this post was submitted on 23 May 2024
1018 points (99.0% liked)

Science Memes

11047 readers
3264 users here now

Welcome to c/science_memes @ Mander.xyz!

A place for majestic STEMLORD peacocking, as well as memes about the realities of working in a lab.

Rules

  1. Don't throw mud. Behave like an intellectual and remember the human.
  2. Keep it rooted (on topic).
  3. No spam.
  4. Infographics welcome, get schooled.

This is a science community. We use the Dawkins definition of meme.

Research Committee

Other Mander Communities

Science and Research

Biology and Life Sciences

Physical Sciences

Humanities and Social Sciences

Practical and Applied Sciences

Memes

Miscellaneous

founded 2 years ago
MODERATORS
[email protected]
[email protected]
[email protected]
1018
Name & shame. :) (mander.xyz)
submitted 5 months ago* (last edited 5 months ago) by [email protected] to c/[email protected]
95 comments fedilink hide all child comments
 
you are viewing a single comment's thread
view the rest of the comments
[–] [email protected] 190 points 5 months ago (18 children)

Dude. Couldn't even proofread the easy way out they took

[–] [email protected] 105 points 5 months ago (12 children)

This almost makes me think they're trying to fully automate their publishing process. So, no editor in that case.

Editors are expensive.

[–] [email protected] 19 points 5 months ago (1 children)

If they really want to do it, they can just run a local language model trained to proofread stuff like this. Would be way better

[–] [email protected] 12 points 5 months ago (1 children)

This is exactly the line of thinking that lead to papers like this being generated.

[–] [email protected] 1 points 5 months ago (2 children)

I don't think so. They are using AI from a 3rd party. If they train their own specialized version, things will be better.

[–] [email protected] 11 points 5 months ago (1 children)

Here is a better idea: have some academic integrity and actually do the work instead of using incompetent machine learning to flood the industry with inaccurate trash papers whose only real impact is getting in the way of real research.

[–] [email protected] 3 points 5 months ago (2 children)

There is nothing wrong with using AI to proofread a paper. It's just a grammar checker but better.

[–] [email protected] 4 points 5 months ago (1 children)

You can literally use tools to check grammar perfectly without using AI. What the LLM AI does is it predict what word comes next in a sequence, and if the AI is wrong as it often is then you've just attempted to publish a paper with halucinations wasting the time and effort of so many people because you're greedy and lazy.

[–] [email protected] 2 points 5 months ago (1 children)

AI does better at checking for grammar and clarity of message. It's just a fact. I've made comparisons myself using a grammar checker on an essay vs AI and AI corrected it and made it much better.

[–] [email protected] -3 points 5 months ago (1 children)

AI doesn't do anything better than a human being. Human Beings are the training data, an AI that mimics it 98% is still less accurate than the humans. If you suck so much at writing papers then you're just below average as a human being who writes papers and using tools will never remedy that without introspection and a desire to improve.

[–] [email protected] 2 points 5 months ago

You said that "you can literally use tools to check grammar perfectly" I've responded to that claim. No mention of humans. You seem to be projecting

[–] [email protected] 4 points 5 months ago* (last edited 5 months ago) (1 children)

Proofreading involves more than just checking grammar, and AIs aren't perfect. I would never put my name on something to get published publicly like this without reading it through at least once myself.

[–] [email protected] 4 points 5 months ago

I entirely agree. You should read through something you'll publish.

[–] [email protected] 1 points 5 months ago

That's not necessarily true. General-purpose 3rd party models (chatgpt, llama3-70b, etc) perform surprisingly good in very specific tasks. While training or finetuning your specialized model should indeed give you better results, the crazy amount of computational resources and specialized manpower needed to accomplish it makes it unfeasible and unpractical in many applications. If you can get away with an occational "as an AI model...", you are better off using existing models.

load more comments (10 replies)
load more comments (15 replies)