1 code implementation • 25 Mar 2023 • Fengyin Lin, Mingkang Li, Da Li, Timothy Hospedales, Yi-Zhe Song, Yonggang Qi
This paper studies the problem of zero-short sketch-based image retrieval (ZS-SBIR), however with two significant differentiators to prior art (i) we tackle all variants (inter-category, intra-category, and cross datasets) of ZS-SBIR with just one network (``everything''), and (ii) we would really like to understand how this sketch-photo matching operates (``explainable'').
no code implementations • 24 Feb 2023 • Ruchika Chavhan, Henry Gouk, Jan Stuehmer, Calum Heggan, Mehrdad Yaghoobi, Timothy Hospedales
Contrastive self-supervised learning methods famously produce high quality transferable representations by learning invariances to different data augmentations.
no code implementations • 23 Feb 2023 • Minyoung Kim, Da Li, Timothy Hospedales
We tackle the domain generalisation (DG) problem by posing it as a domain adaptation (DA) task where we adversarially synthesise the worst-case target domain and adapt a model to that worst-case domain, thereby improving the model's robustness.
no code implementations • 15 Dec 2022 • Royson Lee, Rui Li, Stylianos I. Venieris, Timothy Hospedales, Ferenc Huszár, Nicholas D. Lane
Recent image degradation estimation methods have enabled single-image super-resolution (SR) approaches to better upsample real-world images.
no code implementations • 8 Dec 2022 • Zicheng Liu, Da Li, Javier Fernandez-Marques, Stefanos Laskaridis, Yan Gao, Łukasz Dudziak, Stan Z. Li, Shell Xu Hu, Timothy Hospedales
Federated learning has been predominantly concerned with collaborative training of deep networks from scratch, and especially the many challenges that arise, such as communication cost, robustness to heterogeneous data, and support for diverse device capabilities.
1 code implementation • 4 Oct 2022 • Yongshuo Zong, Yongxin Yang, Timothy Hospedales
In this work, we introduce MEDFAIR, a framework to benchmark the fairness of machine learning models for medical imaging.
1 code implementation • 1 Aug 2022 • Panagiotis Eustratiadis, Henry Gouk, Da Li, Timothy Hospedales
This paper investigates a family of methods for defending against adversarial attacks that owe part of their success to creating a noisy, discontinuous, or otherwise rugged loss landscape that adversaries find difficult to navigate.
no code implementations • 17 Jul 2022 • Ruchika Chavhan, Henry Gouk, Jan Stühmer, Timothy Hospedales
Providing invariances in a given learning task conveys a key inductive bias that can lead to sample-efficient learning and good generalisation, if correctly specified.
no code implementations • 15 Jul 2022 • Ondrej Bohdal, Da Li, Shell Xu Hu, Timothy Hospedales
We study the highly practical but comparatively under-studied problem of latent-domain adaptation, where a source model should be adapted to a target dataset that contains a mixture of unlabelled domain-relevant and domain-irrelevant examples.
no code implementations • 15 Jun 2022 • Adrian El Baz, Ihsan Ullah, Edesio Alcobaça, André C. P. L. F. Carvalho, Hong Chen, Fabio Ferreira, Henry Gouk, Chaoyu Guan, Isabelle Guyon, Timothy Hospedales, Shell Hu, Mike Huisman, Frank Hutter, Zhengying Liu, Felix Mohr, Ekrem Öztürk, Jan N. van Rijn, Haozhe Sun, Xin Wang, Wenwu Zhu
Although deep neural networks are capable of achieving performance superior to humans on various tasks, they are notorious for requiring large amounts of data and computing resources, restricting their success to domains where such resources are available.
1 code implementation • 5 Apr 2022 • Calum Heggan, Sam Budgett, Timothy Hospedales, Mehrdad Yaghoobi
Currently available benchmarks for few-shot learning (machine learning with few training examples) are limited in the domains they cover, primarily focusing on image classification.
