Final Answer:
The hardware innovation that tailors architecture to meet computation needs on a domain, such as matrix multiplication in machine learning, is known as Domain-Specific Architecture.
Step-by-step explanation:
Domain-Specific Architecture is a hardware innovation designed to optimize computational efficiency for specific tasks within a given domain. In the context of machine learning, tasks like matrix multiplication, which are prevalent in algorithms such as neural networks, can be highly resource-intensive. Domain-Specific Architecture tailors the hardware to excel in these specific computations, offering improved performance and energy efficiency compared to general-purpose architectures. This specialized approach enhances the overall efficiency and speed of computations within targeted domains, contributing to advancements in fields like machine learning.