Spatial-then-Temporal Self-Supervised Learning for Video Correspondence

Learning temporal correspondence from unlabeled videos is of vital importance in computer vision, and has been tackled by different kinds of self-supervised pretext tasks. For the self-supervised learning, recent studies suggest using large-scale video datasets despite the training cost. We propose a spatial-then-temporal pretext task to address the training data cost problem. The task consists of two steps. First, we use contrastive learning from unlabeled still image data to obtain appearance-sensitive features. Then we switch to unlabeled video data and learn motion-sensitive features by reconstructing frames. In the second step, we propose a global correlation distillation loss to retain the appearance sensitivity learned in the first step, as well as a local correlation distillation loss in a pyramid structure to combat temporal discontinuity. Experimental results demonstrate that our method surpasses the state-of-the-art self-supervised methods on a series of correspondence-based tasks. The conducted ablation studies verify the effectiveness of the proposed two-step task and loss functions.

Making Byzantine Decentralized Learning Efficient

Decentralized-SGD (D-SGD) distributes heavy learning tasks across multiple machines (a.k.a., {\em nodes}), effectively dividing the workload per node by the size of the system. However, a handful of \emph{Byzantine} (i.e., misbehaving) nodes can jeopardize the entire learning procedure. This vulnerability is further amplified when the system is \emph{asynchronous}. Although approaches that confer Byzantine resilience to D-SGD have been proposed, these significantly impact the efficiency of the process to the point of even negating the benefit of decentralization. This naturally raises the question: \emph{can decentralized learning simultaneously enjoy Byzantine resilience and reduced workload per node?} We answer positively by proposing \newalgorithm{} that ensures Byzantine resilience without losing the computational efficiency of D-SGD. Essentially, \newalgorithm{} weakens the impact of Byzantine nodes by reducing the variance in local updates using \emph{Polyak's momentum}. Then, by establishing coordination between nodes via {\em signed echo broadcast} and a {\em nearest-neighbor averaging} scheme, we effectively tolerate Byzantine nodes whilst distributing the overhead amongst the non-Byzantine nodes. To demonstrate the correctness of our algorithm, we introduce and analyze a novel {\em Lyapunov function} that accounts for the {\em non-Markovian model drift} arising from the use of momentum. We also demonstrate the efficiency of \newalgorithm{} through experiments on several image classification tasks.

Using Deep Image Prior to Assist Variational Selective Segmentation Deep Learning Algorithms

Variational segmentation algorithms require a prior imposed in the form of a regularisation term to enforce smoothness of the solution. Recently, it was shown in the Deep Image Prior work that the explicit regularisation in a model can be removed and replaced by the implicit regularisation captured by the architecture of a neural network. The Deep Image Prior approach is competitive, but is only tailored to one specific image and does not allow us to predict future images. We propose to incorporate the ideas from Deep Image Prior into a more traditional learning algorithm to allow us to use the implicit regularisation offered by the Deep Image Prior, but still be able to predict future images.

CHORE: Contact, Human and Object REconstruction from a single RGB image

While most works in computer vision and learning have focused on perceiving 3D humans from single images in isolation, in this work we focus on capturing 3D humans interacting with objects. The problem is extremely challenging due to heavy occlusions between human and object, diverse interaction types and depth ambiguity. In this paper, we introduce CHORE, a novel method that learns to jointly reconstruct human and object from a single image. CHORE takes inspiration from recent advances in implicit surface learning and classical model-based fitting. We compute a neural reconstruction of human and object represented implicitly with two unsigned distance fields, and additionally predict a correspondence field to a parametric body as well as an object pose field. This allows us to robustly fit a parametric body model and a 3D object template, while reasoning about interactions. Furthermore, prior pixel-aligned implicit learning methods use synthetic data and make assumptions that are not met in real data. We propose a simple yet effective depth-aware scaling that allows more efficient shape learning on real data. Our experiments show that our joint reconstruction learned with the proposed strategy significantly outperforms the SOTA. Our code and models will be released to foster future research in this direction.

A Novel Explainable Out-of-Distribution Detection Approach for Spiking Neural Networks

Research around Spiking Neural Networks has ignited during the last years due to their advantages when compared to traditional neural networks, including their efficient processing and inherent ability to model complex temporal dynamics. Despite these differences, Spiking Neural Networks face similar issues than other neural computation counterparts when deployed in real-world settings. This work addresses one of the practical circumstances that can hinder the trustworthiness of this family of models: the possibility of querying a trained model with samples far from the distribution of its training data (also referred to as Out-of-Distribution or OoD data). Specifically, this work presents a novel OoD detector that can identify whether test examples input to a Spiking Neural Network belong to the distribution of the data over which it was trained. For this purpose, we characterize the internal activations of the hidden layers of the network in the form of spike count patterns, which lay a basis for determining when the activations induced by a test instance is atypical. Furthermore, a local explanation method is devised to produce attribution maps revealing which parts of the input instance push most towards the detection of an example as an OoD sample. Experimental results are performed over several image classification datasets to compare the proposed detector to other OoD detection schemes from the literature. As the obtained results clearly show, the proposed detector performs competitively against such alternative schemes, and produces relevance attribution maps that conform to expectations for synthetically created OoD instances.

