Archaeological Research at the Intersection of Physical and Artificial Realities

Summary

This is an abstract from the "Thinking of Acronyms: a Kohler Obsession? Papers in Honor of Timothy A. Kohler (TAKO)" session, at the 90th annual meeting of the Society for American Archaeology.

The proliferation of artificial intelligence coupled with the accessibility of consumer-level computing equipment that can analyze big data has heralded a new paradigm of research in the hard and social sciences. While archaeology is often reticent to broadly adopt the newest technologies, a suite of recent publications highlights the effort researchers are making to rapidly explore the potential applications of unsupervised machine learning as a means of knowledge production. Unsupervised machine learning, by design, independently manages how a model understands, weighs, and analyzes the suite of information it is provided. The broad application of unsupervised machine learning algorithms to legacy archaeological and other datasets underscores the importance of coupling on-the-ground fieldwork efforts in the physical world to effectively understand the information being input to a model, the ability to verify the output, and to understand the interpretive value of the artificially derived results. This paper presents results from an unsupervised machine learning model to identify road features in lidar datasets across the northern U.S. Southwest, an initial fieldwork effort to verify the results, and explores the interpretive value of a regional dataset identified through the computational power of artificial intelligence.

This Resource is Part of the Following Collections

• Society for American Archaeology 90th Annual Meeting, Denver, CO (2025)   •
• Thinking of Acronyms: a Kohler Obsession? Papers in Honor of Timothy A. Kohler (TAKO)

Keywords

Record Identifiers

Abstract Id(s): 50476