Abstract
In-memory computing (IMC) is a paradigm that enables neural network inference by computing analog matrix-vector multiplications (MVM) directly in memory crossbar arrays, with the potential for energy efficiency gains over conventional von Neumann architectures. This work presents a simulation framework for N-ary crossbar architectures that retrieves MVM results with minimal implementation assumptions.
We successfully inferred XOR and MNIST classification tasks using a simulated crossbar array of (4x4) 4-states magnetic tunnel junctions (MTJ). The MNIST accuracy reached 93.56% (vs. 97.56% software baseline). PCA dimensionality reduction was shown to drastically lower the number of required operations and improve the software baseline, for only a modest reduction in crossbar inference accuracy.
We identified weight quantization as the primary error source and studied its impact alongside systematic non-idealities and random noise. We find that cell-specific random noise is less detrimental than systematic errors due to averaging across the array. Finally, we demonstrate an optimal number of states per cell that balances quantization error against resistance state resolution to minimize total MVM error.
Blogger's Review: This paper provides an in-depth exploration of the effects of multibit quantization on neural network inference in in-memory computing architectures. The experimental data and theoretical support highlight the potential for future advancements in efficiency and accuracy. The application of PCA dimensionality reduction also presents a novel approach for efficient inference.