Perceiver-Actor: A Multi-Task Transformer for Robotic Manipulation

Transformers have revolutionized vision and natural language processing with their ability to scale with large datasets. But in robotic manipulation, data is both limited and expensive. Can we still benefit from Transformers with the right problem formulation? We investigate this question with PerAct, a language-conditioned behavior-cloning agent for multi-task 6-DoF manipulation. PerAct encodes language goals and RGB-D voxel observations with a Perceiver Transformer, and outputs discretized actions by "detecting the next best voxel action". Unlike frameworks that operate on 2D images, the voxelized observation and action space provides a strong structural prior for efficiently learning 6-DoF policies. With this formulation, we train a single multi-task Transformer for 18 RLBench tasks (with 249 variations) and 7 real-world tasks (with 18 variations) from just a few demonstrations per task. Our results show that PerAct significantly outperforms unstructured image-to-action agents and 3D ConvNet baselines for a wide range of tabletop tasks.

Soft Diffusion: Score Matching for General Corruptions

We define a broader family of corruption processes that generalizes previously known diffusion models. To reverse these general diffusions, we propose a new objective called Soft Score Matching that provably learns the score function for any linear corruption process and yields state of the art results for CelebA. Soft Score Matching incorporates the degradation process in the network and trains the model to predict a clean image that after corruption matches the diffused observation. We show that our objective learns the gradient of the likelihood under suitable regularity conditions for the family of corruption processes. We further develop a principled way to select the corruption levels for general diffusion processes and a novel sampling method that we call Momentum Sampler. We evaluate our framework with the corruption being Gaussian Blur and low magnitude additive noise. Our method achieves state-of-the-art FID score $1.85$ on CelebA-64, outperforming all previous linear diffusion models. We also show significant computational benefits compared to vanilla denoising diffusion.

MR Image Denoising and Super-Resolution Using Regularized Reverse Diffusion

Patient scans from MRI often suffer from noise, which hampers the diagnostic capability of such images. As a method to mitigate such artifact, denoising is largely studied both within the medical imaging community and beyond the community as a general subject. However, recent deep neural network-based approaches mostly rely on the minimum mean squared error (MMSE) estimates, which tend to produce a blurred output. Moreover, such models suffer when deployed in real-world sitautions: out-of-distribution data, and complex noise distributions that deviate from the usual parametric noise models. In this work, we propose a new denoising method based on score-based reverse diffusion sampling, which overcomes all the aforementioned drawbacks. Our network, trained only with coronal knee scans, excels even on out-of-distribution in vivo liver MRI data, contaminated with complex mixture of noise. Even more, we propose a method to enhance the resolution of the denoised image with the same network. With extensive experiments, we show that our method establishes state-of-the-art performance, while having desirable properties which prior MMSE denoisers did not have: flexibly choosing the extent of denoising, and quantifying uncertainty.

FedDropoutAvg: Generalizable federated learning for histopathology image classification

Federated learning (FL) enables collaborative learning of a deep learning model without sharing the data of participating sites. FL in medical image analysis tasks is relatively new and open for enhancements. In this study, we propose FedDropoutAvg, a new federated learning approach for training a generalizable model. The proposed method takes advantage of randomness, both in client selection and also in federated averaging process. We compare FedDropoutAvg to several algorithms in an FL scenario for real-world multi-site histopathology image classification task. We show that with FedDropoutAvg, the final model can achieve performance better than other FL approaches and closer to a classical deep learning model that requires all data to be shared for centralized training. We test the trained models on a large dataset consisting of 1.2 million image tiles from 21 different centers. To evaluate the generalization ability of the proposed approach, we use held-out test sets from centers whose data was used in the FL and for unseen data from other independent centers whose data was not used in the federated training. We show that the proposed approach is more generalizable than other state-of-the-art federated training approaches. To the best of our knowledge, ours is the first study to use a randomized client and local model parameter selection procedure in a federated setting for a medical image analysis task.

Deep Unfolding of the DBFB Algorithm with Application to ROI CT Imaging with Limited Angular Density

This paper addresses the problem of image reconstruction for region-of-interest (ROI) computed tomography (CT). While model-based iterative methods can be used for such a problem, their practicability is often limited due to tedious parameterization and slow convergence. In addition, inadequate solutions can be obtained when the retained priors do not perfectly fit the solution space. Deep learning methods offer an alternative approach that is fast, leverages information from large data sets, and thus can reach high reconstruction quality. However, these methods usually rely on black boxes not accounting for the physics of the imaging system, and their lack of interpretability is often deplored. At the crossroads of both methods, unfolded deep learning techniques have been recently proposed. They incorporate the physics of the model and iterative optimization algorithms into a neural network design, leading to superior performance in various applications. This paper introduces a novel, unfolded deep learning approach called U-RDBFB designed for ROI CT reconstruction from limited data. Few-view truncated data are efficiently handled thanks to a robust non-convex data fidelity function combined with sparsity-inducing regularization functions. Iterations of a block dual forward-backward (DBFB) algorithm, embedded in an iterative reweighted scheme, are then unrolled over a neural network architecture, allowing the learning of various parameters in a supervised manner. Our experiments show an improvement over various state-of-the-art methods, including model-based iterative schemes, deep learning architectures, and deep unfolding methods.