On Tuesday 12th December 2023, I had the opportunity to attend the AI2023 conference held in Cambridge. During those 3 days, I participated in a series of workshops led by AI practitioners, listen to keynotes delivered by professors, ask questions during panel discussions, and watch presentations of posters and refereed papers covering technical aspects and applications of Artificial Intelligence. However, for myself, the highlight of this conference was a presentation of my paper written jointly with Dr. Ken McGarry from the University of Sunderland. You can download the paper and keynote in the links below.
- Paper: View the Paper on Springer website
- Keynote: Download AI2023 Presentation (1MB)
I want to say thank you to Dr. Ken McGarry, who encouraged me to transform my master's dissertation into conference paper, and for his invaluable support, guidance and insights.
In this article:
How it all started
In January 2023, I completed my master's in Computer Science with Data Science, achieving distinction in a subject that I knew nothing about just 2 years earlier. This great personal achievement proves that anything is possible, as long as you put your head into it, commit the time, and keep working to achieve your goal. However, this journey is not for everyone. Studying part-time alongside full-time work and other life commitments is a challenge. Studying for 2 years in the evenings, and weekends, and squeezing any free time, just to keep up with learning material, assignments, and the research project.
I decided to study Data Science because according to Harvard Business Review, this discipline was the "sexiest job of the 21st century," and with a promise of making a lot of money and a little encouragement, I decided to jump on this bandwagon. However, money wasn't the only enticement. I was also inspired by two books: Ten Arguments for Deleting Your Social Media Accounts Right Now by Jaron Lanier, and Homo Deus: A Brief History of Tomorrow by Yuval Noah Harrari, who coined an interesting concept of Dataism.
Since my options were constrained by the availability of time and money, I decided to enroll in Computer Science with Data Science online course offered by the University of Sunderland, which was both convenient and affordable. The course helped me to learn about the principles behind the discipline, and also explore technical aspects and applications of Data Analytics, Big Data, and Machine Learning. This is also where I was assigned a project supervisor, who after marking my dissertation reached out with an offer to get my project reworked, and submit it to a conference.
In my research project, I applied a quantitative approach of geospatial sentiment analysis to reveal sentiment and evaluate machine learning methods using tweets regarding the public opinion about the war in Ukraine posted on Twitter. I was also considering a research project in astronomy, for example, I was interested in applications of computer vision in the detection of objects in the photographs taken by space telescopes, which would nicely coincide with the launch of the James Webb Space Telescope. Eventually, I decided to drop that idea and carry on with geospatial sentiment analysis primarily because I have prior experience working with Twitter API, I enjoy creating and reading maps, and also the war in Ukraine provided a unique opportunity to study the dynamics of the public opinion in cyberspace regarding war-related information.
Writing the paper
I was excited by the prospect of publishing my work in a journal. We decided to apply for academic access to Twitter, which according to their documentation, would give us access to historical data, which would enable us to access tweets posted on Twitter since the Russian invasion in February 2022. However, we never heard anything back from Twitter and we had to work with 2,893 tweets collected through extended access that I received in 2021. The key takeaway from writing your paper is that it's not as simple as writing a Word document. Everything has to be written in .tex format, which to uninitiated eyes, can be challenging.
Conference Day 1
During the first morning of the AI2023 conference, we were greeted by the organisers and given an opportunity to meet with other delegates who arrived to the conference mainly from the UK institutions but also some from the European institutions in Germany and Slovakia. The first day of the conference featured a series of workshops running in parallel streams. The workshops encompassed areas such as development of scientific models and their application in social sciences, the rise of the Industry 4.0, and the role of AI in manufacturing and supply chains, as well as AI challenges in healthcare, and exploration of challenges of testing AI, esp. in adaptive system and safety critical environments.
Conference Day 2
The second day of the AI2023 conference kicked-off with the first keynote lecture, in which Prof. Robert Stevens from the Manchester University, demonstrated "how Machine Learning may benefit from knowledge representation and symbollic Artificial Intelligence to help make the vast resource of what we know about biology more accessible" (Stevens, 2023). The lecture was followed by the presentation of the best refereed technical paper that provided insights on explainations for hybrid artificial intelligence. On this day, I also presented my own paper, that you can download on Springer website. I was the first person to present a refereed paper during the Machine Learning applications session. The rest of day followed by more presentations of papers and posters, finishing with the first panel discussion about "how is AI going to change society in the coming decade? How do we ensure that change is for the good?"
I wasn't impressed with the answers provided by the panel of experts, however, the audience raised interesting questions, some arguing that AI will indeed change society, and there is no way out of this accelerating change. Some brought up the analogy of the trolley problem, others praised AI for the multidute of opportunities it will provide. However, I was concerned about negative impact of deep fakes on the society, and I argued that since we teach AI using the content generated by people, we need to take responsibility for what we post online. Not many people could see my point, though. The day ended with pre-drinks and gala dinner at the dinning hall at Peterhouse College.
Conference Day 3
More content to come.
