Approximate Floating-Point Operations with Integer Units by Processing in the Logarithmic Domain | IEEE Conference Publication | IEEE Xplore