Anthony Gitter
Anthony Gitter
I'm not attempting to keep up with the nucleotide models (Evo, LucaOne, DNABert2, HyenaDNA, Boshar 2024 Nucleotide Transformer, etc.) or multi-modal PLMs.
The goal of #2 was to greatly simplify the installation by reducing the number of commands someone needs to run to create the environment. This updates to readme to use...
This new article is relevant for the references: Guiding questions to avoid data leakage in biological machine learning applications https://doi.org/10.1038/s41592-024-02362-y
We often get questions about regression models, which brings the current classification-only scope into question. We should discuss whether there is a way to add some regression support into the...
The decision tree lesson asks which hyperparameter is important for the `simulated_t_cell/simulated_t_cells_7.csv` dataset. It is not clear what the expected answer is. class_weight could be relevant because of the class...
Given the rise of LLMs, should we define self-supervised learning in the Intro lesson or glossary? It would be a slight distraction from the main focus of classification. On the...
The Carpentries instructions that we link to no longer use Anaconda, they migrated to conda-forge for the conda distribution. We should do the same.