Bounding overwatch

About: Bounding overwatch is a research topic. Over the lifetime, 966 publications have been published within this topic receiving 15156 citations.

Range bounding with Taylor models for global optimization

Abstract: Taylor models provide enclosures of functional dependencies by a polynomial and an interval remainder bound that scales with a high power of the domain width, allowing a far-reaching suppression of the dependency problem. For the application to range bounding, one observes that the resulting polynomials are more well-behaved than the original function; in fact, merely naively evaluating them in interval arithmetic leads to a quadratic range bounder that is frequently noticeably superior to other second order methods. However, the particular polynomial form allows the use of other techniques. We review the linear dominated bounder (LDB) and the quadratic fast bounder (QFB). LDB often allows an exact bounding of the polynomial part if the function is monotonic. If it does not succeed to provide an optimal bound, it still often provides a reduction of the domain simultaneously in all variables. Near interior minimizers, where the quadratic part of the local Taylor model is positive semidefinite, QFB minimizes the quadratic contribution to the lower bound of the function, avoiding the infamous cluster effect for validated global optimization tasks. Some examples of the performance of the bounders for unconstrained global optimization problems are given, beginning with various common toy problems of the community, and also including a rather challenging Lennard-Jones problem.

Locating Objects Without Bounding Boxes

Abstract: Recent advances in convolutional neural networks (CNN) have achieved remarkable results in locating objects in images. In these networks, the training procedure usually requires providing bounding boxes or the maximum number of expected objects. In this paper, we address the task of estimating object locations without annotated bounding boxes which are typically hand-drawn and time consuming to label. We propose a loss function that can be used in any fully convolutional network (FCN) to estimate object locations. This loss function is a modification of the average Hausdorff distance between two unordered sets of points. The proposed method has no notion of bounding boxes, region proposals, or sliding windows. We evaluate our method with three datasets designed to locate people's heads, pupil centers and plant centers. We outperform state-of-the-art generic object detectors and methods fine-tuned for pupil tracking.

Universal Lesion Detection and Classification Using Limited Data and Weakly-Supervised Self-training

Abstract: Radiologists identify, measure, and classify clinically significant lesions routinely for cancer staging and tumor burden assessment. As these tasks are repetitive and cumbersome, only the largest lesion is identified leaving others of potential importance unmentioned. Automated deep learning-based methods for lesion detection have been proposed in literature to help relieve their tasks with the publicly available DeepLesion dataset (32,735 lesions, 32,120 CT slices, 10,594 studies, 4,427 patients, 8 body part labels). However, this dataset contains missing lesions, and displays a severe class imbalance in the labels. In our work, we use a subset of the DeepLesion dataset (boxes + tags) to train a state-of-the-art VFNet model to detect and classify suspicious lesions in CT volumes. Next, we predict on a larger data subset (containing only bounding boxes) and identify new lesion candidates for a weakly-supervised self-training scheme. The self-training is done across multiple rounds to improve the model’s robustness against noise. Two experiments were conducted with static and variable thresholds during self-training, and we show that sensitivity improves from 72.5% without self-training to 76.4% with self-training. We also provide a structured reporting guideline through a “Lesions” sub-section for entry into the “Findings” section of a radiology report. To our knowledge, we are the first to propose a weakly-supervised self-training approach for joint lesion detection and tagging in order to mine for under-represented lesion classes in the DeepLesion dataset.

On Adaptive Influence Maximization Under General Feedback Models

Abstract: The classic influence maximization problem explores the strategies for deploying cascades such that the total influence is maximized, and it assumes that the seed nodes that initiate the cascades are computed prior to the diffusion process. In its adaptive version, the seed nodes are allowed to be launched in an adaptive manner after observing certain diffusion results. In this article, we provide a systematic study on the adaptive influence maximization problem, focusing on the algorithmic analysis of the general feedback models. We introduce the concept of regret ratio to characterize the key trade-off in designing adaptive seeding strategies, based on which we present the approximation analysis for the well-known greedy policy. In addition, we provide analysis concerning improving the efficiencies and bounding the regret ratio. Finally, we propose several future research directions.