research ADELIA Automatic Differentiation for Efficient Laplace Inference Approximations DaCe AD Unifying high-performance automatic differentiation for ML and scientific computing ICON acceleration 2025 ACM Gordon Bell Prize for Climate Modelling PerfDojo Automated ML library generation for heterogeneous architectures LOOPer A learned automatic code optimizer for polyhedral compilers dataset LOOPerSet A large-scale dataset for data-driven polyhedral compiler optimization