Researchers from the Friedrich Miescher Institute for Biomedical Research and Heidelberg University A paper titled "Surrogate gradients for analog neuromorphic computing" published in PNAS in January 2022 revealed a way to bypass this problem. The team demonstrated that an artificial intelligence algorithm called spike neural networks - which uses the brain's unique communication signals, known as spikes - can learn together with chips how to compensate for device mismatches. This paper represents a significant step towards using AI for simulated neuromorphic computing.