Robust quantum gates using smooth pulses and physics-informed neural networks

The presence of decoherence in quantum computers necessitates the suppression of noise. Dynamically corrected gates via specially designed control pulses offer a path forward, but hardware-specific experimental constraints can cause complications. Here, we present a widely applicable method for obtaining robust smooth pulses which is not based on a sampling approach and does not need any assumptions with regards to the underlying statistics of the experimental noise for both quasistatic and broadband noise. We demonstrate the capability of our approach in the context of spin qubits and transmon by finding smooth shapes which suppress the effects of noise within the logical subspace as well as leakage out of that subspace.