Ranked #1 on Few-Shot Audio Classification on NSynth
no code implementations • 1 Feb 2022 • Da Li, Henry Gouk, Timothy Hospedales
However much of the work in general purpose DG is heuristically motivated, as the DG problem is hard to model formally; and recent evaluations have cast doubt on existing methods' practical efficacy -- in particular compared to a well tuned empirical risk minimisation baseline.
no code implementations • 9 Nov 2021 • Minyoung Kim, Timothy Hospedales
In essence, the MAP solution is approximated by the LDA estimate, but to take the GP prior into account, we adopt the prior-norm adjustment to estimate LDA's shared variance parameters, which ensures that the adjusted estimate is consistent with the GP prior.
no code implementations • 26 Oct 2021 • Adrian Bulat, Jean Kossaifi, Sourav Bhattacharya, Yannis Panagakis, Timothy Hospedales, Georgios Tzimiropoulos, Nicholas D Lane, Maja Pantic
We propose defensive tensorization, an adversarial defence technique that leverages a latent high-order factorization of the network.
no code implementations • ICLR 2022 • Hae Beom Lee, Hayeon Lee, Jaewoong Shin, Eunho Yang, Timothy Hospedales, Sung Ju Hwang
Many gradient-based meta-learning methods assume a set of parameters that do not participate in inner-optimization, which can be considered as hyperparameters.
no code implementations • ICLR 2022 • Ayan Das, Yongxin Yang, Timothy Hospedales, Tao Xiang, Yi-Zhe Song
Learning meaningful representations for chirographic drawing data such as sketches, handwriting, and flowcharts is a gateway for understanding and emulating human creative expression.
no code implementations • 29 Sep 2021 • Boyan Gao, Henry Gouk, Yongxin Yang, Timothy Hospedales
We take a different approach, and explore the impact of the ERM loss function on out-of-domain generalisation.
no code implementations • ICLR 2022 • Lucas Deecke, Timothy Hospedales, Hakan Bilen
A fundamental shortcoming of deep neural networks is their specialization to a single task and domain.
no code implementations • 13 Sep 2021 • Miguel Jaques, Martin Asenov, Michael Burke, Timothy Hospedales
This paper introduces V-SysId, a novel method that enables simultaneous keypoint discovery, 3D system identification, and extrinsic camera calibration from an unlabeled video taken from a static camera, using only the family of equations of motion of the object of interest as weak supervision.
1 code implementation • 15 Jul 2021 • Rui Li, Ondrej Bohdal, Rajesh Mishra, Hyeji Kim, Da Li, Nicholas Lane, Timothy Hospedales
We use our MetaCC benchmark to study several aspects of meta-learning, including the impact of task distribution breadth and shift, which can be controlled in the coding problem.
1 code implementation • NeurIPS 2021 • Ondrej Bohdal, Yongxin Yang, Timothy Hospedales
Gradient-based meta-learning and hyperparameter optimization have seen significant progress recently, enabling practical end-to-end training of neural networks together with many hyperparameters.
1 code implementation • 17 Jun 2021 • Ondrej Bohdal, Yongxin Yang, Timothy Hospedales
The problem is especially noticeable when using modern neural networks, for which there is a significant difference between the confidence of the model and the probability of correct prediction.
1 code implementation • CVPR 2021 • Ayan Das, Yongxin Yang, Timothy Hospedales, Tao Xiang, Yi-Zhe Song
Analysis of human sketches in deep learning has advanced immensely through the use of waypoint-sequences rather than raster-graphic representations.
2 code implementations • ICCV 2021 • Xueting Zhang, Debin Meng, Henry Gouk, Timothy Hospedales
Current state-of-the-art few-shot learners focus on developing effective training procedures for feature representations, before using simple, e. g. nearest centroid, classifiers.
Ranked #6 on Few-Shot Image Classification on Meta-Dataset
cross-domain few-shot learning Few-Shot Image Classification
no code implementations • 1 Jan 2021 • Lucas Deecke, Timothy Hospedales, Hakan Bilen
A fundamental shortcoming of deep neural networks is their specialization to a single task and domain.
no code implementations • 9 Dec 2020 • Chenyang Zhao, Timothy Hospedales
In reinforcement learning, domain randomisation is an increasingly popular technique for learning more general policies that are robust to domain-shifts at deployment.
no code implementations • 2 Dec 2020 • Jiechao Guan, Zhiwu Lu, Tao Xiang, Timothy Hospedales
By transferring knowledge learned from seen/previous tasks, meta learning aims to generalize well to unseen/future tasks.
no code implementations • 17 Oct 2020 • Panagiotis Eustratiadis, Henry Gouk, Da Li, Timothy Hospedales
Stochastic Neural Networks (SNNs) that inject noise into their hidden layers have recently been shown to achieve strong robustness against adversarial attacks.
