00:00.00 Paul Hi welcome back to the archeech podcast episode 2 7 today. We're talking with Marcus Eberl about a recent article of his about about computer vision and Lithic micro-debotage and Marcus just before we went to break before we got a little glitchy there. Your. You're talking about some of the shortcomings of doing this kind of work with a more traditional manual process of using a microscope and counting particles by hand so much so that you gave up on that initially. Yeah, the first time through with it. Um, so aside from just being slow and and maybe. 00:29.36 Markus Yeah. 00:37.79 Paul Inaccurate are there any other major shortcomings to that kind of more traditional approach and how do you address those with the with the computer vision approach. 00:45.60 Markus Yeah, yeah, well thanks for asking that I mean I think another shortcoming is also like the inter-observer error. So like if I want to train other people to do this type of analysis. How well will. 00:54.11 archpodnet M. 00:55.37 Paul Ah. 01:01.93 Markus Different people see the same thing in the same sort and sample and again I mean I have colleagues like isaac ola who really ah trained and other observers. But this has been a problem in the past. How do we really get a standardized observation that different people. 01:13.98 Paul Is. 01:21.30 Markus See and classify the same things and I think especially with micro debotage. We easily run the risk of that people different people see different things and and count different things. 01:35.41 archpodnet We interviewed isaac on this show. Actually yeah yeah for sure I'm wondering another thing I was interested in in what we were seeing here is how the ground surface influences the accumulation of microdebotage like if you've got a really smooth. 01:35.84 Paul Yeah, that's a big problem across a bunch of things. 01:41.85 Markus Um. 01:53.10 Paul Um, and I. 01:55.17 archpodnet Clean Surface I mean you can tend to sweep all that out and have anything whereas if you're just on the ground you know and you've got more crevices and things. So so just tell us a little bit about that in your experience in this collection and what you can find. 01:58.45 Markus Um, watching us are. 02:06.71 Markus You know I mean this is another really critical aspect I mean for example in the area where I'm working with classic maya The upper levels of society had the nicer houses often with very smooth stack of floors and a colleague of mine. 02:18.60 archpodnet M. 02:20.40 Paul N. 02:25.31 Markus Did some experiments with macroabbotage and she found that very little actually stuck to the stuco floors I mean they were very hard and you wouldn't really expect to find macroabbotage in whatever crevices you have there because the surfaces tend to be too. Dense and too smooth for like like a good analysis of Microabbotage on the other hand, the area where I'm working So as a my archeologist I tend to specialize on the non-elite section of the population where we have mostly. Ah. 02:45.72 archpodnet Um. 03:01.37 Markus Pebble floors possibly filled with packed soil in these types of environments microabbotage tends to stick much better I mean so we have here like areas that couldn't be where you could clean the larger flakes off very easily. 03:04.31 Paul Oh. 03:19.67 Markus But I think where people would struggle to clean off micro debotage because it's really, it's its sinks into this oil. 03:24.37 archpodnet Um, well I mean 2 points on that first off I mean just in our own society today I mean you tend to see that. 03:35.80 archpodnet I don't know how to say this in any way that's not pc but certain levels of society tend to just have more things around anyway and don't clean up as much to begin with right? So I have a feeling that that's probably similar in other cultures and along those lines. What are the chances. The people in the elite you know of of Maya Society in particular are even. Flint napping and not just having things created for them. 03:56.99 Markus I Mean that's actually really one of the aspects that I and my colleagues are really interested in because at this moment we don't really know I mean again, you know it makes sense what you say you know like like oh if you're like a lord I mean you won't be napping your own things. 04:05.19 archpodnet Um. 04:13.92 Paul And. 04:15.27 Markus But the point is you know it's really hard to prove that point and that's what I find fascinating about archaeology in general and this approach in particular you know I mean I can go back and say you know if I work in a palace you know do I have. 04:18.36 archpodnet Right. 04:22.62 Paul Ah. 04:29.51 Markus Any evidence of microabittage or is there a potential that you know a few stray flints are still visible. So for me, it's more like a ah question to be asked instead of assuming. Okay, we you know people at that level. They don't care about the napping chart or whatever. 04:36.64 archpodnet M. Um, yeah. 04:48.80 Markus So yeah. 04:49.31 archpodnet Um, okay. 04:50.86 Paul I have a follow up to that then um so a traditional method of trying to identify workshops would be looking at larger ah flakes for example or could be across any of variety of different kinds of artifacts of any kind of workshop that you're looking at. But obviously you'd be looking for Lithic flakes. 04:59.85 Markus Um, of. 05:08.51 Paul If you're looking for these ah for these flint napping workshops in absence of larger flakes is looking at the ah the microdebotage is that a ah reliable indicator of workshops. 