Over the last decade there has been an extensive evolution in the Artificial Intelligence (AI) field. Modern radiation oncology is based on the exploitation of advanced computational methods aiming to personalization and high diagnostic and therapeutic precision. The quantity of the available imaging data and the increased developments of Machine Learning (ML), particularly Deep Learning (DL), triggered the research on uncovering "hidden" biomarkers and quantitative features from anatomical and functional medical images. Deep Neural Networks (DNN) have achieved outstanding performance and broad implementation in image processing tasks. Lately, DNNs have been considered for radiomics and their potentials for explainable AI (XAI) may help classification and prediction in clinical practice. However, most of them are using limited datasets and lack generalized applicability. In this study we review the basics of radiomics feature extraction, DNNs in image analysis, and major interpretability methods that help enable explainable AI. Furthermore, we discuss the crucial requirement of multicenter recruitment of large datasets, increasing the biomarkers variability, so as to establish the potential clinical value of radiomics and the development of robust explainable AI models.
We present the quantum κ-deformation of BMS symmetry, by generalizing the lightlike κ-Poincaré Hopf algebra. On the technical level our analysis relies on the fact that the lightlike κ-deformation of Poincaré algebra is given by a twist and the lightlike deformation of any algebra containing Poincaré as a subalgebra can be done with the help of the same twisting element. We briefly comment on the physical relevance of the obtained κ-BMS Hopf algebra as a possible asymptotic symmetry of quantum gravity.In the recent years there is a surge of interest in BMS symmetry at null infinity of asymptotically flat spacetimes [1], [2], [3]. This renewed interest in the seemingly exotic aspects of classical general relativity was fueled by the discovery of a surprising and close relationship between asymptotic symmetries and soft gravitons theorem [4] that, as it turned out, has its roots in Ward identities for supertranslations [5], [6]. Moreover, the gravitational memory effect [7] appears to be related to the two [8], so that there is a triangle of interrelationships, which vertices are BMS symmetry, Weinberg's soft graviton theorem, and the memory effect. The extensive discussion of these effects can be found in recent reviews [9] and [10].It has also been argued recently [11] that the similarity of null infinity and black hole horizon suggests that the charges associated with the BMS symmetry might be present at the black hole horizon, so that the black hole might have an infinite number of hairs, so-called soft hairs, in addition to the three classical one: mass, charge, and angular momentum. It was argued in [11] and [12] that the presence of these charges may help solving the black hole information paradox.In this letter we will investigate the properties of a κ-deformed generalization of the BMS symmetry. There are several reasons to be interested in such a generalization. The main one is that it is believed that investigations of κ-deformation might shed some light on the properties of elusive quantum gravity. The deformation parameter of κ-deformation of Poincaré algebra [13]-[17], see [18] for a recent review and more references, has dimension of mass, and therefore it is natural to identify it with Planck mass, or inverse Planck length, which in turn suggests a possible close relationship between this deformation and quantum gravity. In fact, it was shown in [19]
Evaluating, explaining, and visualizing high-level concepts in generative models, such as variational autoencoders (VAEs), is challenging in part due to a lack of known prediction classes that are required to generate saliency maps in supervised learning. While saliency maps may help identify relevant features (e.g., pixels) in the input for classification tasks of deep neural networks, similar frameworks are understudied in unsupervised learning. Therefore, we introduce a new method of obtaining saliency maps for latent representations of known or novel high-level concepts, often called concept vectors in generative models. Concept scores, analogous to class scores in classification tasks, are defined as dot products between concept vectors and encoded input data, which can be readily used to compute the gradients. The resulting concept saliency maps are shown to highlight input features deemed important for high-level concepts. Our method is applied to the VAE's latent space of CelebA dataset in which known attributes such as "smiles" and "hats" are used to elucidate relevant facial features. Furthermore, our application to spatial transcriptomic (ST) data of a mouse olfactory bulb demonstrates the potential of latent representations of morphological layers and molecular features in advancing our understanding of complex biological systems. By extending the popular method of saliency maps to generative models, the proposed concept saliency maps help improve interpretability of latent variable models in deep learning.Codes to reproduce and to implement concept saliency maps: https://github.com/lenbrocki/concept-saliency-maps
BMS symmetry is a symmetry of asymptotically flat spacetimes in vicinity of the null boundary of spacetime and it is expected to play a fundamental role in physics. It is interesting therefore to investigate the structures and properties of quantum deformations of these symmetries, which are expected to shed some light on symmetries of quantum spacetime. In this paper we discuss the structure of the algebra of extended BMS symmetries in 3 and 4 spacetime dimensions, realizing that these algebras contain an infinite number of distinct Poincaré subalgebras, a fact that has previously been noted in the 3 dimensional case only. Then we use these subalgebras to construct an infinite number of different Hopf algebras being quantum deformations of the BMS algebras. We also discuss different types of twist-deformations and the dual Hopf algebras, which could be interpreted as noncommutative, extended quantum spacetimes.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
hi@scite.ai
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.