An Unsupervised Deep Learning Model to Discover Visual Similarity Between Sketches for Visual Analogy Support

2020 ◽  
Author(s):  
Zijian Zhang ◽  
Yan Jin
Keyword(s):  
Visual Similarity ◽  
Computational Method ◽  
Shape Features ◽  
Clustering Model ◽  
Extensive Evaluation ◽  
Latent Space ◽  
Unsupervised Deep Learning ◽  
Similar Appearance ◽  
The Impact ◽  
Visual Analogy

Abstract Visual analogy has been recognized as an important cognitive process in engineering design. Human free-hand sketches provide a useful data source for facilitating visual analogy. Although there has been research on the roles of sketching and the impact of visual analogy in design, little work has been done aiming to develop computational tools and methods to support visual analogy from sketches. In this paper, we propose a computational method to discover visual similarity between sketches, considering the following practical application: Given a sketch drawn by a designer that reflects the designer’s rough idea in mind, our goal is to identify the shape similar sketches that can stimulate the designer to make more and better visual analogies. The first challenge in doing so is how to discover the similar shape features embedded in sketches from various categories. To address this challenge, we propose a deep clustering model to learn a latent space which can reveal underlying shape features for multiple categories of sketches and cluster sketches simultaneously. An extensive evaluation of the clustering performance of our proposed method has been carried out in different configurations. The results have shown that the proposed method can discover sketches that have similar appearance, provide useful explanations of the visual relationship between different sketch categories, and has the potential to generate visual stimuli to enhance designers’ visual imageries.

Toward Computer Aided Visual Analogy Support (CAVAS): Augment Designers Through Deep Learning

2021 ◽  
Author(s):  
Zijian Zhang ◽  
Yan Jin
Keyword(s):  
Visual Information ◽  
Visual Cues ◽  
Shape Features ◽  
Computer Tool ◽  
Analogical Thinking ◽  
Visual Knowledge ◽  
Clustering Model ◽  
Latent Space ◽  
Computer Aided ◽  
Visual Analogy

Abstract The goal of this research is to develop a computer-aided visual analogy support (CAVAS) framework that can augment designers’ visual analogical thinking by providing relevant visual cues or sketches from a variety of categories and stimulating the designer to make more and better visual analogies at the ideation stage of design. The challenges of this research include what roles a computer tool should play in facilitating visual analogy of designers, what the relevant and meaningful visual analogies are at the sketching stage of design, and how the computer can capture such meaningful visual knowledge from various categories through analyzing the sketches drawn by the designers. A visual analogy support framework and a deep clustering model, called Cavas-DL, are proposed to learn a latent space of sketches that can reveal the shape patterns for multiple categories of sketches and at the same time cluster the sketches to preserve and provide category information as part of visual cues. The latent space learned serves as a visual information representation that captures the learned shape features from multiple sketch categories. The distance- and overlap-based similarities are introduced and analyzed to identify long- and short-distance analogies. Extensive evaluations of the performance of our proposed methods are carried out with different configurations, and the visual presentations of the potential analogical cues are explored. The evaluation results and the visual organizations of information have demonstrated the potential of the usefulness of the Cavas-DL model.

scholarly journals Extracting a biologically latent space of lung cancer epigenetics with variational autoencoders

2019 ◽  
Vol 20 (S18) ◽  
Author(s):  
Zhenxing Wang ◽  
Yadong Wang
Keyword(s):  
Lung Cancer ◽  
Dna Methylation ◽  
Deep Learning ◽  
Malignant Tumors ◽  
Methylation Data ◽  
Cancer Epigenetics ◽  
Latent Space ◽  
Unsupervised Deep Learning ◽  
Latent Representations ◽  
Model Training

Abstract Background Lung cancer is one of the most malignant tumors, causing over 1,000,000 deaths each year worldwide. Deep learning has brought success in many domains in recent years. DNA methylation, an epigenetic factor, is used for model training in many studies. There is an opportunity for deep learning methods to analyze the lung cancer epigenetic data to determine their subtypes for appropriate treatment. Results Here, we employ variational autoencoders (VAEs), an unsupervised deep learning framework, on 450K DNA methylation data of TCGA-LUAD and TCGA-LUSC to learn latent representations of the DNA methylation landscape. We extract a biologically relevant latent space of LUAD and LUSC samples. It is showed that the bivariate classifiers on the further compressed latent features could classify the subtypes accurately. Through clustering of methylation-based latent space features, we demonstrate that the VAEs can capture differential methylation patterns about subtypes of lung cancer. Conclusions VAEs can distinguish the original subtypes from manually mixed methylation data frame with the encoded features of latent space. Further applications about VAEs should focus on fine-grained subtypes identification for precision medicine.

