YES; and if you have not heard it yet, the release of ChatGPT 3.5 has pushed the AI/ML revolution into high gear and it has arrived for real in research and education… To take stock of what is going on Harvard and MIT banded together in a two day workshop, locations fairly shared between the two big universities in Cambridge.
There were many interesting presentations which should be available in youtube soon as well as ion the indico web pages. Among others the presentation of A2rchi, our own little open source end to end framework to facilitate the use of large language models. One of them which particularly caught my attention was the presentation of ‘squirrel AI’ where a personalized tutor is used in the K12 school environment to enhance the learning with personalized content presented to each student individually to focus the teaching on the areas which were least understood by the students and thus ‘force-replicating’ the teacher which can in a common class room only address the full class of students, each with different strength and weaknesses in the material. The not surprising effect is that this approach much improves the efficiency of the teaching. Also promising results were shown the day before comparing AI-supported teaching compared to the active learning approach in introductory physics by Greg Kestin and Kelly Miller.
During the dinner on Thursday, the organizing committee and a number of speakers had a long and fruitful discussion about the future and what we should do about the opportunities and challenges in the integration of AI/ML into education and research. The unanimous, maybe slightly biased, sentiment was that the revolution has started and that we have to work together and make sure to be part of it. We all need to discuss and work together to determine how to best integrate this new tool into our education, administration and research at the universities. Therefore, we concluded that a yearly workshop like this one was important. To be able to keep in step with the fast paced progress in AI/ML and at the same time to be more inclusive it was decided to call for the next workshop in the summer, possible expand it in terms of duration, the contributing universities and the organization (introducing parallel sessions to address specific topics more deeply). More on the topic by the co-organizer of the workshop in this post.