Publications

Published:

Journal papers Workshop papers

Peer-reviewed conference papers:

  • Revisiting character-level adversarial attacks for Language Models.

    Elias Abad Rocamora, Yongtao Wu, Fanghui Liu, Grigorios Chrysos, Volkan Cevher
    International Conference on Machine Learning (ICML), 2024.
     PDF  Code
    We introduce a character-level adversarial attack for text classifiers.

  • Going beyond compositional generalization, DDPMs can produce zero-shot interpolation.

    Justin Deschenaux, Igor Krawczuk, Grigorios Chrysos, Volkan Cevher
    International Conference on Machine Learning (ICML), 2024.
     PDF  Code
    We study the ability of diffusion models to generate samples beyond their training distribution focusing on the interpolation abilities.

  • Multilinear Operator Networks.

    Yixin Cheng, Grigorios Chrysos, Markos Georgopoulos, Volkan Cevher
    International Conference on Learning Representations (ICLR), 2024.
     PDF  Code
    We introduce a family of networks that rely on multilinear operations and capture high-degree interactions of the input elements. This family of networks, called MONet, performs on par with modern architectures on image recognition and beyond.

  • Generalization of Scaled Deep ResNets in the Mean-Field Regime.

    Yihang Chen, Fanghui Liu, Yiping Lu, Grigorios Chrysos, Volkan Cevher
    International Conference on Learning Representations (ICLR), 2024.
     PDF  Code
    We investigate the generalization properties of deep and wide ResNet models in the mean-field regime.

  • Robust NAS under adversarial training: benchmark, theory, and beyond.

    Yongtao Wu, Fanghui Liu, Carl-Johann Simon-Gabriel, Grigorios Chrysos, Volkan Cevher
    International Conference on Learning Representations (ICLR), 2024.
     PDF  Code We release a benchmark for searching adversarially robust networks, while we also establish the generalization bounds for searched architectures under multi-objective adversarial training.

  • Efficient local linearity regularization to overcome catastrophic overfitting.

    Elias Abad Rocamora, Fanghui Liu, Grigorios Chrysos, Pablo Olmos, Volkan Cevher
    International Conference on Learning Representations (ICLR), 2024.
     PDF  Code
    We propose a regularization term to mitigate catastrophic overfitting, which emerges in Single-step adversarial training.

  • Maximum Independent Set: Self-Training through Dynamic Programming.

    Lorenzo Brusca*, Lars C.P.M. Quaedvlieg*, Stratis Skoulakis*, Grigorios Chrysos, Volkan Cevher
    Conference on Neural Information Processing Systems (NeurIPS), 2023.
     PDF  Code  Project page
    We design a framework for estimating the maximum independent set (MIS) without using supervised samples, inspired by dynamic programming.

  • On the Convergence of Encoder-Only Shallow Transformers.

    Yongtao Wu, Fanghui Liu, Grigorios Chrysos, Volkan Cevher
    Conference on Neural Information Processing Systems (NeurIPS), 2023.
     PDF
    We study the convergence of shallow (single-block) transformers under different scalings and initializations using the realistic softmax-based transformer.

  • Benign Overfitting in Deep Neural Networks under Lazy Training.

    Zhenyu Zhu, Fanghui Liu, Grigorios Chrysos, Francesco Locatello, Volkan Cevher
    International Conference on Machine Learning (ICML), 2023.
     PDF
    In this work, we study the three interrelated concepts of overparameterization, benign overfitting, and the Lipschitz constant of DNNs and connect them in the lazy training regime.

  • Regularization of polynomial networks for image recognition.

    Grigorios Chrysos, Bohan Wang, Jiankang Deng, Volkan Cevher
    Computer Vision and Pattern Recognition Conference (CVPR), 2023.
     PDF
    We demonstrate how polynomial networks (without elementwise activation functions) can benefit from regularization schemes to reach the performance of standard neural networks. Then, we introduce a new class of polynomial networks that achieve even higher degree of expansions by using these additional regularization schemes.

  • Robustness in deep learning: The good (width), the bad (depth), and the ugly (initialization).

    Zhenyu Zhu, Fanghui Liu, Grigorios Chrysos, Volkan Cevher
    Conference on Neural Information Processing Systems (NeurIPS), 2022.
     PDF
    We explore the interplay of the width, the depth and the initialization(s) on the average robustness of neural networks with both theoretical bounds and empirical validation.

  • Generalization Properties of NAS under Activation and Skip Connection Search.

    Zhenyu Zhu, Fanghui Liu, Grigorios Chrysos, Volkan Cevher
    Conference on Neural Information Processing Systems (NeurIPS), 2022.
     PDF
    Using our theoretical guarantees of neural architecture search (NAS) under various activation functions and residual connections, we design an effective train-free algorithm for NAS.

