)
David Pattison
Machine Learning Engineer,
Agent Oriented Software
Accurate Automated Identification of Noxious Weed Species
Weeds cost the Australian agriculture sector $4.8 billion each year. The biggest challenge to targeted, automated intervention is precise identification and localisation. This session shares the deep learning techniques utilised to achieve accurate identification whilst being invariant to a range of environmental factors.
About David:
David Pattison graduated from the University of Cambridge with a degree in Information and Computer Engineering. Under the guidance of Adrian Weller, director of machine learning research, he completed a research project focused on employing deep learning techniques to greatly reduce the duration of MRI scans. As a machine learning engineer at AOS, David currently works on novel deep learning architectures for autonomous navigation in unmapped off-road environments as well as the identification and localisation of invasive weed species. These technologies constitute part of AOS’s capabilities to use automated vehicles to precisely apply herbicide to noxious weeds, improving pasture productivity. David’s areas of interest include novel convolutional neural network architectures, data-efficient machine learning practices, reinforcement learning and sustainable agriculture.
Weeds cost the Australian agriculture sector $4.8 billion each year. The biggest challenge to targeted, automated intervention is precise identification and localisation. This session shares the deep learning techniques utilised to achieve accurate identification whilst being invariant to a range of environmental factors.
About David:
David Pattison graduated from the University of Cambridge with a degree in Information and Computer Engineering. Under the guidance of Adrian Weller, director of machine learning research, he completed a research project focused on employing deep learning techniques to greatly reduce the duration of MRI scans. As a machine learning engineer at AOS, David currently works on novel deep learning architectures for autonomous navigation in unmapped off-road environments as well as the identification and localisation of invasive weed species. These technologies constitute part of AOS’s capabilities to use automated vehicles to precisely apply herbicide to noxious weeds, improving pasture productivity. David’s areas of interest include novel convolutional neural network architectures, data-efficient machine learning practices, reinforcement learning and sustainable agriculture.
Sessions