AI models for Biology
A good website keeping us all current for biological AI models. https://bio.rodeo There are not that many open models but the website lists their “openness” and their citation.
-
-
Fitness landscape for antibodies (Flab2)
There have been many AI models to predict the structure of proteins, especially antibodies. There are several key aspects to developing antibodies as drugs beyond binding to the target of interest. These are : thermostability, expression, aggregation, binding affinity, pharmacokinetics, polyreactivity and immunogenicity. There have been many models available such as IgLM, ProGen2, Chai-1, ESM2,…
-
AI job search
With AI enabling many activities in the job hunt process, it is expected that many job applicants and executive at hiring companies use AI based tools for the process. The “automated” process used to be enabled by keywords which was the way that the candidates were selected from a large pool but with the availability…
-
AI | Biology | Pharmaceutical
AWS Target Discovery agent
AI agents are everywhere accomplishing activities. Drug Discovery agents have been coming on fast and the scale at which these agents approach drug discovery is phenomenal. There is a great example of a AWS agent that was SPECification built using Kiro from scratch in 3 weeks that delivers drug targets. Amazing to see that it…
-
Noetik – Data driven models
Came across an interesting company called Noetik.ai https://www.noetik.ai/ Their platform and models are using spatial biology data to develop AI models that could help in drug discovery. They have partnered with GSK to in Colorectal cancer and Non small cell lung cancer and GSK will license their OCTO models for their research. The image is…
-
AI in subjective domains
I am exploring how AI can scale expert feedback in subjective domains with a hobby project. After visiting art museums around the world, I kept coming back to a familiar challenge: how do you consistently evaluate something inherently subjective? After having created medical devices in pain, the idea was worth exploring. It’s a question that…
-
Tools for critique of art – open source
Good publication to review material: CognArtive: Large Language Models for Automating Art Analysis and Decoding Aesthetic Elements at https://arxiv.org/html/2502.04353v1 And the wonderful visualization to analyze the data over the years that was derived from WikiArt. https://cognartive.github.io Autocritic: an open-source system that uses classic art theory to evaluate images and steer generative systems. It distills historical…
-
LoRa in AI image generators
This stands for Low Rank adaptation. In AI terms, this is a small add-on to the model that makes it do specific things. It is usually used in the context of stable diffusion. A large model takes very long to train with significant resources. However, sometimes, you need specific details – for example you could…
-
Text from Image
Starting with a base image, it is possible to generate a description. An AI (Qwen-3-VL) simple description : This is a simple, hand-drawn black-and-white sketch of a person’s upper body. Here’s a breakdown of the key elements: Hair: Long, flowing hair depicted with two smooth, curved lines extending from the top of the head down…
-
Engaging website
There are many ways that websites show their landing pages. Usually there are a lot of graphics. The Harvard innovation lab does it wonderfully with their landing page for one of their challenges showing particles which are responsive to mouse movement. Very dynamic and very intriguing! https://innovationlabs.harvard.edu/presidents-innovation-challenge/finalists The developer’s console tells you a bit about…
