Arash Vahdat

I am a senior research scientist at NVIDIA research specializing in machine learning and computer vision. Prior to NVIDIA, I was a senior research scientist at D-Wave Systems Inc. where I worked on deep generative learning with discrete latent variable models (ICML’18NIPS’18arXiv’18, arXiv’19) and weakly supervised learning from noisy labels (NIPS’17ECCV’14, ICCV’13a). Prior to D-Wave, I was a PhD student under Greg Mori’s supervision working on latent variable frameworks for visual analysis (ICCV’13bNIPS’13NIPS’12, ECCV’12), then a research faculty member at Simon Fraser University, where I led several research projects at the intersection of deep learning and video analysis (CVPR’16aCVPR’16bICPR’16). I grew up in Iran and completed my undergrad in computer engineering at Sharif University of Technology.

At D-Wave, I solely developed Akros the robust image labeling framework that was presented at the main NIPS conference in 2017. In Dec 2017, a variant of this framework won the CATARACTS surgical video analysis challenge in collaboration with Siemense Healthinears. This framework later became one of the founding services for, D-Wave’s machine learning business unit. Akros later was extended to object detection (ICCV’19) and semantic object segmentation (CVPR’20).

At D-Wave, I also developed new frameworks for training deep generative models with discrete latent variables. These frameworks, known as discrete variational autoencoders (DVAEs), have pushed the state-of-the-art in this area (DVAE++DVAE#, DVAE##), and have expanded our understanding of the existing frameworks (arXiv’18).

At NVIDIA, I am working on a broad range of problems including neural architecture search, discrete latent variable models, representation learning and deep generative learning.

E-mail: avahdat [at] sfu [dot] ca
Google Scholar


  • Area Chair:
    • ICLR
  •  Reviewer:
    • NeurIPS, ICML
    • PAMI, Pattern Recognition

Research Interests

  • Neural architecture search
  • Representation learning
  • Probabilistic deep learning, generative learning, variational inference, energy-based models
  • Semi-supervised learning, structured latent variable models
  • Training from noisy labels
  • Human-in-the-loop machine learning