More details can be found in the slides that I will unload to indico
•Two published papers:
•One paper on accelerating temporal graphs, Learning on graphs’22 best paper award
•Semi-supervised GNN for pileup mitigation, the European Physics Journal C
•One top-1 rank at 3D object detection benchmark: BEVFusion
•Submitted two papers:
•FlatFormer, an efficient point cloud transformer model, submitted to CVPR’22
•Interpretable Point cloud models, submitted to ICLR’23
•On-going works:
•FlatFormer, further improvements
•Point-cloud OOD benchmarks
•Graph OOD/ domain adaptation paper to be submitted to ICML’23