A COMPARISON OF SCORING METRICS FOR PREDICTING THE NEXT NAVIGATION STEP WITH MARKOV MODEL-BASED SYSTEMS

2010 ◽  
Vol 09 (04) ◽  
pp. 547-573 ◽  
Author(s):  
JOSÉ BORGES ◽  
MARK LEVENE
Keyword(s):  
Markov Model ◽  
Prediction Accuracy ◽  
Prediction Models ◽  
Markov Models ◽  
Real Data ◽  
Absolute Error ◽  
Brier Score ◽  
Data Sets ◽  
Extensive Evaluation ◽  
The Impact

The problem of predicting the next request during a user's navigation session has been extensively studied. In this context, higher-order Markov models have been widely used to model navigation sessions and to predict the next navigation step, while prediction accuracy has been mainly evaluated with the hit and miss score. We claim that this score, although useful, is not sufficient for evaluating next link prediction models with the aim of finding a sufficient order of the model, the size of a recommendation set, and assessing the impact of unexpected events on the prediction accuracy. Herein, we make use of a variable length Markov model to compare the usefulness of three alternatives to the hit and miss score: the Mean Absolute Error, the Ignorance Score, and the Brier score. We present an extensive evaluation of the methods on real data sets and a comprehensive comparison of the scoring methods.

scholarly journals Computational Test Design for High-Speed Liquid Impact and Dispersal

2011 ◽  
Author(s):  
Alexander L. Brown ◽  
Kurt E. Metzinger
Keyword(s):  
High Speed ◽  
Water Channel ◽  
Computational Method ◽  
Computational Fluid Mechanics ◽  
Reacting Flow ◽  
Validation Data ◽  
Sled Test ◽  
The Impact ◽  
Scale Data ◽  
Limited Scope

Transportation accidents frequently involve liquids dispersing in the atmosphere. An example is that of aircraft impacts, which often result in spreading fuel and a subsequent fire. Predicting the resulting environment is of interest for design, safety, and forensic applications. This environment is challenging for many reasons, one among them being the disparate time and length scales that must be resolved for an accurate physical representation of the problem. A recent computational method appropriate for this class of problems has been developed for modeling the impact and subsequent liquid spread. This involves coupling a structural dynamics code to a turbulent computational fluid mechanics reacting flow code. Because the environment intended to be simulated with this capability is difficult to instrument and costly to test, the existing validation data are of limited scope, relevance, and quality. A rocket sled test is being performed where a scoop moving through a water channel is being used to brake a pusher sled. We plan to instrument this test to provide appropriate scale data for validating the new modeling capability. The intent is to get high fidelity data on the break-up and evaporation of the water that is ejected from the channel as the sled is braking. These two elements are critical to fireball formation for this type of event involving fuel in the place of water. We demonstrate our capability in this paper by describing the pre-test predictions which are used to locate instrumentation for the actual test. We also present a sensitivity analysis to understand the implications of length scale assumptions on the prediction results.

scholarly journals Congenital hypothyroidism and thyroid dyshormonogenesis: a case report of siblings with a newly identified mutation in thyroperoxidase

2016 ◽  
Vol 29 (5) ◽  
Author(s):  
David P. Sparling ◽  
Kendra Fabian ◽  
Lara Harik ◽  
Vaidehi Jobanputra ◽  
Kwame Anyane-Yeboa ◽  
...  
Keyword(s):  
Thyroid Cancer ◽  
Congenital Hypothyroidism ◽  
Genetic Diagnosis ◽  
Management Decisions ◽  
Whole Exome ◽  
Long Term Follow Up ◽  
Thyroid Dyshormonogenesis ◽  
Similar Appearance ◽  
The Impact

