EAA2020: Session #464

There has been much hype on the use of applicability of Machine Learning (ML) and Artificial Intelligence (AI). However, its fundamental use is to extract complex patterns which humans do no necessarily see. These methods, especially when multiple are stacked, provide the possibility to highlight patterns previously unknown to the observer. Multi-model approaches, predominantly used in finance and to provide consumers with personalised suggestions are especially powerful in large and fuzzy datasets and in combination with an expert’s opinion reduce the amount of theory based assumptions and approximations within a model.
Whilst providing explanatory insights and subtle new ideas their current framework has to be re-evaluated to harness their power for research in the fields of humanities. Their potential, especially in archaeological research is tremendous and has been explored, e.g. to re-assemble 3D pottery shreds into an object, visual recognition of script or coins, and other object based approaches.
In our session we aim to discuss how ML and AI model outcomes can be used to negotiate different narratives of the past and how to design a framework, to ensure reproducible and documented research. We particularly welcome theoretical and methodical papers as well as case studies.

social theory, artificial intelligence, methodology

 Session #464 (PDF)

Chiara Girotto¹

Henry Price²

Machine learning approaches for optimized semi-automatic analysis of ancient grains

Bringing Machine Learning to Taphonomy. Identifying carnivore tooth marks in bone surface with ML algorithms: crocodiles and wolves.

Patterns of Trauma - Using AI to distinguish interpersonal violence from accidental injury

Difficulties tracing and interpreting patterns in compositional data of metal artefacts. Why are the more complex methods not always useful?

The seven deadly sins of modelling

Reality is what I recognise? Choice of factors and ML in social archaeology

More than the eye can see - Machine Learning and rock art