Deliverable 4.2 - Standard ML Tasks

This report explores the implementation and evaluation of a n-ary AI inference framework using multistate magnetic tunnel junctions (M²TJs) as core components in analog in-memory computing (IMC) architectures for neural network inference. As traditional von Neumann architectures face limitations due to the memory-computation separation, IMC offers a promising alternative by performing matrix-vector multiplications (MVMs) directly within memory arrays, thereby reducing data transfer overhead and enhancing energy efficiency.