First blog entry

My writing skills have been a little rusty. Hence, I try my best to maintain of this website as a way to share some reflections of what I'm up to.

After nearly 2 years of the pandemic, it has been super clichéd to even mention how it has impacted our lives. I graduated during the pandemic; I gave a lot of interviews from my study room (which looks eerily like a storage room with empty boxes, books, and papers all over); I got a job, and I have been working in the exact same room during these 2 years. Many changes have happened, and nearly all of them happened online. How cool.

After 6 years of exploring the wonderful world of linguistics, I got to experience many aspects of languages. At my university in Vietnam, I was trained in theoretical linguistics (some morphology, syntax, phonetics, semantics, pragmatics), and applied linguistics (discourse, language teaching). In 2016, I was introduced the machine translation model, and that was my first contact with natural language generation. It was used in computer-assisted translation program. In 2019, the model of neural machine translation was introduced at the company I was working for, and I was so impressed by how potential it can be, and most of the time, how terribly wrong it could go with little corpus input.

When I applied to NUS Linguistics for my graduate studies, I proposed a simple decision tree model in neural machine translation to predict gender and pronouns in Vietnamese using acoustic signals. As I look back, the idea was shallow, but at least, it was what I wanted to do at the time. During my time at NUS, I learnt more about sociolinguistics, phonetics, and language variation. Thanks to the foundational courses, I finally understood more and more these linguistic phenomena, and how to use computational methods and machine learning to tackle them. I spent my time researching more computational sociolinguistics, as well as enrolling myself in some practical courses in NLP and machine learning to finally understand how ML/DL practitioners deal with language data. It was fantastic. I was encouraged to do sentiment analysis and Vietnamese acoustic modeling (something I have always dreamt of doing) with the help of my supervisor at NUS. It was so great. Working with sound data is challenging, but very interesting. That said, it was time I returned to the NLP aspect that deals with text data.

The field is so vast. Even I could not imagine myself writing a research proposal about machine translation, went on to do speech, and now working with text again. During my first 4 months at work, not a single day goes by without me learning a little more about machine learning and natural language processing. It is either reading the recently published research papers, or practicing building some simple models of DL on Udemy. I am not a professional programmer, so I try my best to practice coding every day. I have always been fascinated by the world of programming, so it comes to me quite naturally. After 6 years, I finally got to work with numbers, data, spreadsheets, codes, and languages. I am so excited, and grateful for the opportunity given at work. My teammates also encourage me to code more. They are such pros, and I want to code effectively like them one day soon.