05:19.45 Markus You know and we have you know and this is going to be an interesting discussion in the future. So for some researchers they would like to see both meaning that we have here microdibotage and we have some kind of evidence like visible evidence for. 05:24.78 Paul O. 05:39.10 Markus For a workshop and that could be of course larger flakes but you know this could be also like specific other tools that were used for flintnapping that could be a specific layout and I personally would veer more towards that I mean to really I mean I think ah. 05:45.69 Paul No. 05:57.90 Markus Clusters of microabpitage can be very useful but I would like to see you know, additional evidence. For example, you know I mean in the area where I'm working these are larger residential groups with multiple buildings grouped around the plaza and it would be great to see like. Elevated concentrations of Microabbotage in specific areas of this group but I would like to see additional evidence. For example, you know do we have other types of tools that that we find for example in the mittance of this group that would point out. Oh we have here very likely a stonenapper who is working there. 06:20.11 Paul Um. 06:34.19 Markus So for me again, you know I mean as I said you know I'm not against slow data. So you know I am perfectly happy to do an extensive excavation look in 2 mittens and then to really use. For example, my macroabbotage analysis with machine learning. To compliment that with my excavations of mittens and others and at least that's my gut feeling right now. 06:52.90 Paul Um. 06:55.78 archpodnet Okay, yeah, you know man I got to go back to this elite thing real quick because I just ah first off I Love how much a simple question. 07:04.95 Markus Paint. 07:08.46 archpodnet Can actually put a lot more things into play it just lead you down this path because I'm thinking. Okay, so if they are flint napping and somebody's cleaning up after them right? Whether it's them or somebody else cleaning up after them. You think you would see some sort of commonplace that they would sweep up all this debris from inside the house dust skin. Lithic Debotage. Whatever it is and then dump it in a place. So now you're getting to where is their trash collection for this you know elite place and could you analyze that to see well in this what looks like household trash. We have micro debotage. 07:40.75 Markus Now you're You're absolutely right? And and the the problem often is so there have been several archaeologists who did like experiments or ethno archeological studies. The problem with this that modern of Stonenappers. And I say stonenappers because in the area where I'm working people would be both using flint and Obsidian So these stonenappers they often go out of their way to dispose of the trash. 08:03.20 archpodnet Ah, right. 08:09.38 Markus So it's not something you know where you simply you know where you just walk behind the house and you see oh they must be themitten and you know this is where I have to excavate I mean at least in some of these modern cases. Stonenappers really knew you know they wouldn't have their family handling these sharp debris and they really walked like. 08:09.72 archpodnet Um. 08:25.78 Paul Um. 08:29.17 Markus Hundreds of meters away from the area where they were working so in the sense that would then add the challenge. You know as an archaeologist where do I find the trash and especially the workshop ah trash. It could be a different place than where they would put the. 08:41.16 archpodnet Now. Sure. Okay, well I think maybe we should get into the you know part of the actual title of this with the machine learning stuff I'm interested because you mentioned. 08:45.99 Markus Kitchen Debris You know all the vegetables and fruits that were thrown out. 08:58.10 Markus Um, your apple like the line. 09:00.46 archpodnet Twenty years ago you're you're interested in Microdebotage. You're looking through a microscope at soil samples and you're and you're figuring all this out so presumably, you've been thinking about this throughout and it's been. You know one of those things that's kind of like plaguing your mind when do you think or when did machine learning become viable enough for you to look at it and start saying. You know what? this could be a tool where we could actually solve this problem is it with this paper or is it was it gen of that a little bit earlier. 09:25.52 Markus Actually it goes back a little bit earlier and really started with the particle analyzer and this was and I don't really I don't even know what got me there but this was like a late night googling accident. 09:31.18 archpodnet M. 09:37.22 archpodnet The. 09:40.79 Markus And because I was always interested in different ways in which I could describe a flake you know I mean like how could I get like quantitative data on a flake and as I said like ah some googling accident late night brought me to particle analyzers and just to explain that like these are machines used. 09:44.00 archpodnet Yeah. 09:55.66 Paul Oh. 10:00.60 Markus By modern companies for like a quality control. So the company where I bought this machine from. They told me oh they sell it to glasspeed Manufacturers So where people have. 10:05.56 archpodnet Ah. 10:13.26 Markus Ten Thousand Ten Thousand class beats and they want to know whether this run of class beats is exactly you know, according to whatever specification so they were actually super surprised and delighted to have an archaeologist knock on the door and say oh this is exactly what I want to use for my work and. And and this was really the thing where I realized I mean the company invited me to send them 2 samples. You know to to test the capacities of their machine and when I got the