arXiv subjects

Tristan McKinney

Publications and source records attributed to Tristan McKinney.

4 recordsLinked to original sources

arXiv2022

Improving self-supervised representation learning via sequential adversarial masking

Recent methods in self-supervised learning have demonstrated that masking-based pretext tasks extend beyond NLP, serving as useful pretraining objectives in computer vision. However, existing approaches apply random or ad hoc masking strategies that limit the difficulty of the reconstruction task and, consequently, the strength of the learnt representations. We improve upon current state-of-the-art work in learning adversarial masks by proposing a new framework that generates masks in a sequential fashion with different constraints on the adversary. This leads to improvements in performance on various downstream tasks, such as classification on ImageNet100, STL10, and CIFAR10/100 and segmentation on Pascal VOC. Our results further demonstrate the promising capabilities of masking-based approaches for SSL in computer vision.

cs.CV↗

arXiv2021

Exploiting Invariance in Training Deep Neural Networks

Inspired by two basic mechanisms in animal visual systems, we introduce a feature transform technique that imposes invariance properties in the training of deep neural networks. The resulting algorithm requires less parameter tuning, trains well with an initial learning rate 1.0, and easily generalizes to different tasks. We enforce scale invariance with local statistics in the data to align similar samples at diverse scales. To accelerate convergence, we enforce a GL(n)-invariance property with global statistics extracted from a batch such that the gradient descent solution should remain invariant under basis change. Profiling analysis shows our proposed modifications takes 5% of the computations of the underlying convolution layer. Tested on convolutional networks and transformer networks, our proposed technique requires fewer iterations to train, surpasses all baselines by a large margin, seamlessly works on both small and large batch size training, and applies to different computer vision and language tasks.

cs.CV↗

arXiv2021

The Challenge of Variable Effort Crowdsourcing and How Visible Gold Can Help

We consider a class of variable effort human annotation tasks in which the number of labels required per item can greatly vary (e.g., finding all faces in an image, named entities in a text, bird calls in an audio recording, etc.). In such tasks, some items require far more effort than others to annotate. Furthermore, the per-item annotation effort is not known until after each item is annotated since determining the number of labels required is an implicit part of the annotation task itself. On an image bounding-box task with crowdsourced annotators, we show that annotator accuracy and recall consistently drop as effort increases. We hypothesize reasons for this drop and investigate a set of approaches to counteract it. Firstly, we benchmark on this task a set of general best-practice methods for quality crowdsourcing. Notably, only one of these methods actually improves quality: the use of visible gold questions that provide periodic feedback to workers on their accuracy as they work. Given these promising results, we then investigate and evaluate variants of the visible gold approach, yielding further improvement. Final results show a 7% improvement in bounding-box accuracy over the baseline. We discuss the generality of the visible gold approach and promising directions for future research.

cs.HC↗

arXiv2018

Effective Field Theory of 2D van Hove Singularities

We study the effective field theory of 2D fermions with a short-range interaction in the presence of a van Hove singularity. We find that there are additional divergences associated with the singularity that necessitate regularization beyond the usual Wilsonian cut-off. In the full theory these divergences are cut off by the finite size of the Brillouin zone. This leads to a UV/IR mixing and causes the RG equation for the coupling constant to have an explicit dependence on the ratio of the Wilsonian cut-off to the bandwidth. We discuss the properties of the superconducting ground state and the transport properties of the normal state and show that the latter are approximately described by the marginal Fermi liquid scenario. To leading order, our results are universal in the sense that they do not depend upon the nature of the non-van Hove portion of the Fermi surface. We also comment on the van Hove scenario of high-Tc superconductivity.

cond-mat.str-el↗