Neural networks have been responsible for most of the top-performing AI systems of the past decade, but they tend to be big, which means they tend to be slow. That’s a problem for systems like Alexa, which depend on neural networks to process spoken requests in real time. In natural-language-understanding (NLU) applications, most of a neural network’s size comes from a huge lookup table that corre
![New method for compressing neural networks better preserves accuracy](https://cdn-ak-scissors.b.st-hatena.com/image/square/ca16d3ca2e9164146f4a2a0829538a2df8a7f9d3/height=288;version=1;width=512/https%3A%2F%2Fassets.amazon.science%2Fdims4%2Fdefault%2Fce84994%2F2147483647%2Fstrip%2Ftrue%2Fcrop%2F1200x630%2B0%2B0%2Fresize%2F1200x630%21%2Fquality%2F90%2F%3Furl%3Dhttp%253A%252F%252Famazon-topics-brightspot.s3.amazonaws.com%252Fscience%252F32%252F80%252Fc230480c4f60a534bc077755bae7%252Famazon-science-og-image-squid.png)