- Unlocking Slot Attention by Changing Optimal Transport Costs. ICML 2023
- CrossSplit: Mitigating Label Noise Memorization through Data Splitting. ICML 2023
- Equivariance with Learned Canonicalization Functions. ICML 2023
- Multiset-equivariant set prediction with approximate implicit differentiation. ICLR 2022
- Better set representations for relational reasoning. NeurIPS 2020
- Deep set prediction networks. NeurIPS 2019
- FSPool: Learning set representations with featurewise sort pooling. ICLR 2020
- Learning Representations of Sets through Optimized Permutations. ICLR 2019
- Learning to count objects in natural images for visual question answering. ICLR 2018
Research Experience
Position: Research Scientist at Samsung - SAIT AI Lab, Montreal
Background
Research interests: deep learning with structured objects (like sets) and their equivariance properties. Currently a research scientist at Samsung - SAIT AI Lab, Montreal, which is located within Mila.