  • Sound and Complete Verification of Polynomial Networks.

    Elias Abad Rocamora, Mehmet Fatih Sahin, Fanghui Liu, Grigorios Chrysos, Volkan Cevher
    Conference on Neural Information Processing Systems (NeurIPS), 2022.
     PDF  Code
    We propose a branch and bound algorithm for certifying polynomial networks against (adversarial) attacks.

  • Extrapolation and Spectral Bias of Neural Nets with Hadamard Product: a Polynomial Net Study.

    Yongtao Wu, Zhenyu Zhu, Fanghui Liu, Grigorios Chrysos, Volkan Cevher
    Conference on Neural Information Processing Systems (NeurIPS), 2022.
     PDF  Code
    We study the extrapolation and spectral bias of neural networks with Hadamard products from a neural tangent kernel perspective.

  • Augmenting Deep Classifiers with Polynomial Neural Networks.

    Grigorios Chrysos* Markos Georgopoulos*, Jiankang Deng, Jean Kossaifi, Yannis Panagakis, Anima Anandkumar
    European Conference on Computer Vision (ECCV), 2022.
     PDF  Code
    We express modern architectures (e.g., residual and non-local networks) in the form of different degree polynomials of the input. This enables us to design extensions of successful architectures that perform favorably in various benchmarks.

  • Cluster-guided Image Synthesis with Unconditional Models.

    Markos Georgopoulos, James Oldfield, Grigorios Chrysos, Yannis Panagakis
    Computer Vision and Pattern Recognition Conference (CVPR), 2022.
     PDF
    We study controllable generation in unsupervised GAN models by leveraging clusters in the representation space of the generator. We show that these clusters, which capture semantic attributes, can be used for conditioning the generator.

  • The Spectral Bias of Polynomial Neural Networks.

    Moulik Choraria, Leello Tadesse Dadi, Grigorios Chrysos, Julien Mairal, Volkan Cevher
    International Conference on Learning Representations (ICLR), 2022.
     PDF
    We study the spectral bias of polynomial networks and compare it with the spectral bias of standard neural nets using kernel approximations.

  • Controlling the Complexity and Lipschitz Constant improves Polynomial Nets.

    Zhenyu Zhu, Fabian Latorre, Grigorios Chrysos, Volkan Cevher
    International Conference on Learning Representations (ICLR), 2022.
     PDF
    We provide sample complexity results and bounds on the Lipschitz constant of polynomial networks, which we use to construct a regularization scheme that improves the robustness against adversarial noise.

  • Conditional Generation Using Polynomial Expansions.

    Grigorios Chrysos, Markos Georgopoulos, Yannis Panagakis
    Conference on Neural Information Processing Systems (NeurIPS), 2021.
     PDF  Code
    We propose a polynomial expansion with respect to two (or more) variables, which is applied to conditional image generation.

  • Poly-NL: Linear Complexity Non-local Layers with Polynomials.

    Francesca Babiloni, Ioannis Marras, Filippos Kokkinos, Jiankang Deng, Grigorios Chrysos, Stefanos Zafeiriou
    International Conference on Computer Vision (ICCV), 2021.
     PDF
    We cast non-local blocks as special cases of third degree polynomial functions. In addition, we propose a new non-local block that builds on this polynomial perspective but has more efficient operations, i.e., we aim to retain the expressivity of non-local layers while maintaining a linear complexity.

  • Unsupervised Controllable Generation with Self-Training.

    Grigorios Chrysos, Jean Kossaifi, Zhiding Yu, Anima Anandkumar
    International Joint Conference on Neural Networks (IJCNN), 2021.
     PDFOral.
    We modify the GAN architecture to achieve interpretable generation without using any supervision.

  • Reconstructing the Noise Manifold for Image Denoising.

    Ioannis Marras, Grigorios Chrysos, Ioannis Alexiou, Gregory Slabaugh, Stefanos Zafeiriou
    European Conference on Computer Vision (ECCV), 2020.
     PDF
    We propose learning the noise variance manifold along with typical image-to-image translation to obtain improved denoising.

  • Multilinear Latent Conditioning for Generating Unseen Attribute Combinations.

    Markos Georgopoulos, Grigorios Chrysos, Maja Pantic, Yannis Panagakis
    International Conference on Machine Learning (ICML), 2020.
     PDF
    We extend conditional VAE to capture multiplicative interactions of the (annotated) attributes in the latent space. This enables generating images with unseen attribute combinations during training.

  • Π-nets: Deep Polynomial Neural Networks.

