no code implementations • 24 Oct 2022 • Tzu-Jui Julius Wang, Jorma Laaksonen, Tomas Langer, Heikki Arponen, Tom E. Bishop
Moreover, in other V-L downstream tasks considered, our WFH models are on par with models trained with paired V-L data, revealing the utility of unpaired data.
no code implementations • 11 May 2020 • Heikki Arponen, Tom E. Bishop
We address (ii) via a differentiable estimate of the KL divergence between network outputs and a binary target distribution, resulting in minimal information loss when the features are rounded to binary.
no code implementations • 12 Aug 2019 • Heikki Arponen, Tom E. Bishop
Using class labels to represent class similarity is a typical approach to training deep hashing systems for retrieval; samples from the same or different classes take binary 1 or 0 similarity values.
no code implementations • ICLR Workshop LLD 2019 • Heikki Arponen, Tom E Bishop
Using class labels to represent class similarity is a typical approach to training deep hashing systems for retrieval; samples from the same or different classes take binary 1 or 0 similarity values.
no code implementations • 6 Sep 2017 • Heikki Arponen, Matti Herranen, Harri Valpola
We prove an exact relationship between the optimal denoising function and the data distribution in the case of additive Gaussian noise, showing that denoising implicitly models the structure of data allowing it to be exploited in the unsupervised learning of representations.