Imageomics / Imageomics/FuncaPalooza-2025

SAEs for qualitative beetle traits

Open
#2 4 comments 0 reactions 0 assignees View on GitHub
project idea
Dominant language
No language data
Stars
5
Forks
0
PR merge metrics
No merged PRs in 30d

Description

I would love to train sparse autoencoders on beetles for a couple reasons:

1. Visually, beetles are not very diverse (compared to the rest of the tree of life). Do SAEs work on them? (ML)
2. Species ID can be pretty challenging and this could find useful *morphological* traits for species ID, which might connect to functional traits. (Bio)

Below are some interesting/relevant SAE traits discovered by an SAE trained on iNat21 (which is quite diverse).

Image

Image

Image

Contributor guide

No contributing guide indexed for this repository

Research direction

No files, tests, or entry points are mentioned. Start by reviewing the linked iNat21 SAE traits and define the beetle image data, SAE training approach, and evaluation needed to determine whether qualitative morphological traits emerge for species identification.

Written by the indexing model from the issue text.

Assessment

Domain
bioinformatics, machine-learning
Issue type
Feature
Difficulty
5/5
Estimated time
Over a week
Activity status
Stale
Clarity
Needs clarification
Newbie friendliness
25/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.