Megan Peters, Konrad Kording and the people behind neuromatch are building an online summer school for computational neuroscience (July 13 – 31st). If you’re reading this blog – chances are you should be a student in this class or a TA. It’s going to be awesome! 3 straight weeks of computational neuro education. Designed for PhD student or postdoc experimentalists in mind, what sets it apart is matching, the same technology behind neuromatch. Become part of small tight-knit group of trainees with similar interests – bond, share, learn. But wait, there’s more:
- Low TA to student ratio (6-10)
- Lectures with world-class computational neuroscientists
- Tutorials in Python, cloud-based in Google Colab: take your theoretical knowledge and apply to your own data
- Inclusive: 100$ tuition (waivable)
I loved summer school (CSHL computational neuroscience: vision). I made some life-long friends and have memories to last a lifetime. We want it to be better than a real-life summer school. Better by being more inclusive, matching you with more diverse people, offering multiple levels of support, and being accessible to people at different stages in their career.
Teaching at scale
Most real-life elite education is expensive and exclusive. In online education, it’s possible to scale to huge audiences for a fraction of the cost, and let everybody in – including people of disadvantaged backgrounds. MOOCs can have excellent learning materials – better production values, materials that resist the test of time, scannable lectures. Training educators in the latest methods can be done at scale.
The biggest disadvantage of MOOCs is high attrition – up to 95% of people that sign up for MOOCs drop out. New models are appearing that have the accessibility of online learning and the critical social component of real-life learning. At Stanford, Chris Piech and Mehran Sahami started Code in Place, an effort to bring CS education to people at home during covid19 times. It’s a massive effort – 8,000 students from all over the world accompanied by 800 section leaders (including me!). Several factors have kept dropout to a minimum (< 20% after 2 weeks):
- A low TA-to-student ratio
- matching TAs to students based on learner characteristic (age, location)
- engaging materials
- a very high standard for TAs (Top affiliations are Stanford, Google and MIT)
- a tough preparatory assignment to make sure the students entering class were motivated
I was very excited when Konrad announced that they were working on a summer school, fresh off my Code in Place experience, so I contacted him and have joined the team to help (in a small way) organize this thing. There are a handful of experiments in teaching computational neuroscience at scale, the most successful in my opinion being the Neuronal Dynamics online book, tutorials, and lectures. Imagine that, plus TAs, your own wolf pack, live Q&A – community. That’s what we’re aiming for.