I'm very late to sharing this news: Deep Learning for RNA design is published in Science and chosen as the cover article!
Now over 3 years ago, gRNAde started as a side-project in my CS PhD because I wanted to do something real beyond benchmarks.
Its been an amazing journey
On the citizen science platform Eterna, #AI methods matched expert human players at inventing sequences that fold into an RNA pseudoknot, a structure in which loops pair with distant parts of the same molecule.
This opens a route to automatically designing RNA therapeutics,
After coming back to Singapore, I’m convinced we have all the ingredients to become a global leader in Embodied AI: world-class talent, a growing industry, strong resources, strategic advantages, and most importantly, the ambition to make it happen.
My group at NUS is looking
I still remember years ago GPT4 was a (pretty poor) baseline in @nc_frey@samuel_stanton_ papers at Prescient - this is now possible with prompting, reasoning about biology and geometry!
Many drugs work by binding to a specific target in the body and blocking or changing what it does. An important first step in the drug development process is designing a molecule that can bind tightly to its target. Traditionally, that's meant weeks or months of expert work per
Feel kind of sad. Obviously it was said that as soon as benchmarks go public the labs train on the test set or RL on them. But this reminds me of school, rote learning for doing well in exams, etc.
But we also took a chance to have a look at some in-the-wild scheming, reward seeking, etc. examples, and dumped it in appendix.
1) Summarizer unfaithfulness
Reasoning summaries often omit important information from the original trace.
Here, Opus 4.8 realizes it knows the
