We’ve had a lot of trouble with AI lately, so I’ll make an e-review and we’ll write something about it.
AI has 4 basic areas:
1. Machine learning: linear regression, tree models, support vector methods (SVM), clustering (k-means), neural networks, hungry learning (K-NN)… Use in Python, R, Matlab
2. Deep learning: convolutional neural networks (CNN), recurrent neural networks (RNN), generative models (GAN, VAE) Use in TensorFlow, PyTorch, Keras
3. Natural language processing: tokenization/lemmatization, TF-IDF, transformers and BERT, GPT (generative pretrained transformer)
4. Computer vision: convolutional neural networks (CNN), detection algorithms (R-CNN, YOLO), generative adversarial networks (GAN)
Phew, that’s enough, and just getting a general overview and a mathematical foundation is a long time job. So let’s take a deep dive and see how AI and archeology are doing, here are a few projects:
open-archeo – primarilly not about AI but good list of resources (software and data)
ArcheoNet – use of convolutional neural networks
ArchAI – looks modern and has good PR. Mainly detection of structures and recognition of LIDAR images. This video is worth of seeing…
OpenArcheo – AI based analysis of archaeological semantics