1 code implementation • ECCV 2020 • Kaiyang Zhou, Yongxin Yang, Timothy Hospedales, Tao Xiang
This explicitly increases the diversity of available training domains and leads to a more generalizable model.
Ranked #51 on Domain Generalization on PACS
1 code implementation • 6 Jul 2020 • Ivana Balažević, Carl Allen, Timothy Hospedales
In this work, we propose a probabilistically principled general approach to SSL that considers the distribution over label predictions, for labels of different complexity, from "one-hot" vectors to binary vectors and images.
1 code implementation • ECCV 2020 • Ayan Das, Yongxin Yang, Timothy Hospedales, Tao Xiang, Yi-Zhe Song
The study of neural generative models of human sketches is a fascinating contemporary modeling problem due to the links between sketch image generation and the human drawing process.
2 code implementations • 15 Jun 2020 • Ondrej Bohdal, Yongxin Yang, Timothy Hospedales
In particular, we study the problem of label distillation - creating synthetic labels for a small set of real images, and show it to be more effective than the prior image-based approach to dataset distillation.
no code implementations • 10 Jun 2020 • Carl Allen, Ivana Balažević, Timothy Hospedales
Much progress has been made in semi-supervised learning (SSL) by combining methods that exploit different aspects of the data distribution, e. g. consistency regularisation relies on properties of $p(x)$, whereas entropy minimisation pertains to the label distribution $p(y|x)$.
no code implementations • CVPR 2021 • Miguel Jaques, Michael Burke, Timothy Hospedales
Learning low-dimensional latent state space dynamics models has been a powerful paradigm for enabling vision-based planning and learning for control.
no code implementations • 1 Jun 2020 • Lucas Deecke, Timothy Hospedales, Hakan Bilen
While recent techniques in domain adaptation and multi-domain learning enable the learning of more domain-agnostic features, their success relies on the presence of domain labels, typically requiring manual annotation and careful curation of datasets.
1 code implementation • 26 May 2020 • Shreyank N Gowda, Panagiotis Eustratiadis, Timothy Hospedales, Laura Sevilla-Lara
We treat this as a grouping problem by exploiting object proposals and making a joint inference about grouping over both space and time.
One-shot visual object segmentation reinforcement-learning +5
no code implementations • 13 May 2020 • Stephane Doncieux, Nicolas Bredeche, Léni Le Goff, Benoît Girard, Alexandre Coninx, Olivier Sigaud, Mehdi Khamassi, Natalia Díaz-Rodríguez, David Filliat, Timothy Hospedales, A. Eiben, Richard Duro
Robots are still limited to controlled conditions, that the robot designer knows with enough details to endow the robot with the appropriate models or behaviors.
1 code implementation • 11 May 2020 • Xinwang Liu, En Zhu, Jiyuan Liu, Timothy Hospedales, Yang Wang, Meng Wang
We propose a simple yet effective multiple kernel clustering algorithm, termed simple multiple kernel k-means (SimpleMKKM).
1 code implementation • 11 Apr 2020 • Timothy Hospedales, Antreas Antoniou, Paul Micaelli, Amos Storkey
We survey promising applications and successes of meta-learning such as few-shot learning and reinforcement learning.
no code implementations • ECCV 2020 • Da Li, Timothy Hospedales
Therefore we propose an online shortest-path meta-learning framework that is both computationally tractable and practically effective for improving DA performance.
Meta-Learning Multi-Source Unsupervised Domain Adaptation +1
no code implementations • 3 Apr 2020 • Da Li, Yongxin Yang, Yi-Zhe Song, Timothy Hospedales
In DG this means encountering a sequence of domains and at each step training to maximise performance on the next domain.
Ranked #62 on Domain Generalization on PACS
no code implementations • 12 Mar 2020 • Kaiyang Zhou, Yongxin Yang, Timothy Hospedales, Tao Xiang
This is achieved by having a learning objective formulated to ensure that the generated data can be correctly classified by the label classifier while fooling the domain classifier.
