Can ChatGPT be used to identify a tartan?

[jsrnephdoc]

I General and Generative are their own sets within that sphere.

I didn't know that they were not one and the same. So, if I were looking for a Generative AI answer to my question, I'd be making the following assumptions:

1. The generative AI database would recognize that the needed dataset would include numerical data such as thread counts and sett size
2. Those would be prioritiezed over inferences about color, unless a dataset that specified color based on pantone numbers or something equivalent existed.
3. only THEN could a generative AI engine answer my question, and even then it's more complicated, because the "look" of a tartan, including sett size depends on other factors as well, such as how thick the individual threads are and how tightly they're woven.

And my guess is that that sort of cataloging has not been done.

[bookish]

I didn't know that they were not one and the same. So, if I were looking for a Generative AI answer to my question, I'd be making the following assumptions:

1. The generative AI database would recognize that the needed dataset would include numerical data such as thread counts and sett size
2. Those would be prioritiezed over inferences about color, unless a dataset that specified color based on pantone numbers or something equivalent existed.
3. only THEN could a generative AI engine answer my question, and even then it's more complicated, because the "look" of a tartan, including sett size depends on other factors as well, such as how thick the individual threads are and how tightly they're woven.

And my guess is that that sort of cataloging has not been done.

You know, your points are allowing for me to get my thoughts in order a little better. I probably have given the wrong impression about thread count and color.

1. I think this approach is a non-AI way of identifying tartan. If you had a relational database of sett colors and threadcount, and an algorithm that analyzes images of tartan to approximate the size of color blocks, then you could determine if a picture of a tartan was more likely tartan x or tartan y.
2. I think it is enough to try to say, "this block of color is likely within the [general color name] color space." Though, this probably is a problem for the non-AI way of trying to identify a tartan with an algorithm.
3. AI is trained on images in different ways, but generally AI doesn't know or care what the context of an image is. It doesn't know what tartan is, and doesn't relate context to its determination. It just says, "I have been trained to say that when something kinda looks like this, it is tartan. This is more yellow, so there is a likelihood that it is Macleod." It doesn't look up "loud Macleod" in a database, it just has to be told a lot what is and what isn't Macleod, and probably has a statistical weight that says when something kinda looks like this, it is probably Macleod.

I bet someone in the mill industry is trying to train a model on tartan...

I have a friend who actually works on AI stuff, so I'll ask him about this use case.

[OC Richard]

The tartan is Ayrshire Gold, a 2023 fashion design by Clyde Kilts, an Ayrshire based retailer.

Attachment 44445

As if one wasn't bad enough, they also do an Ayrshire Silver.

Attachment 44446

Thanks Peter!

As we know some of the big firms involved in Kilt Hire trot out entire lines of tartans from time to time, and I had guessed that it might be from one of those. I didn't know about Clyde Kilts.

Witness MacGregor And MacDuff's line of "Mist" tartans https://macgregorandmacduff.co.uk/co...20Mist%20Kilts

and McCall's "Pride" tartans https://www.mccalls.co.uk/pride-pages/weathered-pride/

and Houston's "Bute" tartans https://www.kiltsforhire.com/search?q=bute

[bookish]

Since I brought up that there are different types of AI or machine learning, I didn't want to leave it at that. So that anyone can look into it themselves, I asked my friend, Dr. Shearman about this. My questions were:

1. Is AI really the right tool, or would a more deterministic set of algorithms be better?
2. What methods of training an AI on the images be best?
3. What general class of AI would this fall under?

His responses were:
1. Yep, AI or ML (machine learning) is the right tool for this task.
2. Assuming you don't have any a priori structure to your images, I would try fine tuning a vision language model.
3. Vision language models are what you want. Inputs: (image, text) -> text. Such that (img, "does your image contain a person wearing a kilt?") -> "yes"

He went on to explain that there is a way to segment the image to focus only on the tartan, and use other image processing for the identification. You can also fine tune the model at the final stages of development on targeted examples, which requires a lot of labeled data. He also stated that a small convolutional neural network might be able to solve this problem directly.

