LAMP

Category: Publications

  • Attribution-aware Weight Transfer: A Warm-Start Initialization for Class-Incremental Semantic Segmentation

    Dipam Goswami, René Schuster, Joost van de Weijer, Didier Stricker Read Full Paper → In class-incremental semantic segmentation (CISS), deep learning architectures suffer from the critical problems of catastrophic forgetting and semantic background shift. Although recent works focused on these issues, existing classifier initialization methods do not address the background shift problem and assign the same initialization weights […]

  • Positive Pair Distillation Considered Harmful: Continual Meta Metric Learning for Lifelong Object Re-Identification

    Kai Wang, Chenshen Wu, Andy Bagdanov, Xialei Liu, Shiqi Yang, Shangling Jui, Joost van de Weijer Read Full Paper → Lifelong object re-identification incrementally learns from a stream of re-identification tasks. The objective is to learn a representation that can be applied to all tasks and that generalizes to previously unseen re-identification tasks. The main challenge is that at inference time […]

  • Attention Distillation: self-supervised vision transformer students need more guidance

    Kai Wang, Fei Yang, Joost van de Weijer Read Full Paper → Self-supervised learning has been widely applied to train high-quality vision transformers. Unleashing their excellent performance on memory and compute constraint devices is therefore an important research topic. However, how to distill knowledge from one self-supervised ViT to another has not yet been explored. Moreover, the […]

  • Class-incremental learning: survey and performance evaluation on image classification

    Marc Masana, Xialei Liu, Bartlomiej Twardowski, Mikel Menta, Andrew D. Bagdanov, Joost van de Weijer Read Full Paper → For future learning systems, incremental learning is desirable because it allows for: efficient resource usage by eliminating the need to retrain from scratch at the arrival of new data; reduced memory usage by preventing or limiting the amount of data required […]

  • One paper accepted at International Conference on Learning Representations (ICLR) 2022

  • Distilling GANs with Style-Mixed Triplets for X2I Translation with Limited Data

    Yaxing Wang, Joost van de weijer, Lu Yu, SHANGLING JUI Read Full Paper → Conditional image synthesis is an integral part of many X2I translation systems, including image-to-image, text-to-image and audio-to-image translation systems. Training these large systems generally requires huge amounts of training data. Therefore, we investigate knowledge distillation to transfer knowledge from a high-quality unconditioned generative model […]

  • Towards Exemplar-Free Continual Learning in Vision Transformers: an Account of Attention, Functional and Weight Regularization

    Francesco Pelosin, Saurav Jha, Andrea Torsello, Bogdan Raducanu, Joost van de Weijer Read Full Paper → In this paper, we investigate the continual learning of Vision Transformers (ViT) for the challenging exemplar-free scenario, with special focus on how to efficiently distill the knowledge of its crucial self-attention mechanism (SAM). Our work takes an initial step towards a surgical investigation […]

  • Continually Learning Self-Supervised Representations with Projected Functional Regularization

    Alex Gomez-Villa, Bartlomiej Twardowski, Lu Yu, Andrew D. Bagdanov, Joost van de Weijer Read Full Paper → Recent self-supervised learning methods are able to learn high-quality image representations and are closing the gap with supervised approaches. However, these methods are unable to acquire new knowledge incrementally — they are, in fact, mostly used only as a pre-training phase over […]

  • Area Under the ROC Curve Maximization for Metric Learning

    Bojana Gajić, Ariel Amato, Ramon Baldrich, Joost van de Weijer, Carlo Gatta Read Full Paper → Most popular metric learning losses have no direct relation with the evaluation metrics that are subsequently applied to evaluate their performance. We hypothesize that training a metric learning model by maximizing the area under the ROC curve (which is […]

  • Incremental Meta-Learning via Episodic Replay Distillation for Few-Shot Image Recognition

    Kai Wang, Xialei Liu, Andy Bagdanov, Luis Herranz, Shangling Jui, Joost van de Weijer Read Full Paper → Most meta-learning approaches assume the existence of a very large set of labeled data available for episodic meta-learning of base knowledge. This contrasts with the more realistic continual learning paradigm in which data arrives incrementally in the form of tasks containing disjoint […]