2024 Reflections

2024 will always be an exceptional and unforgettable year for me. It stands out as one of the most unique namely because we spent nearly the entirety of 2024 living in Edinburgh, Scotland. It would be impossible to compress our year into a post of the highlights, but here we go.

About a year ago, I was flying back to Scotland to start my second semester at the University of Edinburgh. The second semester was a smooth continuation of the first great semester and continued to be a great experience. I learned so much about machine learning, linguistics, and speech. Another thing about 2024 was that I learned about academia in a deeper and meaningful way. Sometime in the second semester, I started to seriously consider a PhD at KTH in Stockholm this year. The experiences and people that I met in my program, my professors, everyone I met in the CSTR group at Edinburgh, and just the very interesting topics I was learning about got me quite interested in pursuing more education. 

During the semester, I took several classes, namely, Machine Translation, ASR, Speech Synthesis, and Second Language Acquisition. It was in this schedule that I really fell in love with speech synthesis. I wrote about that problem and how it is interesting to me here. I also have a post about my unit selection voice project here.

As part of my interests, I started working on my personal project regarding reading literature in a foreign language (GitHub repo). My entrepreneurial interests were matched by several of my classmates. We would sometimes meet as a small group and talk about ideas. So that has been a fun part of the year. Through this project, I have learned a lot about second language learning and also Flutter development. I hope to, one day, take it somewhere and help people leverage literature in their language learning journey.

The only part of the year that could possibly be construed as unfortunate is that my summer was spent mostly on campus and indoors. My dissertation project kept me very busy, not to mention Victoria's dissertation that also she occasionally asked for help on. My dissertation didn't get a blog post, but it did make it onto my GitHub here. 

The summer was filled with fun though. I was able to participate in several class parties. One of which was hosted by Victoria and I on our flat on Advocates Close. It was to celebrate the American Independence Day. We had a BBQ and a long night of fun, mostly hanging out and chatting with friends. In the summer, I also bought a golf bag and clubs from Facebook Marketplace and started golfing with my friend Jacob. I eventually bought a golf membership pass to play on the Edinburgh Leisure courses. 2024 can be identified in my life's timeline as the year I started golfing. It is a very fun and relaxing activity. My best round was at Carrick Knowe golf course and I broke 100 for 18.

I made a lot of friends in 2024, from school, from the community, and from church. I took on a bit of responsibility at church as well, I was called as a councilor in the Elder's Quorum in January. This was a great opportunity to interact with people in a more meaningful way and to get to know people in a different way. I am very grateful for the chance I had to work with the Elders. It taught me a lot about myself and prioritizing things in my life. I also hope that as I transition back to the States that I can continue in church service and also get better at magnifying my calling.

We had many wonderful visitors throughout the year, including, everyone from the Hansen side, and my parents. We were able to see so much of Scotland because of our visitors. We visited, several times, the Isle of Skye. We also have been to St. Andrews, Glasgow, Stirling, Inverness, Pitlochry, North Berwick, and many more places. Our last visitors, my parents, came to celebrate graduation day. It was so nice to have family make the effort to come visit and give us a good excuse to take trips.

The time between dissertation and graduation was full of applications and golf and stress. Unfortunately, 2024 was also marked with a stagnated tech market, so finding work has been difficult. Despite this, I have been able to do contract work with two companies, Flashlight Learning and Literal. Both in the EdTech space, and I am doing random things/AI work for both companies. It has been a blessing to pay bills. I have also learned a lot. I have been doing a lot of python code, data processing, prompt engineering, and speech synthesis (including voice conversion) for Flashlight. And implementing a py package called booknlp for Literal. I hope to continue working with both companies into 2025.

Reflections:

I enjoy tutoring and pushing myself intellectually. I hope to continue doing both things.

Scotland is a special place, and will always have a bit of my heart.

Take the leap of faith. It is worth it!

Learnings:

Where to begin. Seq2Seq, acoustic v language modeling, phonemes v grapheme v all the emes, and accents are hard to define.

Pytorch!

Kilts are extremely comfortable and manly and cool.

Goals:

2023: Graduate✅/get a job⏳/start a company⏳

2023: Read a few books for fun and a few for personal development✅ I read Atomic Habits, the first 4.5 books of the Red Rising series, and lots of academic articles. 

2024: I would like to launch a Beta version of my language learning application. And start an officially licensed company.

2024: I want to read more. I hate putting a number on it, but If I read 8 books, I think I would be happy.

2024: I hope to play some rounds of golf and make a friend in Logan who will go with me.

2024: I want to make something nice of wood.

Random Facts:

I visited a few new countries this year, continuing my streak. I visited England to visit both London and Preston. I also visited Northern Ireland and Ireland on a trip.

I was lucky enough to get to return to both Sweden and Norway this year. 

Love is Blind is kinda trash, but its so hard to look away.

Haggis and Blood Pudding get a bad wrap. I think they are great!

Comments

Post a Comment

Popular posts from this blog

Phonetic Posteriorgrams

Image
For my MSc dissertation, I am trying to synthesize speech with variable L2 accent strength. There are a few ways to approach this. The "fastest" and "easiest" way is to have the perfect dataset that has reliable labels for accent strength on each utterance. There are two problems with this approach. First, this dataset would be incredibly difficult and expensive to create. Second, it is difficult to get a trusted label for "accent strength" that is also consistent across the dataset. What does accent strength even mean and how would it be measured non-subjectively? I was talking about this problem with one of the PhD candidates in the CSTR group, and she told me to look into phonetic posteriorgrams. Since then, I have tried to read a bit about them. I wanted to explain them in this post.  What is a Phonetic Posteriorgram? A phonetic posteriorgram (PPG) is well described in it's name. Phonetic indicates that we are talking about phonemes, or speech cate...
Read more

The Hard Part About Speech

Image
The rise of Large Language Models (LLM) in recent years has certainly made a lot of people see the world differently. I believe that many people are more interested in statistics, machine learning, computer science, and linguistics because of these fantastic advancements, myself included. As I have studied intensely at the University of Edinburgh over the last year, I have realized more about what these kinds of models do, and how they relate to the bigger picture of modeling functions in general. All of these models are doing some form of function approximation/prediction task. Language models like GPT or BERT are simply predicting the next word token given the context, Automatic Speech Recognizers (ASR) like Whisper are predicting the text output given a waveform, Text to Speech (TTS) models like Amazon Polly are trying to predict the spectrogram or waveform given the textual input and who it should sound like. There is always a task, a dependency, data, and a metric to measure that ...
Read more

Language Identification Project

Image
I wanted to complete some practice problems and build a small basic portfolio of small tech projects that showcase some of my recent learning at the University of Edinburgh. I have taken several courses about natural language processing (NLP) at uni.  NLP is a very wide term for a variety of different tasks, for example, text classification, named entity recognition, machine translation, language modeling, etc. This project was just a simple text classification task. I wanted to input text and have a model predict which language it is. Find the code on my  Github , and my steps that I took to create the project below. Hope this helps facilitate your learning too. Step 1: Find data Foundational to any project involving any sort of prediction. You need data and preferably a lot of it. I wanted to predict language, and one corpus of language data that I already knew about from school was the Europarl Dataset . This is a corpus of EU meetings that are translated into the languages...
Read more