Much of this is specialized knowledge that I would have to look into more to understand (like what a "convolutional neural network" is). So, I think I'll leave it at that for folks to investigate if this is really a problem space you find interesting. I don't know if the finished product would justify the cost in time and money to develop, but there we are.

[jsrnephdoc]

Much of this is specialized knowledge that I would have to look into more to understand (like what a "convolutional neural network" is). So, I think I'll leave it at that for folks to investigate if this is really a problem space you find interesting. I don't know if the finished product would justify the cost in time and money to develop, but there we are.

So, now we have TWO problems, one quite generic but possibly VERY interesting; i.e., could an ambitious AI wrangler develop an algorithm that would do a decent job of answering the "what tartan is this?" question (that might be of general use to retailers, customers, and even tartan historians).

The other, far more practical, is "who weaves the Ashyre Gold tartan, and how would one go about obtaining a kilt made from it?" The answer to the second part of that one is probably obvious, except that I know nothing about the retailer that Peter MacDonald identified, or who sews for them, what fabric weights are available, etc., and the email contact I found for the retailer has not triggered any response as yet.

[figheadair]

So, now we have TWO problems, one quite generic but possibly VERY interesting; i.e., could an ambitious AI wrangler develop an algorithm that would do a decent job of answering the "what tartan is this?" question (that might be of general use to retailers, customers, and even tartan historians).

The other, far more practical, is "who weaves the Ashyre Gold tartan, and how would one go about obtaining a kilt made from it?" The answer to the second part of that one is probably obvious, except that I know nothing about the retailer that Peter MacDonald identified, or who sews for them, what fabric weights are available, etc., and the email contact I found for the retailer has not triggered any response as yet.

In reverse order; Clyde Kilts appear to be an online retailer and so you would need to contact them. House of Edgar have a link page for Clyde Kilts which suggests that they might be the weavers.

As for the first question. I've yet to be convinced that a machine can, or will ever be able to do what a human can without access to all the source information and then be able to account for the variables of shade and settings.

[jsrnephdoc]

In reverse order; Clyde Kilts appear to be an online retailer and so you would need to contact them. House of Edgar have a link page for Clyde Kilts which suggests that they might be the weavers.

I've learned that it's Lochcarron, and that it's 16 oz fabric. A bit of a surprise, but they're making neckties as well from that heavyweight cloth (evident, of course, from the video of the singer before the Tartan Day Parade.

As for the first question. I've yet to be convinced that a machine can, or will ever be able to do what a human can without access to all the source information and then be able to account for the variables of shade and settings.

There, I suspect, despite almost NO comprehension of the magnitude of the problem nor the tools immediately available to deal with it, I suspect you're underestimating the LLMs. Over the past 4o years, the amount of data that a relatively inexpensive desktop computer can look at simultaneously has increased from 400,000 Bytes to 800,000,000,000,000 (or thereabouts). Our ability to scan that much information on paper or using our own memory, of course, has not. So, obstacles such as thread count, sett size (which can change just because of the thickness of the yarn made to weave cloth) eventually won't be obstacles, and the computer won't "forget" something it's taken into account already in adding more information to an ongoing analysis.

That said, of course I need to defer to others far more knowledgeable than I (barely beginning my own dabbling) about AI who are contributing to this thread as to whether my rumination are in any way actually descriptive of the issues.

[bookish]

There, I suspect, despite almost NO comprehension of the magnitude of the problem nor the tools immediately available to deal with it, I suspect you're underestimating the LLMs. Over the past 4o years, the amount of data that a relatively inexpensive desktop computer can look at simultaneously has increased from 400,000 Bytes to 800,000,000,000,000 (or thereabouts). Our ability to scan that much information on paper or using our own memory, of course, has not. So, obstacles such as thread count, sett size (which can change just because of the thickness of the yarn made to weave cloth) eventually won't be obstacles, and the computer won't "forget" something it's taken into account already in adding more information to an ongoing analysis.

