Keeping up with data — Week 45 reading list

A curious mind wandering through the world of data

Source: waitbutwhy.com

A friend shared a link to the State of AI Report with me last Sunday. It is one of those reports that hooks you in; you find yourself clicking on the source links, searching new terms, unfamiliar concepts and new thinking.

Cutting to the chase, this week’s reading list is about AI.

1. State of AI Report 2020: Report covering the developments in AI during the last 12 months from all possible angles. 177 slides full of information means that everyone finds something to their interest. My takeaways are: AI development is (very) expensive; NLP is still the hottest area of AI; AI is mostly closed source; corporate-driven academic brain drain is significant; and PyTorch is fast outpacing TensorFlow in research papers. Plus, the report is a fantastic source of fun facts for parties 🤓. (stateof.ai)
2. Artificial Intelligence — The Revolution Hasn’t Happened Yet: Covering the challenge of bringing together computers and humans in ways that enhance human life. An eye-opening article about the opportunities of augmenting human intelligence and creativity and making the human environment more supportive, interesting and safe that goes far beyond a human-imitative AI. A call for a new — human-centric — engineering discipline. Intellectually demanding but rewarding article. (Michael I. Jordan @ Medium)
3. The Next Generation Of Artificial Intelligence: Unsupervised learning independent (or rather less dependent) of laborious data labelling process; federated learning preserving data privacy; transformers for NLP but also computer vision; neural network compressions enabling running algorithms on devices at the edge; generative AI creating content; “system 2” reasoning performing for analytical tasks. Executive overview of what to expect in the next couple of years. (Forbes, part 1 and part 2)
4. The AI Revolution: The Road to Superintelligence: From current artificial narrow intelligence, to artificial general intelligence (understanding why Braveheart was great but The Patriot was terrible), to artificial superintelligence and how it might impact us. Reading this article reminded me of Max Tegmark’s book Life 3.0 that was written two years after this article. Similar synopsis but I find the philosophising easier when accompanied with jokes and comic strips. (Wait but why, part 1 and part 2)

Often disregarded as a buzzword, AI (whether it’s the right term or not) is a topic worth exploring beyond the hype. I, for one, certainly enjoy reading and learning about the technicalities of ML algorithms, engineering challenges or ethical, sociological or economic impacts.