Quantization has become a mainstream compression technique for reducing model size,
computational requirements, and energy consumption for modern deep neural networks …

Edge device deep neural network (DNN) training is practical to improve model adaptivity for
unfamiliar datasets while avoiding privacy disclosure and huge communication cost …

Providing high-quality predictions is no longer the sole goal for neural networks. As we live
in an increasingly interconnected world, these models need to match the constraints of …

Recently, low bit-width quantization (eg, INT8) has been commonly used in deep neural
network inference acceleration, but fewer researchers have focused on low-precision …

This paper presents an energy-efficient FP4/FP8 hybrid-precision training processor.
Through hardware-software co-optimization, the design efficiently implements all general …

Quantization-aware training (QAT) receives extensive popularity as it well retains the
performance of quantized networks. In QAT, the contemporary experience is that all …

Continual learning is a desirable feature in many modern machine learning applications,
which allows in-field adaptation and updating, ranging from accommodating distribution …

Abstract As Deep Neural Networks (DNNs) continue to increase in complexity, the
computational demands of their training have become a significant bottleneck. Low …

Quantization-aware training (QAT) has been demonstrated to not only reduce computational
cost and storage footprint, but well retain the performance of full-precision neural networks …

Foundational to the digital economy, data and its accompanying analytical models today
play a critical role in the gaining of new insights from the digital world and the facilitation of …