Search papers, labs, and topics across Lattice.
This chapter provides a foundational overview of hypercomplex widely linear processing techniques tailored for quaternion machine learning. It covers augmented statistics for quaternion-valued random processes, widely linear models to leverage these statistics, and quaternion calculus/algebra for algorithm development. The chapter also addresses mean square estimation for practical applications and includes illustrative examples to promote understanding and adoption of quaternion-based methods.
Unlock the potential of 3D rotation modeling in machine learning with a practical guide to quaternion-based signal processing techniques.
Numerous attempts have been made to replicate the success of complex-valued algebra in engineering and science to other hypercomplex domains such as quaternions, tessarines, biquaternions, and octonions. Perhaps, none have matched the success of quaternions. The most useful feature of quaternions lies in their ability to model three-dimensional rotations which, in turn, have found various industrial applications such as in aeronautics and computergraphics. Recently, we have witnessed a renaissance of quaternions due to the rise of machine learning. To equip the reader to contribute to this emerging research area, this chapter lays down the foundation for: - augmented statistics for modelling quaternion-valued random processes, - widely linear models to exploit such advanced statistics, - quaternion calculus and algebra for algorithmic derivations, - mean square estimation for practical considerations. For ease of exposure, several examples are offered to facilitate the learning, understanding, and(hopefully) the adoption of this multidimensional domain.
