artemmoskalev.bsky.social
Re-imagining drug discovery with AI 🧬. Deep Learning ⚭ Geometry. Previously PhD at the University of Amsterdam. https://amoskalev.github.io/
23 posts
1,130 followers
513 following
Regular Contributor
Conversation Starter
comment in response to
post
1. Hyperbolic NNs for sequence modeling: jobs.jnj.com/en/jobs/2506234550w
2. Causal Inference and Bayesian Optimization: jobs.jnj.com/en/jobs/2506234553w
3. Multi-modal Sequence, Structure & Interaction modeling: jobs.jnj.com/en/jobs/2506234539w
Apply and reach out to me if interested! 😁
comment in response to
post
If you need to pick a neural network and train it on RNA data, our work provides guidelines on which method works best in which conditions.
comment in response to
post
What did we learn? In the presence of severe noise simple sequence transformer without any geometry works the best but it requires much more data to converge! At the same time, 3D Geometric GNNs are the most vulnerable to geometric noise.
comment in response to
post
We study different types of neural networks on various types of RNA representations: 1D vs 2D vs 3D. We evaluate property prediction performance, noise robustness, data efficiency and OOD noise generalization.
comment in response to
post
HELM: Hierarchical Encoding for mRNA Language Modeling is simple to use and only requires changing a loss function!
Joint work with brilliant Mangal Prakash and our amazing interns Mehdi Yazdani-Jahromi and Junjue Xu.
5/5
comment in response to
post
Interesting observation: the improvement from HELM is most prominent on datasets with codon-usage bias! The entropy of codon distribution negatively correlates with relative improvement from learning with hierarchy.
4/5
comment in response to
post
HELM integrates codon-level hierarchy through hierarchical cross entropy, resulting 8% improvement in mRNA property prediction across various types and species. We can also generate more diverse sequences while maintaining biological plausibility.
3/5
comment in response to
post
Why is the hierarchy a key? mRNA sequences have a codon-level hierarchy, where synonymous codons encode the same amino acid but impact expression, stability, and therapeutic properties differently. Learning this hierarchy is essential for understanding mRNA sequences.
2/5
comment in response to
post
Who knew artist would take the first hit
comment in response to
post
🤚🤚🤚 @artemmoskalev.bsky.social Thanks!
comment in response to
post
😮😮
comment in response to
post
Long context models
comment in response to
post
Hi! Can you add me to the pack? ✋
comment in response to
post
🤚🤚🤚
comment in response to
post
🤚
comment in response to
post
Dankuwel! Can you add me? 🤚
comment in response to
post
I think it only tells that the reviewer didn’t put good effort into formulating and writing an initial review.