no code implementations • 3 Oct 2023 • Samyadeep Basu, Mehrdad Saberi, Shweta Bhardwaj, Atoosa Malemir Chegini, Daniela Massiceti, Maziar Sanjabi, Shell Xu Hu, Soheil Feizi
From both the human study and automated evaluation, we find that: (i) Instruct-Pix2Pix, Null-Text and SINE are the top-performing methods averaged across different edit types, however {\it only} Instruct-Pix2Pix and Null-Text are able to preserve original image properties; (ii) Most of the editing methods fail at edits involving spatial operations (e. g., changing the position of an object).
no code implementations • 18 Jul 2023 • Samyadeep Basu, Maziar Sanjabi, Daniela Massiceti, Shell Xu Hu, Soheil Feizi
On the challenging Winoground compositional reasoning benchmark, our method improves the absolute visio-linguistic performance of different CLIP models by up to 7%, while on the ARO dataset, our method improves the visio-linguistic performance by upto 3%.
no code implementations • 4 Apr 2023 • Samyadeep Basu, Daniela Massiceti, Shell Xu Hu, Soheil Feizi
Through our controlled empirical study, we have two main findings: (i) Fine-tuning just the LayerNorm parameters (which we call LN-Tune) during few-shot adaptation is an extremely strong baseline across ViTs pre-trained with both self-supervised and supervised objectives, (ii) For self-supervised ViTs, we find that simply learning a set of scaling parameters for each attention matrix (which we call AttnScale) along with a domain-residual adapter (DRA) module leads to state-of-the-art performance (while being $\sim\!$ 9$\times$ more parameter-efficient) on MD.
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.
no code implementations • 1 Sep 2022 • Yangtao Wang, Xi Shen, Yuan Yuan, Yuming Du, Maomao Li, Shell Xu Hu, James L Crowley, Dominique Vaufreydaz
This method also achieves competitive results for unsupervised video object segmentation tasks with the DAVIS, SegTV2, and FBMS datasets.
Ranked #4 on Unsupervised Instance Segmentation on COCO val2017
no code implementations • 15 Jul 2022 • Ondrej Bohdal, Da Li, Shell Xu Hu, Timothy Hospedales
Recognizing that device's data are likely to come from multiple latent domains that include a mixture of unlabelled domain-relevant and domain-irrelevant examples, we focus on the comparatively under-studied problem of latent domain adaptation.
1 code implementation • 27 Jun 2022 • Yingyi Chen, Shell Xu Hu, Xi Shen, Chunrong Ai, Johan A. K. Suykens
This decomposition provides three insights: (i) it shows that over-fitting is indeed an issue for learning with noisy labels; (ii) through an information bottleneck formulation, it explains why the proposed feature compression helps in combating label noise; (iii) it gives explanations on the performance boost brought by incorporating compression regularization into Co-teaching.
Ranked #10 on Image Classification on Clothing1M (using extra training data)
no code implementations • 10 Jun 2022 • Minyoung Kim, Da Li, Shell Xu Hu, Timothy M. Hospedales
Recent sharpness-aware minimisation (SAM) is known to find flat minima which is beneficial for better generalisation with improved robustness.
1 code implementation • CVPR 2022 • Shell Xu Hu, Da Li, Jan Stühmer, Minyoung Kim, Timothy M. Hospedales
To this end, we explore few-shot learning from the perspective of neural network architecture, as well as a three stage pipeline of network updates under different data supplies, where unsupervised external data is considered for pre-training, base categories are used to simulate few-shot tasks for meta-training, and the scarcely labelled data of an novel task is taken for fine-tuning.
Ranked #1 on Few-Shot Image Classification on Meta-Dataset
1 code implementation • 28 Apr 2021 • Yingyi Chen, Xi Shen, Shell Xu Hu, Johan A. K. Suykens
On Clothing1M, our approach obtains 74. 9% accuracy which is slightly better than that of DivideMix.
Ranked #12 on Image Classification on Clothing1M (using extra training data)
2 code implementations • ICLR 2020 • Shell Xu Hu, Pablo G. Moreno, Yang Xiao, Xi Shen, Guillaume Obozinski, Neil D. Lawrence, Andreas Damianou
The evidence lower bound of the marginal log-likelihood of empirical Bayes decomposes as a sum of local KL divergences between the variational posterior and the true posterior on the query set of each task.
Ranked #12 on Few-Shot Image Classification on CIFAR-FS 5-way (1-shot)
2 code implementations • CVPR 2019 • Sungsoo Ahn, Shell Xu Hu, Andreas Damianou, Neil D. Lawrence, Zhenwen Dai
We further demonstrate the strength of our method on knowledge transfer across heterogeneous network architectures by transferring knowledge from a convolutional neural network (CNN) to a multi-layer perceptron (MLP) on CIFAR-10.