ไม่สามารถเล่นวิดีโอนี้
ขออภัยในความไม่สะดวก
HOW AI is changing Science | DeepMind’s AlphaFold 3 Explained
ฝัง
- เผยแพร่เมื่อ 24 พ.ค. 2024
- In 2021, researchers made a ground-breaking discovery using AI’s deep learning algorithms. They were able to predict the 3D structure of almost all proteins known to humans from just their amino acid building blocks. This was an incredible achievement because proteins are fundamental to our very existence.
In today’s video, we will explore AlphaFold, the AI model developed by Google DeepMind and Isomorphic Labs. We will discuss how it works, why its latest version, AlphaFold 3, released just a few weeks ago, is creating such excitement, and what this might mean for us.
If you’re new here, my name is Roberta. I am a medical doctor and science writer. If you are a returning viewer, thank you so much for coming back!
Resources:
nature.com/art...
www.tandfonlin...
archive.org/de...
web.archive.or...
www.rcsb.org/
blog.google/te...
www.bloomberg....
• Google DeepMind CEO on...
deepmind.googl...
golgi.sandbox....
• AlphaFold: The making ...
deepmind.googl...
deepmind.googl...
www.forbes.com...
www.theguardia...
www.quantamaga...
Some of my previous videos:
The science of sleep and wellbeing:
• The Science of Sleep &...
Gut microbiome and health
• How to change your gut...
How to maximise the benefits of reading aloud
• How to maximise the be...
Your Brain on Travel
• Your Brain on Travel
Why is hay fever getting worse and how to better cope with it
• Why is hay fever getti...
How to be happy according to science:
• How to be happy accord...
3 daily habits for a healthier life
• 3 daily habits for a h...
Ultra-processed foods
• How ultra-processed fo...
A bit about me:
• From Medicine to Youtu...
Thank you for stopping by, consider subscribing and I will be seeing you in the next one!
DeepMind's work on AlphaFold and GNoME (a model used to discover hundreds of thousands of stable materials) are what make me excited about the future of AI. It's upsetting that the sorts of models which are going to receive the most funding, and attention, are models which can be turned into a product. However, it's these science oriented models that, to me, demonstrate the true potential of AI to help humanity advance. Great video, thank you
Very clear, to the point explanation! Thank you for sharing this video! Absolutely love your content, keep up the good work!
Thank you so much!!
Your videos deserve way more attention than they're getting.
This is fantastic stuff!
Thank you so much for your support!!
I’m not surprised that we made such quick progress with the implementation of AI in this filed. I think the overall problem will remain with implementation and accessibility. The healthcare systems are already overwhelmed as they are.
AI could relieve a great deal of the burden on the healthcare systems, and not just drug discovery programs. GPT-4 is already capable of correctly diagnosing many conditions and prescribing the correct treatment. That's a big deal in communities that don't have enough doctors.
This video is so well put together. I hope people find it fast. Please keep up the good work!
Thank you so much!!
If AI could be fed data of proteins and conditions needed to form protein crystals (to enable X-ray crystallography), I wonder if AI could predict what conditions are necessary to form protein crystal, to enable confirmation of protein structure through X-ray crystallography.
Swell robotics everywhere, Ai jobloss is the only thing I worry about anymore. Anyone else feel the same?
The phenomenon of AI hallucinations demonstrates that vastly improved performance of large language models has gone nowhere towards solving the honesty versus creativity conundrum that we also have as humans. That’s my take.