A variety of deep learning approaches have been proposed to automatically classify Alzheimer’s disease (AD) from medical images. However, common approaches such as traditional convolutional neural networks (CNNs), lack interpretability and are prone to overfitting when trained on small datasets. As an alternative, significantly less work has explored applying deep learning approaches to region-based features that are commonly attained from atlas partitions of known regions of interest (ROI). In this work, we combine CNNs with graph neural networks (GNNs) to jointly learn an adjacency matrix of connectivity’s between ROIs as a prior for learning meaningful features for AD prediction. We apply our method to the ADNI dataset and systematically inspect the different intermediate layers of our network using t-SNE projections that show strong separation on out-of-sample data. Finally, we show that the edge probabilities alone are sufficient to reach high classification accuracy by training a secondary random forest classifier on the adjacency matrices outputted from our network and illustrate the interpretability properties of the graphs by visualizing the feature importance for all edges.
Numerous deep learning approaches have been proposed to automatically classify Alzheimer’s disease (AD) from medical images. However, common approaches, such as convolutional neural networks (CNNs), lack interpretability and are prone to over-fitting when trained on small datasets. As an alternative, significantly less work has explored applying deep learning approaches to region-based features commonly obtained from atlas partitions of known regions of interest (ROI). In this paper, we propose a self-attention mechanism to jointly learn a graph of ROI connectivity as a prior for learning meaningful features for AD prediction. We apply our method to both the classification of AD subjects from healthy controls and to predict whether mild cognitive impaired (MCI) subjects will progress to AD (pMCI) or not (sMCI). We systematically show that our model’s performance compares well with other common ML approaches for ROI-based methods, such as neural networks and support vector machines. Finally, we perform exploratory graph analysis to illustrate the interpretability properties of the attention graphs and how they can provide insight for scientific discovery.
Real world motion planning often suffers from the need to replan during execution of the trajectory. This replanning can be triggered as the robot fails to properly track the trajectory or new sensory information is provided that invalidates the planned trajectory. Particularly in the case of many occluded obstacles or in unstructured terrain, replanning is a frequent occurrence. Developing methods to allow the robots to replan efficiently allows for greater operation time and can ensure robot mission success. This paper presents a novel approach that updates heuristic weights of a sampling based A* planning algorithm. This approach utilizes parallel instances of this planner to quickly search through multiple heuristic weights within its allotted replanning time. These weights are employed upon triggered replanning to speed up computation time. The concept is tested on a simulated quadrupedal robot LLAMA with realistic constraints on computation time imposed.
In today’s increasingly divided political climate there is a need for a tool that can compare news articles and organizations so that a user can receive a wider range of views and philosophies. NewsAnalyticalToolkit allows a user to compare news sites and their political articles by coverage, mood, sentiment, and objectivity. The user can sort through the news by topic, which was determined using Natural Language Processing (NLP) and Latent Dirichlet Allocation (LDA). LDA is a probabilistic method used to discover latent topics within a series of documents and cluster them accordingly. Each news article can be considered a mix of multiple topics and LDA assigns a set of topics to each with a probability of it pertaining to that topic. For each topic, a user can then discover the coverage, mood, sentiment and objectivity expressed by each author and site. The mood was determined using IBM Watsons ToneAnalyzerV3, which uses linguistic analysis to detect emotional, social and language tones in written text. The analyzer is based on the theory of psycholinguistics, a field of research that explores the relationship between linguistic behavior and psychological theories. The sentiment and objectivity scores were determined using SentiWordNet, which is a lexical database that groups English words into sets of synonyms and assigns sentiment scores to them. The features were combined to plot an interactive graph of how opinionated versus how analytical an article is, so that the user can click through them to get a better understanding of the topic in question.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
INSTITUTIONAL Select your institution to access the SPIE Digital Library.
PERSONAL Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.