Blending genomics with deep learning, making GPU-accelerated computers do complicated stuff without giving them any food or water.
Canberra
Joined July 2018
- Tweets905
- Following217
- Followers377
- Likes612
Eli Niktab retweeted
>renames it to the Department of War
>names himself the Secretary of War
>fights one war
>loses
Eli Niktab retweeted
Changing training data alone can improve deep learning prediction of spatial transcriptomics gene expression from histology images by 38% (without any changes to model architecture).
We've updated our preprint showing this with expanded results: biorxiv.org/content/10.1101/…
🧵👇 1/4
Eli Niktab retweeted
Did you know that networks trained with different learning rates extract different features (and a different number of them!) from the data? Come by our poster at HiLD Workshop #ICML2024 tomorrow to discuss it with @irsadrtdinov!
Paper: openreview.net/forum?id=IID2…
1/6
Eli Niktab retweeted
A student told me, "I don't get the difference between model fitting by:
1. minimum Mean Square Error (MSE) and
2. Maximum Likelihood Estimation (MLE)"
To help this student, I built out this #Python @matplotlib dashboard for fitting a Gaussian parametric distribution to data while calculating & visualizing the training data MSE and likelihood!
Many students have found this helpful, so I shared it on #GitHub @ github.com/GeostatsGuy/DataS… ∀. #DataScience
Eli Niktab retweeted
finally finished the antibody ML primer post
if you are in the protein ML field, not in antibody ML, and would like to understand antibody ML bit more, you’re my ideal audience!
abhishaike.com/2024/04/15/a-…
🧵 🧵 🧵 – at Queens, NY
Eli Niktab retweeted
Eigenvalues of random matrices with iid entries converge to the Wigner circle law. This is universal (does not depend on the law of the entries), as proved by Tao and Vu in 2010. For Gaussian matrices, this defines a determinental point process. en.wikipedia.org/wiki/Circul…
Eli Niktab retweeted
What impact is AI having on innovation in New Zealand today, and what is the potential to disrupt further sectors in future? #Techweek23 #AI #Tech
Eli Niktab retweeted
Read about how the @DevermanLab has brought biologists and computer scientists together to rethink experiments to make the best use of machine learning. The lab came up with results that could accelerate the development of better gene therapies. broad.io/DevermanFeature
Eli Niktab retweeted
Self-organizing maps of Kohonen defines a map between a 2D space to the data, by minimizing a quantization error and a regularity penalty. en.wikipedia.org/wiki/Self-o…
Eli Niktab retweeted
IMAGEBIND: One Embedding Space To Bind Them All.
Learns a joint embedding across six different modalities - images, text, audio, depth, thermal, and IMU data.
An open source project by Meta-FAIR.
Paper: dl.fbaipublicfiles.com/image…
Demo: imagebind.metademolab.com/
Code: github.com/facebookresearch/…
Blog post: ai.facebook.com/blog/imagebi…
Eli Niktab retweeted
Three years ago, I started working on an easy-to-use tool for interpretable machine learning in science. I wanted it to do for symbolic regression what Theano did for deep learning.
Today, I am beyond excited to share with you the paper describing it!
arxiv.org/abs/2305.01582
1.
Eli Niktab retweeted
Our new preprint: Using synthetic biology we answer the question, why does one transcription factor binding site act differently in different contexts? Check out our surprisingly simple model of regulatory DNA.
Great work by Kai Lowell and @rfriedman22!
biorxiv.org/content/10.1101/…
