![\"AI](\"https:\/\/www.europesays.com\/ie\/wp-content\/uploads\/2026\/08\/ai-trained-on-simulated-sites-1.jpg\")
<\/a>Post-wildfire lidar-derived digital surface model (DSM) of the area around Topanga Beach, Los Angeles, California, shown with green-to-brown elevation shading. Data like these provide the raw material researchers use to train AI models to detect archaeological features. Credit: Jason Stoker, Ph.D., U.S. Geological Survey. Public domain.<\/p>\nThe study focused on 12 unusual circular structures in Kisatchie National Forest, Louisiana. The structures looked similar to historic tar kilns, used to extract tar, pitch, and resin from pine trees. Yet their shape differed from known tar kilns found elsewhere in the southeastern United States.<\/p>\n
The difference created a problem for machine learning. A model trained on existing tar kilns would struggle to spot the Louisiana structures. Only 12 examples existed, too few for a reliable training set.<\/p>\n
The team tested two ways to create artificial training<\/a> data. Both methods placed simulated kiln shapes into digital elevation models made from lidar data<\/a>. Lidar records small changes in ground height, helping archaeologists map features hidden beneath forest cover.<\/p>\n The team trained Mask R-CNN models to identify the objects. The models used tiled lidar images, with 80 percent for training, 10 percent for validation, and 10 percent for testing. Each model trained for 15 rounds before its performance leveled off.<\/p>\n The first model produced 183 predictions. After automatic filtering, researchers reviewed 142. The model found nine of the 12 known targets and six other objects worth checking later.<\/p>\n The second model predicted 2,032 objects, with 709 left after filtering. It found all 12 known targets and 11 additional objects worth further study. Yet 686 of the remaining predictions were false positives.<\/p>\n Many false results came from reservoirs, drainage features, and natural mima mounds. Extra filters removed some errors. Researchers used ground shape to tell mound-like forms from the pit-like shape of the targets.<\/p>\n A third model provided a comparison. Researchers trained it with real tar kilns from South Carolina, after changing the elevation data to better match the Louisiana objects. This model found 11 of the 12 targets and seven additional objects worth checking.<\/p>\n Field work changed the main question. Researchers visited 11 of the original targets and tested two with augers. They found no charcoal-rich deposits, charred pine wood, buried wooden tar pipes, or hard clay floors expected at tar kilns.<\/p>\n The structures also sat near a former World War II military<\/a> training site. Based on their location and shape, the team concluded they were more likely linked to military training. A wartime training manual suggested the structures could be howitzer emplacements.<\/p>\n Automated detection still needs field checks. The models were good at finding unusual shapes, but many predictions were wrong. Cleaning the results took less time than a pedestrian survey of the same area.<\/p>\n The study points to a way around a major machine learning limit. Simulated training objects could be produced in minutes, while manual labeling of real sites takes much longer.<\/p>\n The team suggests using such models before detailed image marking or field surveys. Larger simulated datasets with thousands of objects could improve future models.<\/p>\n The main gain for archaeology is speed. A model does not replace field work, but it helps narrow a landscape to fewer places worth checking. In this case, the approach helped researchers find targets and reject their first idea about what those targets were.<\/p>\n Publication:<\/strong> Peck, K., Gravel-Miguel, C., Snitker, G., & Helmer, M. (2026). Using simulated training data to locate archaeological sites with machine learning.\u00a0Advances in Archaeological Practice,\u00a014(2), 179\u2013197. doi:10.1017\/aap.2025.10130<\/a><\/p>\n","protected":false},"excerpt":{"rendered":"Archaeologists have tested a new way to train deep learning models when real examples of a site type…\n","protected":false},"author":2,"featured_media":652822,"comment_status":"","ping_status":"","sticky":false,"template":"","format":"standard","meta":{"footnotes":"","_share_on_mastodon":"0"},"categories":[261],"tags":[291,275888,275889,289,290,18,275890,19,17,1000,52724,82],"class_list":["post-652821","post","type-post","status-publish","format-standard","has-post-thumbnail","category-artificial-intelligence","tag-ai","tag-ai-in-archeology","tag-archaeological-methods","tag-artificial-intelligence","tag-artificialintelligence","tag-eire","tag-geoarchaeology","tag-ie","tag-ireland","tag-lidar","tag-remote-sensing","tag-technology"],"share_on_mastodon":{"url":"https:\/\/pubeurope.com\/@ie\/117146896214276712","error":""},"_links":{"self":[{"href":"https:\/\/www.europesays.com\/ie\/wp-json\/wp\/v2\/posts\/652821","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.europesays.com\/ie\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.europesays.com\/ie\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.europesays.com\/ie\/wp-json\/wp\/v2\/users\/2"}],"replies":[{"embeddable":true,"href":"https:\/\/www.europesays.com\/ie\/wp-json\/wp\/v2\/comments?post=652821"}],"version-history":[{"count":0,"href":"https:\/\/www.europesays.com\/ie\/wp-json\/wp\/v2\/posts\/652821\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/www.europesays.com\/ie\/wp-json\/wp\/v2\/media\/652822"}],"wp:attachment":[{"href":"https:\/\/www.europesays.com\/ie\/wp-json\/wp\/v2\/media?parent=652821"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.europesays.com\/ie\/wp-json\/wp\/v2\/categories?post=652821"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.europesays.com\/ie\/wp-json\/wp\/v2\/tags?post=652821"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}<\/a>Idealized deep learning workflow using historic foundations as an example archaeological object. Credit: K. Peck et al., Advances in Archaeological Practice (2026).<\/p>\n<\/a>(a) Simulated object placement exclusion model on a 256 \u00d7 256-pixel tile, with black areas showing suitable locations for object placement; (b) simulated objects placed on a 256 \u00d7 256-pixel tile, with objects that are too close together removed; (c) collection pit placement rules; and (d) the script generating a tar kiln perimeter (II) around each placed point (I) as a circular buffer, adding random noise (III), and generating the tar kiln interior (IV) and collection pit (V). Credit: K. Peck et al., Advances in Archaeological Practice (2026).<\/p>\n