    Grigorios Chrysos, Stylianos Moschoglou, Giorgos Bouritsas, Yannis Panagakis, Jiankang Deng, Stefanos Zafeiriou
    Computer Vision and Pattern Recognition Conference (CVPR), 2020.
     PDF  Code  Blog post  1-minute video  Poster
    We use a high-order polynomial expansion as a function approximation method. The unknown parameters of the polynomial (i.e., high-order tensors) are estimated using a collective tensor factorization.

  • Robust Conditional Generative Adversarial Networks.

    Grigorios Chrysos, Jean Kossaifi, Stefanos Zafeiriou
    International Conference on Learning Representations (ICLR), 2019.
     PDF  Code  Poster

    The topic of conditional data generation task (e.g., super-resolution) is the focus of this work. We introduce a new pathway in the encoder-decoder generator to improve the synthesized image.

  • Surface Based Object Detection in RGBD Images.

    Siddhartha Chandra, Grigorios Chrysos, Iasonas Kokkinos
    British Machine Vision Conference (BMVC), 2015.
     PDF Oral, acceptance rate: 7%.
    We extend standard object detection pipelines by leveraging depth information and introducing viewpoint based mixture components.

Return to the top

Peer-reviewed journal papers:

  • Federated Learning under Covariate Shifts with Generalization Guarantees.

    Ali Ramezani-Kebrya, Fanghui Liu, Thomas Pethick, Grigorios Chrysos, Volkan Cevher
    Transactions on Machine Learning Research (TMLR), 2023.
     PDF  Code
    We focus on the aspect of federated learning when there are coavariate shifts, which is a realistic scenario in multiple cases. We derive both theoretical guarantees and demonstrate how this can work in imbalanced data settings.

  • Linear Complexity Self-Attention with 3rd Order Polynomials.

    Francesca Babiloni, Ioannis Marras, Filippos Kokkinos, Jiankang Deng, Matteo Maggioni, Grigorios Chrysos, Philip Torr, Stefanos Zafeiriou
    IEEE Transactions on Pattern Analysis and Machine Intelligence (T-PAMI), 2023. (impact factor 2019: 17.861).
     Paper
    We cast self-attention (and non-local blocks) as special cases of third degree polynomial functions. In addition, we propose a new block that builds on this polynomial perspective but it is more computationally efficient, i.e., we aim to retain the expressivity of self-attention/non-local layers while maintaining a linear complexity.

  • Revisiting adversarial training for the worst-performing class.

    Thomas Pethick, Grigorios Chrysos, Volkan Cevher
    Transactions on Machine Learning Research (TMLR), 2023.
     PDF  Code
    We propose a new training method called class focused online learning (CFOL) to reduce the gap between the top-performing and worst-performing classes in adversarial training, resulting in a min-max-max optimization formulation.

  • Deep Polynomial Neural Networks.

    Grigorios Chrysos, Stylianos Moschoglou, Giorgos Bouritsas, Jiankang Deng, Yannis Panagakis, Stefanos Zafeiriou
    IEEE Transactions on Pattern Analysis and Machine Intelligence (T-PAMI), 2021. (impact factor 2019: 17.861)
     Paper Paper (open access)  Code  Blog post  1-minute video
    We propose a new class of architectures that use polynomial expansions to approximate the target functions. We validate the proposed polynomial expansions (i.e. Π-nets) in diverse experiments: data generation, data classifcation, face recognition and non-euclidean representation learning.

  • Tensor Methods in Computer Vision and Deep Learning.

    Yannis Panagakis*, Jean Kossaifi*, Grigorios Chrysos, James Oldfield, Mihalis A. Nicolaou, Anima Anandkumar, Stefanos Zafeiriou
    Proceedings of the IEEE, 2021.
     Paper  Code
    We provide an in-depth review of tensors and tensor methods in the context of representation learning and deep learning, with a particular focus on computer vision applications. We also provide jupyter notebooks with accompanying code.

  • Non-adversarial polynomial synthesis.

    Grigorios Chrysos, Yannis Panagakis
    Pattern Recognition Letters, 2020.
     Paper
    We propose a decoder-only generator that uses a polynomial expansion to synthesize new images.

  • RoCGAN: Robust Conditional GAN.

    Grigorios Chrysos, Jean Kossaifi, Stefanos Zafeiriou
    International Journal of Computer Vision (IJCV), 2020. (impact factor 2019: 11.042)
     Paper (open access)  Code
    We leverage structure in the output domain of a conditional data generation task (e.g., super-resolution) to improve the synthesized image. We experimentally validate that this results in synthesized images more robust to noise. Extension of the conference paper.

  • Motion Deblurring of Faces.

    Grigorios Chrysos, Paolo Favaro, Stefanos Zafeiriou
    International Journal of Computer Vision (IJCV), 2019. (impact factor 2019: 11.042)
     Paper (open access)
    We introduce a framework for tackling motion blur of faces. Our method simulates motion blur using averaging of video frames, while we collect a dataset that contains millions of such frames.

