Add to Favourites
To login click here
This article discusses the need for a model that can handle noisy and partial inputs in associative memory tasks. The authors propose a general approach using a sparse neural code and implement it in a neural network called the sparse quantized Hopfield network (SQHN). They also develop two new memory tasks and show that SQHN outperforms baselines on these tasks and matches or exceeds state-of-the-art on standard associative memory tasks.
