Since the first neural networks, during the 50’s, AI techniques have gained in strength and accuracy, following a chaotic progression, tightly related to the progression of the hardware. With the advent of massively parallel architectures, like GPUs, a new field of research was addressed, that was previously purely theoretical as the researchers did not have access to adequate computing power.
This evolution led to new techniques and new paradigms, especially in Computer Vision. As a regular “customer” of Computer Vision, Intelligent transport systems now integrate these GPUs in order to ensure a better behavior of an autonomous vehicle within a complex environment.
In this talk, Laurent will describe how this evolution took place and how they handle it at Navya.