guy who has claimed to have a strong interest in Computational Languages and Lingustics, embarrasingly, I am only able to speak fluently in only 2 languages: Myanmar (Native Level) and English (Professional Level). However, I am learning my third language, Korean (한글), and I can definitely now say, "안녕하세요"!.