2 Comments
User's avatar
Rainbow Roxy's avatar

Brilliant. I'm wondering how the platform infers those missing cause-and-effect relationships from such incomplete data? That's quite an algoritmhic challenge to solve.

Expand full comment
Natasha Kilroy's avatar

Good question Roxy! I reached out to the author Abhik for his thoughts and this is what he had to say: “I agree, it’s a fascinating challenge. You're right to point out the difficulty.

​My understanding is that the AI approach is a powerful tool for navigating the existing, known data, but it's not quite at the stage of creating new, validated knowledge from incomplete information.

​The core issue is twofold:

​The AI needs more development to mature beyond its current capabilities.

​The underlying data requires massive curation. The current datasets are often too sparse or noisy to reliably support inferring these complex biological relationships.

​So, while it's an impressive start, it also clearly maps out how much work is still needed on the data-gathering ( genomics and proteomics and structural modeling like MD )and curation front.”

I hope this answers your question! Let me know if you’re wondering about anything else :)

Expand full comment