Should iNat have been mentioned?

[joerich]

I was just wondering if iNat should have been given credit in this news release? Did they use iNat’s database?
https://phys.org/news/2024-02-ai-tool-leverage-database-million.html

[tockgoestick]

Here’s the PDF of the actual paper.

iNaturalist is mentioned multiple times, including Table 1.

[joerich]

Thank you. Do you know what iNat21 is?

[pisum]

https://github.com/visipedia/inat_comp/tree/master/2021

[joerich]

Even though iNaturalist is referenced many times in the research paper, to acknowledge and give credit to all the iNaturalist members who have contributed observations to the data set on which this work is based, I believe that Ohio State University should have mentioned that this work used the iNaturalist data set.

I am also wondering if iNaturalist is working with or planning on using this AI model for identifications in the future? It does seem to provide another step on the ladder for computer recognition.

[cthawley]

The paper definitely follows standard citation norms for scientific papers in my eyes. It cites the dataset, existing presentations about iNat CV/data (which are the only ones available, I think), and mentions the iNat team in the Acknowledgements. Looks good to me.

[pfau_tarleton]

The author of the article has a B.A. in Journalism. My experience with journalists writing press releases about scientists’ research is that they rarely get every detail correct–nor include everything even if the scientists themselves want it included. Omission of iNat from the university press release isn’t a big deal in this particular case.

[samantha_knight]

I work in marketing and PR, and if this was written as a press release, it’s entirely possible that they were required to keep it very factual and without opinions. We’ve had several returned to us because the reviewer felt the language was too laudatory.

[joerich]

My question was about the press release in phys.org. The scientific article does reference iNat as needed.

https://phys.org/news/2024-02-ai-tool-leverage-database-million.html

[cthawley]

I don’t think it’s standard practice/required to cite sources in PRs for papers/studies other than the paper being covered (though I sometimes see PRs that lack this which is a major fail by the PR writer…lol). If a study was based exclusively on or is focused on one specific data source, then that might be appropriate (even if not required), but in this case iNat is one of several data sources, so I don’t personally see any issue.