Position: Quantum Kernel Machines Should Move Beyond Scalar-Valued Kernels to Realize Their Potential
A new position paper argues that quantum kernel machines need to evolve beyond simple scalar-valued kernels to unlock their full potential. The authors contend that current scalar-valued approaches fail to leverage quantum resources like entanglement, limiting their advantage over classical methods. They propose a roadmap focusing on more expressive operator-valued kernel frameworks to tackle complex prediction problems and reveal structural dependencies. AI
IMPACT This research suggests a new direction for quantum machine learning, potentially enabling more powerful AI applications by better utilizing quantum computing resources.