Dr Relja Arandjelović (Др Реља Аранђеловић)
I am a Senior Research Scientist at DeepMind.
Previously I was a Postdoctoral research assistant at the WILLOW group (INRIA / École Normale Superiéure), working with Josef Sivic,
and a Postdoctoral research assistant at the
Visual Geometry Group, Department of Engineering Science, University of Oxford, working with Prof. Andrew Zisserman. I obtained a PhD in Engineering from the same group in 2013. Before coming to Oxford I studied at Cambridge where I got my BA and MEng degrees.
News 
-
06 Jun 2019
Outstanding Reviewer award at the CVPR 2019 -
28 May 2019
New paper on arXiv:
Object Discovery with a Copy-Pasting GAN -
28 Mar 2019
DeepMind published a blog post, Towards Robust and Verified AI, which includes my papers On the effectiveness of interval bound propagation for training verifiably robust models and Training verified learners with learned verifiers -
21 Jan 2019
Code available for our paper On the effectiveness of interval bound propagation for training verifiably robust models -
12 Dec 2018
Updates to the projects: code and models for our NeurIPS 2018 paper Neighbourhood Consensus Networks, datasets for our ECCV 2018 workshop paper Compact Deep Aggregation for Set Retrieval and the BMVC 2016 paper Faces in Places: Compound Query Retrieval -
07 Dec 2018
Our paper, On the effectiveness of interval bound propagation for training verifiably robust models won the best paper award at the NeurIPS 2018 workshop on Security in Machine Learning -
12 Nov 2018
Paper accepted for oral presentation at the NeurIPS 2018 workshop on Security in Machine Learning:
On the effectiveness of interval bound propagation for training verifiably robust models -
21 Sep 2018
Paper accepted to ACCV 2018:
GhostVLAD for set-based face recognition -
09 Sep 2018
Our paper, Compact Deep Aggregation for Set Retrieval, won the best paper award at the ECCV 2018 workshop on Compact and Efficient Feature Representation and Learning in Computer Vision -
05 Sep 2018
Paper accepted for spotlight presentation at NeurIPS 2018:
Neighbourhood Consensus Networks