  • The Menpo Benchmark for Multi-pose 2D and 3D Facial Landmark Localisation and Tracking.

    Jiankang Deng, Anastasios Roussos, Grigorios Chrysos, Evangelos Ververas, Irene Kotsia, Jie Shen, Stefanos Zafeiriou
    International Journal of Computer Vision (IJCV), 2019. (impact factor 2019: 11.042)
     Paper (open access)
    A semi-automatic framework is proposed for annotating challenging deformable images and videos.

  • A Comprehensive Performance Evaluation of Deformable Face Tracking ''In-the-Wild''.

    Grigorios Chrysos, Epameinondas Antonakos, Patrick Snape, A. Asthana, Stefanos Zafeiriou
    International Journal of Computer Vision (IJCV), 2018. (impact factor 2019: 11.042)
     Paper (open access)  Code
    We conduct a large-scale study of deformable face tracking `in-the-wild', i.e., with videos captured in unrestricted conditions.

  • IPST: Incremental Pictorial Structures for model-free Tracking of deformable objects.

    Grigorios Chrysos, Epameinondas Antonakos, Stefanos Zafeiriou
    IEEE Transactions on Image Processing (TIP), 2018. (impact factor 2019: 9.34)
     Paper
    We introduce incremental pictorial structures for tracking deformable (part-based) objects, e.g., human body parts or fiducial points in the face.

  • PD2T: Person-specific Detection, Deformable Tracking.

    Grigorios Chrysos, Stefanos Zafeiriou
    IEEE Transactions on Pattern Analysis and Machine Intelligence (T-PAMI), 2017. (impact factor 2019: 17.861)
     Paper
    We propose a framework for extracting object-specific statistics for tracking a (deformable) object.

Return to the top

Workshop papers:

  • Self-Supervised Neural Architecture Search for Imbalanced Datasets.

    Aleksandr Timofeev, Grigorios Chrysos, Volkan Cevher
    International Conference on Machine Learning Workshops (ICMLW), 2021.
     PDF
    We propose a neural architecture search (NAS) framework for real world tasks: (a) in the absence of labels, (b) in the presence of imbalanced datasets, (c) on a constrained computational budget.

  • Unsupervised Controllable Generation with Self-Training.

    Grigorios Chrysos, Jean Kossaifi, Zhiding Yu, Anima Anandkumar
    International Conference on Machine Learning Workshops (ICMLW), 2020.
     PDF
    We modify the GAN architecture to achieve interpretable generation without using any supervision.

  • The 3D Menpo Facial Landmark Tracking Challenge.

    Stefanos Zafeiriou*, Grigorios Chrysos*, Anastasios Roussos*, Evangelos Ververas, J. Deng, George Trigeorgis
    International Conference on Computer Vision Workshops (ICCVW), 2017.
     PDF
    The first large-scale dataset with 3D annotations of facial landmarkrs is introduced.

  • Deep Face Deblurring.

    Grigorios Chrysos, Stefanos Zafeiriou
    Computer Vision and Pattern Recognition Conference Workshops (CVPRW), 2017.
     PDF
    A method for face deblurring is proposed. The method utilizes weak supervision to guide the learning of the deep neural network.

  • The Menpo Facial Landmark Localisation Challenge.

    Stefanos Zafeiriou, George Trigeorgis, Grigorios Chrysos, J. Deng, Jie Shen
    Computer Vision and Pattern Recognition Conference Workshops (CVPRW), 2017.
     PDF
    The first large-scale dataset with annotations of facial landmarkrs in both (semi-)frontal and profile poses is introduced.

  • The First Facial Landmark Tracking in-the-Wild Challenge: Benchmark and Results.

    Jie Shen, Stefanos Zafeiriou, Grigorios Chrysos, Jean Kossaifi, Georgios Tzimiropoulos, Maja Pantic
    International Conference on Computer Vision Workshops (ICCVW), 2015.
     PDF
    The first large-scale dataset for facial landmark tracking is introduced.

  • Offline Deformable Face Tracking in Arbitrary Videos.

    Grigorios Chrysos, Epameinondas Antonakos, Stefanos Zafeiriou, Patrick Snape
    International Conference on Computer Vision Workshops (ICCVW), 2015.
     PDF
    We propose a framework that can extract object-specific statistics and can be used for tracking long sequences of videos.

Thesis

  • Polynomial function approximation and its application to deep generative models

    Grigorios Chrysos
    PhD thesis, Imperial College London.
     PDF
    My PhD thesis, spanning my first work on polynomial networks and deep generative models.

Return to the top