Computational archaeology

Computational archaeology may include the use of geographical information systems (GIS), especially when applied to spatial analyses such as viewshed analysis and least-cost path analysis as these approaches are sufficiently computationally complex that they are extremely difficult if not impossible to implement without the processing power of a computer. Likewise, some forms of statistical and mathematical modelling, and the computer simulation of human behaviour and behavioural evolution using software tools such as Swarm or Repast would also be impossible to calculate without computational aid. The application of a variety of other forms of complex and bespoke software to solve archaeological problems, such as human perception and movement within built environments using software such as University College London's Space Syntax program, also falls under the term 'computational archaeology'. Other examples of computational archaeology include semantic approach towards machine learning, such as data ontology or the CIDOC Conceptual Reference Model, used in the British Museum's ResearchSpace, Arches, and the Global Rock Art Database.

The acquisition, documentation and analysis of archaeological finds at excavations and in museums is an important field having pottery analysis as one of the major topics. In this area 3D-acquisition techniques like structured light scanning (SLS), photogrammetric methods like "structure from motion" (SfM), computed tomography as well as their combinations provide large data-sets of numerous objects for digital pottery research. These techniques are increasingly integrated into the in-situ workflow of excavations. The Austrian subproject of the Corpus vasorum antiquorum (CVA) is seminal for digital research on finds within museums.

Computational archaeology is also known as "archaeological informatics" (Burenhult 2002, Huggett and Ross 2004) or "archaeoinformatics" (sometimes abbreviated as "AI", but not to be confused with artificial intelligence).

Origins and objectives

In recent years, it has become clear that archaeologists will only be able to harvest the full potential of quantitative methods and computer technology if they become aware of the specific pitfalls and potentials inherent in the archaeological data and research process. AI science is an emerging discipline that attempts to uncover, quantitatively represent and explore specific properties and patterns of archaeological information. Fundamental research on data and methods for a self-sufficient archaeological approach to information processing produces quantitative methods and computer software specifically geared towards archaeological problem solving and understanding.

AI science is capable of complementing and enhancing almost any area of scientific archaeological research. It incorporates a large part of the methods and theories developed in quantitative archaeology since the 1960s but goes beyond former attempts at quantifying archaeology by exploring ways to represent general archaeological information and problem structures as computer algorithms and data structures. This opens archaeological analysis to a wide range of computer-based information processing methods fit to solve problems of great complexity. It also promotes a formalized understanding of the discipline's research objects and creates links between archaeology and other quantitative disciplines, both in methods and software technology. Its agenda can be split up in two major research themes that complement each other:

1. Fundamental research (theoretical AI science) on the structure, properties and possibilities of archaeological data, inference and knowledge building. This includes modeling and managing fuzziness and uncertainty in archaeological data, scale effects, optimal sampling strategies and spatio-temporal effects.
2. Development of computer algorithms and software (applied AI science) that make this theoretical knowledge available to the user.

There is already a large body of literature on the use of quantitative methods and computer-based analysis in archaeology. The development of methods and applications is best reflected in the annual publications of the CAA conference (see external links section at bottom). At least two journals, the Italian Archeologia e Calcolatori and the British Archaeological Computing Newsletter, are dedicated to archaeological computing methods. AI Science contributes to many fundamental research topics, including but not limited to:

• advanced statistics in archaeology, spatial and temporal archaeological data analysis
• bayesian analysis and advanced probability models, fuzziness and uncertainty in archaeological data
• scale-related phenomena and scale transgressions
• intrasite analysis (representations of stratigraphy, 3D analysis, artefact distributions)
• landscape analysis (territorial modeling, visibility analysis)
• optimal survey and sampling strategies
• process-based modeling and simulation models
• archaeological predictive modeling and heritage management applications
• supervised and unsupervised classification and typology, artificial intelligence applications
• digital excavations and virtual reality
• computational reproducibility of archaeological research
• archaeological software development, electronic data sharing and publishing

AI science advocates a formalized approach to archaeological inference and knowledge building. It is interdisciplinary in nature, borrowing, adapting and enhancing method and theory from numerous other disciplines such as computer science (e.g. algorithm and software design, database design and theory), geoinformation science (spatial statistics and modeling, geographic information systems), artificial intelligence research (supervised classification, fuzzy logic), ecology (point pattern analysis), applied mathematics (graph theory, probability theory) and statistics.