Research Manuscript: An Approximate World from Neurons to Genomes
Event TypeResearch Manuscript
Approximate Computing for AI/ML
TimeWednesday, December 8th10:30am - 12:00pm PST
DescriptionThe session features recent advancements in applying approximate computing techniques to different computing paradigms pertaining to Artificial Intelligence and Machine Learning, including hyperdimensional systems, spiking neural networks, probabilistic programming, and compressed video analytics. The techniques proposed include new encoding schemes, fixed-point quantization approaches, training methodologies, and compiler frameworks. The techniques have demonstrated significant improvement in both energy consumption and performance across various end-applications.