AbstractThyroid dyshormonogenesis continues to be a significant cause of congenital hypothyroidism. Over time, forms of thyroid dyshormonogenesis can result in goiter, which can lead to difficult management decisions as the pathologic changes can both mimic or lead to thyroid cancer.Herein we describe the cases of two brothers diagnosed with congenital hypothyroidism, with initial findings consistent with thyroid dyshormonogenesis. One brother eventually developed multinodular goiter with complex pathology on biopsy, resulting in thyroidectomy.Whole exome sequencing revealed the brothers carry a novel frameshift mutation in thyroperoxidase; the mutation, while not previously described, was likely both deleterious and pathogenic.These cases highlight the complex pathology that can occur within thyroid dyshormonogenesis, with similar appearance to possible thyroid cancer, leading to complex management decisions. They also highlight the role that a genetic diagnosis can play in interpreting the impact of dyshormonogenesis on nodular thyroid development, and the need for long-term follow-up in these patients.

scholarly journals consensusDE: an R package for assessing consensus of multiple RNA-seq algorithms with RUV correction

2019 ◽  
Author(s):  
Ashley J. Waardenberg ◽  
Matt A. Field
Keyword(s):  
Differential Expression ◽  
Simulated Data ◽  
R Package ◽  
Bioconductor Package ◽  
Rna Seq ◽  
User Input ◽  
Extensive Evaluation ◽  
The Impact ◽  
Transcript Database ◽  
Multiple Algorithms

AbstractExtensive evaluation of RNA-seq methods have demonstrated that no single algorithm consistently outperforms all others. Removal of unwanted variation (RUV) has also been proposed as a method for stabilizing differential expression (DE) results. Despite this, it remains a challenge to run multiple RNA-seq algorithms to identify significant differences common to multiple algorithms, whilst also integrating and assessing the impact of RUV into all algorithms. consensusDE was developed to automate the process of identifying significant DE by combining the results from multiple algorithms with minimal user input and with the option to automatically integrate RUV. consensusDE only requires a table describing the sample groups, a directory containing BAM files or preprocessed count tables and an optional transcript database for annotation. It supports merging of technical replicates, paired analyses and outputs a compendium of plots to guide the user in subsequent analyses. Herein, we also assess the ability of RUV to improve DE stability when combined with multiple algorithms through application to real and simulated data. We find that, although RUV demonstrated improved FDR in a setting of low replication, the effect was algorithm specific and diminished with increased replication, reinforcing increased replication for recovery of true DE genes. We finish by offering some rules and considerations for the application of RUV in a consensus-based setting.consensusDE is freely available, implemented in R and available as a Bioconductor package, under the GPL-3 license, along with a comprehensive vignette describing functionality: http://bioconductor.org/packages/consensusDE/

scholarly journals A Lightweight Passive Human Tracking Method Using Wi-Fi

Sensors ◽  
2022 ◽  
Vol 22 (2) ◽  
pp. 541
Author(s):  
Jian Fang ◽  
Lei Wang ◽  
Zhenquan Qin ◽  
Bingxian Lu ◽  
Wenbo Zhao ◽  
...  
Keyword(s):  
Target Tracking ◽  
Human Motion ◽  
Wireless Devices ◽  
Target Movement ◽  
Channel State ◽  
Time Model ◽  
Tracking Method ◽  
Extensive Evaluation ◽  
Device Free ◽  
The Impact

Target tracking is a critical technique for localization in an indoor environment. Current target-tracking methods suffer from high overhead, high latency, and blind spots issues due to a large amount of data needing to be collected or trained. On the other hand, a lightweight tracking method is preferred in many cases instead of just pursuing accuracy. For this reason, in this paper, we propose a Wi-Fi-enabled Infrared-like Device-free (WIDE) method for target tracking to realize a lightweight target-tracking method. We first analyze the impact of target movement on the physical layer of the wireless link and establish a near real-time model between the Channel State Information (CSI) and human motion. Secondly, we make full use of the network structure formed by a large number of wireless devices already deployed in reality to achieve the goal. We validate the WIDE method in different environments. Extensive evaluation results show that the WIDE method is lightweight and can track targets rapidly as well as achieve satisfactory tracking results.

Sign in / Sign up

Export Citation Format

Share Document