2 replies to this topic

[TheGene]

http://news.mit.edu/...s-research-1209

 

This is a new one. I have seen a few of these machines so far, they try many many many tests with precision and different parameters, and will analyze/sort the data and refine their experiments. This would be a great way to give the machine a narrow goal like freezing and then thawing cells and then seeing if the thawed cells begin to look/behave like the normal cells, because it would find which settings result in closer results to what we want. It would randomly explore parameters, and exploit parameters that work.

 

Is anyone doing this? This sounds like a great way to do millions of experiments accurately with a broad adaptive methodology that we could never keep track of or at least find as quickly.

 

It could test parameters like size of the specimen, age, antifreeze chemicals (that's a BIG thing they test manually at least), humidity, water retention, heat, time to reach a temperature, magnetism, I could go on but the parameters are endless.

Edited by TheGene, 10 December 2019 - 05:43 PM.

[caliban]

I have seen a few of these machines 

 

 

what machines do you mean? The setup in that article is irrelevant to cryo research.

 

If you are asking "do cryobiologists use high-throughput systems, robot-assisted cell culture and machine learning?"  I would suggest the answer is "sometimes (Ref1; Ref2; Ref3) - but not that much".

Why? There isn't much money in the sector, and the types of preparation and analysis required are not that easy to automate.    

[TheGene]

The machines I meant were just some observations; Adam, Eve, the Intelligent Towing Tank, and of course simple partial-automation techniques I'm sure are used more commonly.

 

Good references.

 

You actually bring me back to a point I said - exploit/explore: I agree money is why groups of people make certain decisions. If you made a video game, 'Mario' has a better chance of being sold. New things do too, like a new 'dog toy' you're not bored of yet, but you don't know if it's good, no experience. We need to try new things and take risks and not be stuck on proven ground, and so does the machine we venture to create :)).

Edited by TheGene, 11 December 2019 - 09:16 AM.