- arXivLong Expressive Memory for Sequence ModelingPreprint
- SISCHigher-order Quasi-Monte Carlo Training of Deep Neural NetworksTo appear in SIAM Journal on Scientific Computing.
- ICMLUnICORNN: A recurrent model for learning very long time dependenciesThe 38th International Conference on Machine Learning.
- ICLR OralCoupled Oscillatory Recurrent Neural Network (coRNN): An accurate and (gradient) stable architecture for learning long time dependenciesThe 9th International Conference on Learning Representations.Oral Presentation (top 1% of all submitted papers)
- SINUMEnhancing accuracy of deep learning algorithms by training with low-discrepancy sequencesSIAM Journal on Numerical Analysis.
- Turbo ExpoReproducing Existing Nacelle Geometries With the Free-Form Deformation ParametrizationASME Turbo Expo: Power for Land, Sea, and Air.