That said, of course I need to defer to others far more knowledgeable than I (barely beginning my own dabbling) about AI who are contributing to this thread as to whether my rumination are in any way actually descriptive of the issues.

I think we encounter the question, not can we do it, but should we do it, and at what cost?

If your interest and aim is to have a bit of fun and make something for yourself that works, then you can ignore the rest of my post. If the aim is to create something that is going to be a product that one expects wide adoption to use, then I have some thoughts.

With more compute power, comes more energy consumption. AI requires a LOT of energy, both for training, and for responding to a user's requests. And, for AI products to be profitable, we generally will see "disruption" business practices, which rely on offering inexpensive services to people so that they will adopt them, scale the services to get large-scale adoption from users, and operate at a loss until enough of the market is captured. Once we reach scale, and adoption is so great that people are locked into the user experience, then prices rise, and the good user experience is sold off at a premium, while those who hold onto their free, or less expensive access, get a continually worsening experience. If you look up Cory Doctorow and "user experience" in the same search engine query, you will get a thesis on that. Meanwhile, scale of AI means so many resources are consumed, the carbon footprint increases, and we are sold on its convenience to keep it afloat, etc.

Here is my bias showing, of course. Another aspect to my bias is that searching for simple answers with such a tool seems like overkill when we have great tools already to find the answer. I'd say try out the search functionality at the Scottish Register of Tartans website, and with a registered account, you can request sett information. If the tools seem lacking in some way, and you have a feature request to make it better, they might take feature requests into account. Another option is building a relationship with someone that has specialized knowledge to help you learn to identify tartan better.

I have a lot of concerns about AI, as I have mentioned. Not that it can't or wouldn't be able to do a thing, but should it do every little thing and answer our every little question? You might be underestimating human potential.

[OC Richard]

For sure that's what computers are good at, storing and accessing vast amounts of data.

I don't doubt that it could be done. The devil is in programming the thing to recognise the right things.

For example, it would have to be able to recognise a wide, but not too wide, range of colours as being equivalent for the purposes of tartan recognition.

This throws off humans often enough! How many would be able to see all these kilts as being the same tartan?

(It's Bruce of Kinnaird.)



Or see these as not being the same tartan?

[jsrnephdoc]

For sure that's what computers are good at, storing and accessing vast amounts of data.

I don't doubt that it could be done. The devil is in programming the thing to recognise the right things.

For example, it would have to be able to recognise a wide, but not too wide, range of colours as being equivalent for the purposes of tartan recognition.

This throws off humans often enough! How many would be able to see all these kilts as being the same tartan?

(It's Bruce of Kinnaird.)

I'm coresponding with an absolute wizard when it comes to Apple ecosystem information technology. At first he assumed that "Machine Learning" rather than a "Large Language Model" could be used, but he actually delved into how the registry is constructed and concluded (as you so nicely just demonstrated) that "appearance" would never be enough to identify a tartan. His summary analysis statement concluded

" I'm not sure what you think AI or ML might contribute: you can't map between appearance and thread count. All those lovely old books showing examples of different tartans are largely irrelevant, as you can't judge by the way they look, just as you can't catalogue a person's genome from their face."

And, of course, that's not the end of the issue. I would guess that the number of times some INDIVIDUAL sees someone else wearing a kilt he admires and would like a kilt in the same tartan for himself dramatically exceeds the number of times someone is playing with a computer program that can create a new tartan description hoping to register it, and many, many of the "fashion" tartans are not registered but are still protected somehow by intellectual property law.

And it's also interesting that in the instant case, the selfsame guy who spends a great deal of time plumbing the historical register was the first to recognize the tartan I inquired about, AND did so BASED SOLELY on its appearance!