Memory Mosaics are networks of associative memories working in concert to achieve a prediction task of interest. Like transformers, memory mosaics possess compositional capabilities and in-context learning capabilities. Unlike transformers, memory mosaics achieve these capabilities in comparatively transparent way ("predictive disentanglement"). We illustrate these capabilities on a toy example and also show that memory mosaics perform as well or better than transformers on medium-scale language modeling tasks.
View on arXiv@article{zhang2025_2405.06394,
title={ Memory Mosaics },
author={ Jianyu Zhang and Niklas Nolte and Ranajoy Sadhukhan and Beidi Chen and Léon Bottou },
journal={arXiv preprint arXiv:2405.06394},
year={ 2025 }
}