Learning Active Multimodal Subspaces in the Brain

Here we introduce a multimodal framework to identify subspaces in the human brain that are defined by collective changes in structural and functional measures and are actively linked to demographic, biological and cognitive indicators in a population. We determine the multimodal subspaces using principles of active subspace learning (ASL) and demonstrate its application on a sample learning task (biological ageing) on a Schizophrenia dataset. The proposed multimodal ASL method successfully identifies latent brain representations as subsets of brain regions and connections forming co-varying subspaces in association with biological age. We show that Schizophrenia is characterized by different subspace patterns compared to those in a cognitively normal brain. The multimodal features generated by projecting structural and functional MRI components onto these active subspaces perform better than several PCA-based transformations and equally well when compared to non-transformed features on the studied learning task. In essence, the proposed method successfully learns active brain subspaces associated with a specific brain condition but inferred from the brain imaging data along with the biological/cognitive traits of interest.

Global Optimisation of a Transonic Fan Blade Through AI-Enabled Active Subspaces

The increased need to design higher performing aerodynamic shapes has led to design optimisation cycles requiring high-fidelity CFD models and high-dimensional parametrisation schemes. The computational cost of employing global search algorithms on such scenarios has typically been prohibitive for most academic and industrial environments. In this paper, a novel strategy is presented that leverages the capabilities of Artificial Neural Networks for regressing complex unstructured data, while coupling them with dimensionality reduction algorithms. This approach enables employing global-based optimisation methods on high-dimensional applications through a reduced computational cost. This methodology is demonstrated on the efficiency optimisation of a modern jet engine fan blade with constrained pressure ratio. The outcome is compared against a state-of-the-art adjoint-based approach. Results indicate the strategy proposed achieves comparable improvements to its adjoint counterpart with a reduced computational cost, and can scale better to multi-objective optimisation applications.

Global Optimisation of a Transonic Fan Blade Through AI-Enabled Active Subspaces

The increased need to design higher performing aerodynamic shapes has led to design optimisation cycles requiring high-fidelity CFD models and high-dimensional parametrisation schemes. The computational cost of employing global search algorithms on such scenarios has typically been prohibitive for most academic and industrial environments. In this paper, a novel strategy is presented that leverages the capabilities of Artificial Neural Networks for regressing complex unstructured data, while coupling them with dimensionality reduction algorithms. This approach enables employing global-based optimisation methods on high-dimensional applications through a reduced computational cost. This methodology is demonstrated on the efficiency optimisation of a modern jet engine fan blade with constrained pressure ratio. The outcome is compared against a state-of-the-art adjoint-based approach. Results indicate the strategy proposed achieves comparable improvements to its adjoint counterpart with a reduced computational cost, and can scale better to multi-objective optimisation applications.