Davide Modolo

Bio

I am a senior research scientist in the fields of computer vision and machine learning. My research interests lie in developing artificial intelligent systems that are able to reason about the visual world and describe images and videos in complex semantic descriptions the same way humans do. Currently, I am a senior research scientist at Amazon.com, where I have been working on two exciting projects: Amazon Go and Amazon Rekognition.

Prior to joining Amazon, I was a PhD student at the University of Edinburgh under the supervision of Dr. Vittorio Ferrari, a Master student at the University of Amsterdam and a Bachelor student and the University of Padova.

News

• September 2022: Our paper on Semi-supervised Transformers at scale was accepted for publication at NeurIPS 2022.
• June 2022: Our paper on MaCLR: Motion-aware Contrastive learning on representations for videos was accepted for publication at ECCV 2022.
• April 2022: Our paper on SCVRL: Shuffled contrastive video representation learning was accepted for publication at CVPR 2022 Workshop.
• March 2022: Three papers accepted for publications at CVPR 2022:
  • What to look at and where: Semantic and spatial refined transformer for detecting human-object interactions (ORAL)
  • TubeR: Tubelet transformer for video Action detection (ORAL)
  • Hierarchical self-supervised representation learning for movie understanding
• August 2021: Our paper on Selective Feature Compression for Efficient Activity Recognition Inference was accepted for publication at ICCV 2021.
• June 2021:  New paper on MoDist: Motion Distillation for Self-supervised Video Representation Learning was released on arXiv.
• March 2021: Our paper on Multi-Object Tracking with Hallucinated and Unlabeled Videos was accepted for publication at CVPR 2021 Workshop.
• February 2021: Our paper on SiamMOT: Siamese Multi-Object Tracking was accepted for publication at CVPR 2021.
• October 2020: Our paper on Understanding the impact of mistakes on background regions in crowd counting was accepted for publication at WACV 2021.
• September 2020: Our paper on Exploiting weakly supervised visual patterns to learn from partial annotations was accepted for publication at NeurIPS 2020.
• March 2020: Our paper on Combining detection and tracking for human pose estimation in videos was accepted for publication at CVPR 2020, as ORAL.
• August 2019: Our paper on Action recognition with spatial-temporal discriminative filter banks was accepted for publication at ICCV 2019.
• January 2019: New paper on Scale-Aware Attention Network for Crowd Counting was released on arXiv.
• February 2018: Our paper on Objects as context for detecting their semantic parts was accepted for publication at CVPR 2018.
• October 2017: Our paper on Do Semantic Parts Emerge in Convolutional Neural Networks? was accepted for publication at IJCV.
• June 2017: Our paper on Learning Semantic Part-Based Models from Google Images was accepted for publication at TPAMI.