results back. 10:37.89 Paul Who. 10:38.65 archpodnet Um. 10:44.13 Markus You know they sent me 2 excel spreadsheets I realized okay I'm in trouble because this is more data than I really expected I mean you know they couldn't even send me a spreadsheet just as an attachment I had to go to like Dropbox or a similar service because they were so large I mean like you know, ah they were like. Tens of thousands of ah particles listed there and each particle I have like 40 variables. So this is then where I realized. Okay I mean you know I mean I'm into statistics and I can handle that but just looking at the spreadsheet I realized you know. 11:11.51 archpodnet Well. 11:22.63 Markus Ah, this seems to be like ah something where machine learning or computers in general could do something really interesting and I then link to the data scientists who are the co-authors on this paper who were super delighted because for them. 11:31.65 archpodnet Ah. 11:40.37 Markus This was the ideal data set that they have been looking for. You know, like experimentally produced data where we know what something is and then you just try to see this specific class in a different data set. 11:42.41 Paul Me. 11:56.20 archpodnet Okay. 11:57.89 Markus So they have been amazing. You know like and I've been working with them for the last five years and this is really where the machine learning then took off because they said you know this is a perfect data set to apply machine learning algorithms so in many ways as an archaeologist I really stumbled into it and but. Ah, through various colleagues I really got into machine learning but it's really like this coincidence of various factors that brought me to this topic and into machine learning. 12:29.24 archpodnet Um, nice. 12:29.54 Paul Well, that's one I kind of like that yet again archeologists are picking on some other industry in this case industry to find tools that we want to use I Always joke that our ah the symbol for our field is you know a bricklayers tool. 12:36.60 archpodnet Um. 12:43.93 Markus Yeah, yeah, yeah. 12:45.44 Paul Not even not even something of our own. 12:47.47 archpodnet Um, yeah, exactly exactly oh man. So now. Let's talk about the the machine learning aspect of it. Actually you know what before we get there I'm curious because you said. Ah, small soil sample that you a small sample that you sent this company and they send back tens of thousands of results what is the smallest particle that their scanner can actually find and suss out and identify and then I've got a follow up to that one related to the debotage so that first. 13:18.91 Markus Yeah, so ah so this machine that ultimately with the help of my university I was able to acquire Meshes everything between thirty five microns so this is what 0.0 3 five millimeters and Thirty five millimeters so it really covers a wide range and you know for the Microabbotage analysis I limit that so I filter out all the smallest and the largest particles. So I limit that to. 13:38.93 archpodnet Um. 13:40.12 Paul Are. 13:52.94 Markus Point one two five millimeters to six millimeters so theoretically this machine could measure much more I also have to filter it out because I ran I mean you know when I got the machine for the first time you know I had no clue and I was just playing around with it and. 13:56.20 archpodnet Okay. 13:59.65 Paul Are. 14:11.44 Markus The machine several times broke down because when I just used the entire range that it can measure I literally ended up with millions of particles specifically because I have tiny clay particles that just you know that create tons of data points. 14:26.21 archpodnet Yeah, yeah, well I mean that I guess that would be expected I get that led to another question I was going to ask how small you can call something micro debotage before it starts becoming. 14:26.43 Paul Nah. 14:42.86 archpodnet I mean potentially natural. You know what I mean I mean how do you even know it was human created when you're getting that small. Yeah, but. 14:43.19 Markus Exactly exactly and and you know this is exactly you know where where I'm really working for these issues because exactly as ah, smaller it gets take either the rounder these Microabbotage particles become and they are. 14:44.29 Paul Dust. 14:58.67 archpodnet Sure. 15:02.44 Markus Are very I mean you know I mean they become very hard to distinguish from regular Sand kernels that you have in solar samples. So this is really one of the issues that I'm working on right now where I say if you go too small. It doesn't really. 15:09.90 archpodnet Um. 15:18.85 Markus It becomes very hard to distinguish between these 2 categories and it doesn't make sense to call very small particles microtabbotage because it's really hard to distinguish them from record or soil particles. 15:28.30 archpodnet Yeah I'm sure they experience erosion much faster too. You know the the edges get knocked off things like that just because of their size I would imagine. Yeah so all right? Well let's take a break. 15:36.73 Markus Exactly him. 15:38.60 Paul So. 15:42.15 archpodnet Let's take a break and on the other side. Yeah, we'll wrap up this discussion and and maybe talk a little bit more about the algorithms and the actual methods that you guys used to do this so we'll do that in the meantime I'll mention once again, check out arcpodnet.com/members. To become a member of the archeology podcast network we've got a kaltuo share event coming up. That's the Sunday after this podcast releases about underwater maritime archeology. So check that out if you miss it, you can watch it if you remember anytime you want. We'll be back in a minute.