Edge computing and artificial intelligence (AI), especially deep learning algorithms, are
gradually intersecting to build the novel system, namely edge intelligence. However, the …

While neural networks have advanced the frontiers in many applications, they often come at
a high computational cost. Reducing the power and latency of neural network inference is …

Neural network pruning and quantization techniques are almost as old as neural networks
themselves. However, to date, only ad-hoc comparisons between the two have been …

When quantizing neural networks, assigning each floating-point weight to its nearest fixed-
point value is the predominant approach. We find that, perhaps surprisingly, this is not the …

When training neural networks with simulated quantization, we observe that quantized
weights can, rather unexpectedly, oscillate between two grid-points. The importance of this …

Transformer-based architectures have become the de-facto standard models for a wide
range of Natural Language Processing tasks. However, their memory footprint and high …

In this paper, we propose a number of novel techniques and numerical representation
formats that enable, for the very first time, the precision of training systems to be aggressively …

Mixed-precision Deep Neural Networks (DNNs) provide an efficient solution for hardware
deployment, especially under resource constraints, while maintaining model accuracy …

When quantizing neural networks for efficient inference, low-bit integers are the go-to format
for efficiency. However, low-bit floating point numbers have an extra degree of freedom …

Reducing the numerical precision of data and computation is extremely effective in
accelerating deep learning training workloads. Towards this end, 8-bit floating point …