3 ms·

I send things too iNaturalist all the time, it's great, it really helped me learn about my local fauna. I want to do a project with their API to identify a coup

I send things too iNaturalist all the time, it's great, it really helped me learn about my local fauna. I want to do a project with their API to identify a couple hundred wildflower photos I've been hoarding. Would that work? ( Idea is my wildflower app could send to their models to confirm my original identification)

jw_cook 7mo ago
I've wanted to do something similar, but unfortunately their CV model isn't public and can't be used through their API.
- contingencies 7mo ago
  Yet they shelter under a 'Science' tax-break. It's duplicitous. They should publish their models and build process. If it's not available for replication, it's not science.
- Taipan_Enigma 7mo ago
  Are their models considered to be the best or is there some competition? For plant identification, they blow every other free app I have tried out of the water. It also seems to return the genus of a plant rather than misidentify the species which I find impressive.
  - jw_cook 7mo ago
    IMO it's best-in-class. The next best thing might be google's speciesnet: https://github.com/google/cameratrapai https://github.com/google/cameratrapai
    - rafram 7mo ago
      Any idea how SpeciesNet and iNat’s model compare to BioCLIP 2? https://huggingface.co/imageomics/bioclip-2 https://huggingface.co/imageomics/bioclip-2
- skyberrys 7mo ago
  That's too bad, maybe I can upload it to iNaturalist then reference the entry there. I don't mind if it's duplicated, I just want to be able to improve the location data without sharing the improved location data so publicly.
Matumio 7mo ago
I don't know if it will work, but Pl@ntNet Identify (which I use often) seems to have an API: https://docs.plantnet.org/en/reference/api-plantnet/ https://docs.plantnet.org/en/reference/api-plantnet/
frontierkodiak 7mo ago
Hey, good news! Pollination ecologist + ML guy here; with open models coming soon. You can keep an eye on (gh) polli-labs/linnaeus (a bit stale; I'll rebase on my private repo later tonight-). There's some cool ideas in here to exploit the structure of taxonomic hierarchies to help the model approach recognition how a professional taxonomist might.. so working from coarse to fine, taxonomy-guided label smoothing (distributing alpha mass by taxonomic distance)..and (forthcoming) RL on expert consensus to teach abstention (if an expert could only identify a specimen to genus for some set of inputs; then our model should abstain from a species classification for the same inputs). Unfortunately I am very, very compute-constrained- but shooting for late April/first week of May for insect + flowering plant models. (Other taxa will come later; probably as unified model). I'm working on camera-based (automated) ecological monitoring systems for ~6yrs at this point; it's a really fun problem space! dropped out of grad school to go all-in on automating my favorite job I ever had (pollination ecology field research..watching flowers for visitations!); since I knew I'd always be a mediocre ecologist- but an engineer that happens to care about ecology could be very very valuable to my field. a taxa recognition model turns out to be only a small piece of the system you need to extract structured observational data from cameras in the field :-) Working with one of my partners right now to launch a really cool demo of what's possible these days- Texas folks especially; keep an eye out on wildflower.org around May 1! I'll spill more ink soon but (anyone) please get in touch if you find these things interesting. Or if you'd like to help me out with compute/expenses!
- skyberrys 7mo ago
  This project looks awesome pollination ecologist + ML guy! I starred it and will definitely be exploring using this. I'm actually pretty simple with my needs because I'm after only a few specific wildflowers not the whole ecosystem. But of course it's impressive how comprehensive your project is looking.

jw_cook 7mo ago

I've wanted to do something similar, but unfortunately their CV model isn't public and can't be used through their API.

contingencies 7mo ago
Yet they shelter under a 'Science' tax-break. It's duplicitous. They should publish their models and build process. If it's not available for replication, it's not science.
Taipan_Enigma 7mo ago
Are their models considered to be the best or is there some competition? For plant identification, they blow every other free app I have tried out of the water. It also seems to return the genus of a plant rather than misidentify the species which I find impressive.
- jw_cook 7mo ago
  IMO it's best-in-class. The next best thing might be google's speciesnet: https://github.com/google/cameratrapai https://github.com/google/cameratrapai
  - rafram 7mo ago
    Any idea how SpeciesNet and iNat’s model compare to BioCLIP 2? https://huggingface.co/imageomics/bioclip-2 https://huggingface.co/imageomics/bioclip-2
skyberrys 7mo ago
That's too bad, maybe I can upload it to iNaturalist then reference the entry there. I don't mind if it's duplicated, I just want to be able to improve the location data without sharing the improved location data so publicly.

Matumio 7mo ago

I don't know if it will work, but Pl@ntNet Identify (which I use often) seems to have an API: https://docs.plantnet.org/en/reference/api-plantnet/ https://docs.plantnet.org/en/reference/api-plantnet/

frontierkodiak 7mo ago

Hey, good news! Pollination ecologist + ML guy here; with open models coming soon. You can keep an eye on (gh) polli-labs/linnaeus (a bit stale; I'll rebase on my private repo later tonight-). There's some cool ideas in here to exploit the structure of taxonomic hierarchies to help the model approach recognition how a professional taxonomist might.. so working from coarse to fine, taxonomy-guided label smoothing (distributing alpha mass by taxonomic distance)..and (forthcoming) RL on expert consensus to teach abstention (if an expert could only identify a specimen to genus for some set of inputs; then our model should abstain from a species classification for the same inputs). Unfortunately I am very, very compute-constrained- but shooting for late April/first week of May for insect + flowering plant models. (Other taxa will come later; probably as unified model). I'm working on camera-based (automated) ecological monitoring systems for ~6yrs at this point; it's a really fun problem space! dropped out of grad school to go all-in on automating my favorite job I ever had (pollination ecology field research..watching flowers for visitations!); since I knew I'd always be a mediocre ecologist- but an engineer that happens to care about ecology could be very very valuable to my field. a taxa recognition model turns out to be only a small piece of the system you need to extract structured observational data from cameras in the field :-) Working with one of my partners right now to launch a really cool demo of what's possible these days- Texas folks especially; keep an eye out on wildflower.org around May 1! I'll spill more ink soon but (anyone) please get in touch if you find these things interesting. Or if you'd like to help me out with compute/expenses!

skyberrys 7mo ago
This project looks awesome pollination ecologist + ML guy! I starred it and will definitely be exploring using this. I'm actually pretty simple with my needs because I'm after only a few specific wildflowers not the whole ecosystem. But of course it's impressive how comprehensive your project is looking.