We realized that for many students, this was motivated by post-PhD professional considerations. It's easier to get hired into a tech or data science job if you have a "computational certificate" on your CV. Many of the students weren't particularly interested in models. - ThreadSky

gershbrain.bsky.social • 107 days ago

We realized that for many students, this was motivated by post-PhD professional considerations. It's easier to get hired into a tech or data science job if you have a "computational certificate" on your CV. Many of the students weren't particularly interested in models.

Comments

gershbrain.bsky.social•107 days ago

I wish that I was able to convince the students in my class that the ultimate goal of science isn't accumulation of facts but the compression of facts into theories. Mel summarizes it well here:
https://www.nature.com/articles/nn1100_1183.pdf
Unfortunately, I think this is a minority viewpoint within neuroscience.

steveread.bsky.social•107 days ago

📌

neuralreckoning.bsky.social•106 days ago

I know there are character limits but both views are too limited surely? What about the goal of making predictions? Or of answering counterfactual questions (eg could evolution have found a different solution to xyz)?

gershbrain.bsky.social•107 days ago

I hope there comes a day when there is no separation between theorists and experimentalists. I don't see how we will have good experiments unless the experimentalists take on modeling. And theorists should want test their theories experimentally.

neuralreckoning.bsky.social•106 days ago

I'm not sure I agree. I think it's important that everyone understand at least something of the full pipeline, but that doesn't necessarily mean doing both. For example, there's definitely a place for pure theory that answers questions like "does X imply Y" to support this sort of reasoning.

ruthrosenholtz.bsky.social•107 days ago

I’ve sometimes thought we need a program for this at one of the funding agencies. Collaborative proposals between experimentalists and modelers, or a K-whatever modeling training for experimentalists.

neuroai.bsky.social•107 days ago

katejj.bsky.social•106 days ago

From my perspective as an experimentalist, models are useful but only somewhat. Experimenters are trying to understand mechanisms, and depending on the system, an overarching theoretical model may or may not be useful. We didn't need a model of the kidney to start picking it apart /

katejj.bsky.social•106 days ago

...but then once we could see what was in it and started trying to understand what it does it turns out multiple different models are useful as it does several things. The brain is the same I'm sure

katejj.bsky.social•106 days ago

I use models all the time to help me understand what experimental question to ask, and I assume all experimenters do, but I use them in a simplified way because there are many more models than experiments one can do, and many seem somewhat implausible to those of us in the trenches

aribenjamin.bsky.social•106 days ago

💯 the gap is far larger than it should be.

I suppose this is why I'm doing a postdoc in an experimental lab. But PIs need to change as well, not just postdocs. I wish there were more of a culture of scientific residencies, at all points in one's career.

sterrett-sc.bsky.social•103 days ago

I really like this title 'computer experimentalist' he uses to describe the role of simulations!

Comments

Posting Rules

Reply