A Proof of Concept for Machine Learning-Based Virtual Knapping Using Neural Networks

Abstract

Prehistoric stone tools are an important source of evidence for the study of human behavioural and cognitive evolution. Archaeologists use insights from the experimental replication of stone tools to understand things such as the behaviours and cognitive capacities required to manufacture them. However, such experiments can require large amounts of time and raw materials, and achieving sufficient control of key knapping variables can be difficult. A computer program able to accurately simulate stone tool production would make lithic experimentation faster and more accessible, and may lead to reliable insights into the factors that structure the archaeological record. We present here a proof of concept for a machine learning-based virtual knapping framework capable of quickly and accurately predicting flake removals from 3D cores using a conditional adversarial neural network (CGAN). We programmatically generated a testing dataset of standardised 3D cores with flakes knapped from them. After training, the CGAN accurately predicted the length, volume, width, and shape of these flake removals using the core surface information alone. This demonstrates the feasibility of machine learning for investigating stone tool production virtually. With an increased training sample and validation against archaeological data, virtual knapping could enable fast, cheap, and highly-repeatable virtual lithic experimentation.