Ranked #49 on Domain Generalization on PACS
no code implementations • CVPR 2020 • Juan-Manuel Perez-Rua, Xiatian Zhu, Timothy Hospedales, Tao Xiang
To this end we propose OpeN-ended Centre nEt (ONCE), a detector designed for incrementally learning to detect novel class objects with few examples.
1 code implementation • ECCV 2020 • Yonggang Li, Guosheng Hu, Yongtao Wang, Timothy Hospedales, Neil M. Robertson, Yongxin Yang
In this paper, we propose Differentiable Automatic Data Augmentation (DADA) which dramatically reduces the cost.
Ranked #13 on Data Augmentation on ImageNet
1 code implementation • 2 Mar 2020 • Marija Jegorova, Joshua Smith, Michael Mistry, Timothy Hospedales
Dynamic System Identification approaches usually heavily rely on the evolutionary and gradient-based optimisation techniques to produce optimal excitation trajectories for determining the physical parameters of robot platforms.
no code implementations • 15 Oct 2019 • Marija Jegorova, Antti Ilari Karjalainen, Jose Vazquez, Timothy Hospedales
High-quality realistic sonar data simulation could be of benefit to multiple applications, including training of human operators for post-mission analysis, as well as tuning and validation of autonomous target recognition (ATR) systems for underwater vehicles.
no code implementations • ICLR 2021 • Carl Allen, Ivana Balažević, Timothy Hospedales
Many models learn representations of knowledge graph data by exploiting its low-rank latent structure, encoding known relations between entities and enabling unknown facts to be inferred.
no code implementations • 25 Sep 2019 • Tianyuan Yu, Yongxin Yang, Da Li, Timothy Hospedales, Tao Xiang
Stochastic neural networks (SNNs) are currently topical, with several paradigms being actively investigated including dropout, Bayesian neural networks, variational information bottleneck (VIB) and noise regularized learning.
no code implementations • 16 Aug 2019 • Xun Xu, Shaogang Gong, Timothy Hospedales
To that end, we relax the common assumption that each individual crowd video instance is only associated with a single crowd attribute.
no code implementations • CVPR 2020 • Jean Kossaifi, Antoine Toisoul, Adrian Bulat, Yannis Panagakis, Timothy Hospedales, Maja Pantic
To alleviate this, one approach is to apply low-rank tensor decompositions to convolution kernels in order to compress the network and reduce its number of parameters.
1 code implementation • ICLR 2020 • Miguel Jaques, Michael Burke, Timothy Hospedales
Our approach significantly outperforms related unsupervised methods in long-term future frame prediction of systems with interacting objects (such as ball-spring or 3-body gravitational systems), due to its ability to build dynamics into the model as an inductive bias.
1 code implementation • NeurIPS 2019 • Ivana Balažević, Carl Allen, Timothy Hospedales
Hyperbolic embeddings have recently gained attention in machine learning due to their ability to represent hierarchical data more accurately and succinctly than their Euclidean analogues.
Ranked #34 on Link Prediction on FB15k-237
no code implementations • 28 Jan 2019 • Carl Allen, Timothy Hospedales
Word embeddings generated by neural network methods such as word2vec (W2V) are well known to exhibit seemingly linear behaviour, e. g. the embeddings of analogy "woman is to queen as man is to king" approximately describe a parallelogram.
no code implementations • 7 Nov 2018 • Marija Jegorova, Stéphane Doncieux, Timothy Hospedales
Leveraging our generative policy network, a robot can sample novel behaviors until it finds one that works for a new environment.
no code implementations • EMNLP 2018 • Tanmoy Mukherjee, Makoto Yamada, Timothy Hospedales
Word translation, or bilingual dictionary induction, is an important capability that impacts many multilingual language processing tasks.
no code implementations • ECCV 2018 • Guosheng Hu, Li Liu, Yang Yuan, Zehao Yu, Yang Hua, Zhihong Zhang, Fumin Shen, Ling Shao, Timothy Hospedales, Neil Robertson, Yongxin Yang
To advance subtle expression recognition, we contribute a Large-scale Subtle Emotions and Mental States in the Wild database (LSEMSW).
