no code implementations • 24 May 2023 • Halil Ibrahim Aysel, Xiaohao Cai, Adam Prügel-Bennett
Utilising semantic proportions suggested in this work offers a promising direction for future research in the field of semantic segmentation.
no code implementations • 15 Feb 2022 • Dominic Belcher, Antonia Marcu, Adam Prügel-Bennett
In this paper we show that the expected generalisation performance of a learning machine is determined by the distribution of risks or equivalently its logarithm -- a quantity we term the risk entropy -- and the fluctuations in a quantity we call the training ratio.
1 code implementation • 14 Feb 2022 • Mark Tuddenham, Adam Prügel-Bennett, Jonathan Hare
The optimisation of neural networks can be sped up by orthogonalising the gradients before the optimisation step, ensuring the diversification of the learned representations.
no code implementations • 22 Nov 2021 • Antonia Marcu, Adam Prügel-Bennett
The community lacks theory-informed guidelines for building good data sets.
1 code implementation • NeurIPS 2021 • Antonia Marcu, Adam Prügel-Bennett
Data distortion is commonly applied in vision models during both training (e. g methods like MixUp and CutMix) and evaluation (e. g. shape-texture bias and robustness).
no code implementations • 13 Aug 2021 • Takaki Yamada, Adam Prügel-Bennett, Stefan B. Williams, Oscar Pizarro, Blair Thornton
We demonstrate how the latent representations generated by GeoCLR can be used to efficiently guide human annotation efforts, where the semi-supervised framework improves classification accuracy by an average of 10. 2% compared to the state-of-the-art SimCLR using the same CNN and equivalent number of human annotations for training.
no code implementations • 26 Jul 2021 • Yue Jiao, Jonathon Hare, Adam Prügel-Bennett
We find that contextual representations in language mod-els outperform static word embeddings, when the compositional chain of object is short.
no code implementations • 26 Jul 2021 • Yue Jiao, Jonathon Hare, Adam Prügel-Bennett
Although different paradigms of visual semantic embedding models are designed to align visual features and distributed word representations, it is unclear to what extent current ZSL models encode semantic information from distributed word representations.
no code implementations • 2 May 2021 • Vlad Velici, Adam Prügel-Bennett
This report presents the application of object detection on a database of underwater images of different species of crabs, as well as aerial images of sea lions and finally the Pascal VOC dataset.
no code implementations • 1 May 2021 • Vlad Velici, Adam Prügel-Bennett
Long Short-Term Memory (LSTM) units have the ability to memorise and use long-term dependencies between inputs to generate predictions on time series data.
no code implementations • 28 Nov 2020 • Mark Tuddenham, Adam Prügel-Bennett, Jonathan Hare
Classification problems using deep learning have been shown to have a high-curvature subspace in the loss landscape equal in dimension to the number of classes.
5 code implementations • 27 Feb 2020 • Ethan Harris, Antonia Marcu, Matthew Painter, Mahesan Niranjan, Adam Prügel-Bennett, Jonathon Hare
Finally, we show that a consequence of the difference between interpolating MSDA such as MixUp and masking MSDA such as FMix is that the two can be combined to improve performance even further.
Ranked #3 on Image Classification on Fashion-MNIST
no code implementations • 11 Nov 2019 • Antonia Marcu, Adam Prügel-Bennett
In this paper, a new approach to computing the generalisation performance is presented that assumes the distribution of risks, $\rho(r)$, for a learning scenario is known.
no code implementations • 25 Sep 2019 • Zezhen Zeng, Jonathon Hare, Adam Prügel-Bennett
Variational Auto-Encoders (VAEs) are designed to capture compressible information about a dataset.
1 code implementation • NeurIPS 2019 • Yan Zhang, Jonathon Hare, Adam Prügel-Bennett
Current approaches for predicting sets from feature vectors ignore the unordered nature of sets and suffer from discontinuity issues as a result.
2 code implementations • ICLR 2020 • Yan Zhang, Jonathon Hare, Adam Prügel-Bennett
Traditional set prediction models can struggle with simple datasets due to an issue we call the responsibility problem.
no code implementations • ICLR 2019 • Yue Jiao, Jonathon Hare, Adam Prügel-Bennett
We present an extension of a variational auto-encoder that creates semantically richcoupled probabilistic latent representations that capture the semantics of multiplemodalities of data.
2 code implementations • ICLR 2019 • Yan Zhang, Jonathon Hare, Adam Prügel-Bennett
Representations of sets are challenging to learn because operations on sets should be permutation-invariant.
1 code implementation • ICLR 2018 • Yan Zhang, Jonathon Hare, Adam Prügel-Bennett
Visual Question Answering (VQA) models have struggled with counting objects in natural images so far.
Ranked #32 on Visual Question Answering (VQA) on VQA v2 test-std
no code implementations • ICLR 2018 • Vlad Velici, Adam Prügel-Bennett
Long Short-Term Memory (LSTM) units have the ability to memorise and use long-term dependencies between inputs to generate predictions on time series data.