Combining mixture weight pruning and quantization for small-footprint speech recognition | IEEE Conference Publication | IEEE Xplore