no code implementations • ICLR 2018 • Kunkun Pang, Mingzhi Dong, Yang Wu, Timothy Hospedales
In contrast to this body of research, we propose to treat active learning algorithm design as a meta-learning problem and learn the best criterion from data.
no code implementations • NeurIPS 2019 • Carl Allen, Ivana Balažević, Timothy Hospedales
We show that different interactions between PMI vectors reflect semantic word relationships, such as similarity and paraphrasing, that are encoded in low dimensional word embeddings under a suitable projection, theoretically explaining why embeddings of W2V and GloVe work.
no code implementations • CVPR 2018 • Jifei Song, Kaiyue Pang, Yi-Zhe Song, Tao Xiang, Timothy Hospedales
In this paper, we present a novel approach for translating an object photo to a sketch, mimicking the human sketching process.
1 code implementation • 12 Sep 2017 • Guosheng Hu, Yuxin Hu, Kai Yang, Zehao Yu, Flood Sung, Zhihong Zhang, Fei Xie, Jianguo Liu, Neil Robertson, Timothy Hospedales, Qiangwei Miemie
We propose a novel investment decision strategy (IDS) based on deep learning.
no code implementations • 29 Jun 2017 • Flood Sung, Li Zhang, Tao Xiang, Timothy Hospedales, Yongxin Yang
We propose a novel and flexible approach to meta-learning for learning-to-learn from only a few examples.
2 code implementations • 20 May 2016 • Yongxin Yang, Timothy Hospedales
Our approach applies to both homogeneous and heterogeneous MTL.
no code implementations • 21 Mar 2016 • Guosheng Hu, Xiaojiang Peng, Yongxin Yang, Timothy Hospedales, Jakob Verbeek
To train such networks, very large training sets are needed with millions of labeled images.
no code implementations • 13 Nov 2015 • Xun Xu, Timothy Hospedales, Shaogang Gong
This is a more challenging problem than existing ZSL of still images and/or attributes, because the mapping between video spacetime features of actions and the semantic space is more complex and harder to learn for the purpose of generalising over any cross-category domain shift.
no code implementations • 9 Oct 2015 • Yi Li, Yi-Zhe Song, Timothy Hospedales, Shaogang Gong
We present a generative model which can automatically summarize the stroke composition of free-hand sketches of a given category.
no code implementations • 28 Jul 2015 • Yongxin Yang, Timothy Hospedales
Most visual recognition methods implicitly assume the data distribution remains unchanged from training to testing.
no code implementations • 27 Jul 2015 • Xun Xu, Timothy Hospedales, Shaogang Gong
The growing rate of public space CCTV installations has generated a need for automated methods for exploiting video surveillance data including scene understanding, query, behaviour annotation and summarization.
no code implementations • CVPR 2015 • Yonggang Qi, Yi-Zhe Song, Tao Xiang, Honggang Zhang, Timothy Hospedales, Yi Li, Jun Guo
We propose a perceptual grouping framework that organizes image edges into meaningful structures and demonstrate its usefulness on various computer vision tasks.
no code implementations • 9 Apr 2015 • Guosheng Hu, Yongxin Yang, Dong Yi, Josef Kittler, William Christmas, Stan Z. Li, Timothy Hospedales
In this work, we conduct an extensive evaluation of CNN-based face recognition systems (CNN-FRS) on a common ground to make our work easily reproducible.
no code implementations • 5 Feb 2015 • Xun Xu, Timothy Hospedales, Shaogang Gong
In this framework a mapping is constructed between visual features and a human interpretable semantic description of each category, allowing categories to be recognised in the absence of any training data.
Ranked #24 on Zero-Shot Action Recognition on UCF101
2 code implementations • 30 Jan 2015 • Qian Yu, Yongxin Yang, Yi-Zhe Song, Tao Xiang, Timothy Hospedales
We propose a multi-scale multi-channel deep neural network framework that, for the first time, yields sketch recognition performance surpassing that of humans.
no code implementations • 17 Sep 2014 • Shuxin Ouyang, Timothy Hospedales, Yi-Zhe Song, Xueming Li
Heterogeneous face recognition (HFR) refers to matching face imagery across different domains.