About: Towards Explainable Deep Neural Networks (xDNN)

Facets (new session)
Description
Metadata
Settings
- owl:sameAs
- Inference Rule:

About: Towards Explainable Deep Neural Networks (xDNN) Goto Sponge NotDistinct Permalink

An Entity of Type : schema:ScholarlyArticle, within Data Space : wasabi.inria.fr associated with source document(s)

Attributes	Values
type	Academic Article research paper schema:ScholarlyArticle
isDefinedBy	Covid-on-the-Web dataset
has title	Towards Explainable Deep Neural Networks (xDNN)
Creator	Angelov, Plamen Soares, Eduardo
Source	ArXiv
abstract	In this paper, we propose an elegant solution that is directly addressing the bottlenecks of the traditional deep learning approaches and offers a clearly explainable internal architecture that can outperform the existing methods, requires very little computational resources (no need for GPUs) and short training times (in the order of seconds). The proposed approach, xDNN is using prototypes. Prototypes are actual training data samples (images), which are local peaks of the empirical data distribution called typicality as well as of the data density. This generative model is identified in a closed form and equates to the pdf but is derived automatically and entirely from the training data with no user- or problem-specific thresholds, parameters or intervention. The proposed xDNN offers a new deep learning architecture that combines reasoning and learning in a synergy. It is non-iterative and non-parametric, which explains its efficiency in terms of time and computational resources. From the user perspective, the proposed approach is clearly understandable to human users. We tested it on some well-known benchmark data sets such as iRoads and Caltech-256. xDNN outperforms the other methods including deep learning in terms of accuracy, time to train and offers a clearly explainable classifier. In fact, the result on the very hard Caltech-256 problem (which has 257 classes) represents a world record.
has issue date	2019-12-05(xsd:dateTime)
has license	arxiv
sha1sum (hex)	5660c5ec41ab8e76b21f3ffafa5f4f663914ac7f
resource representing a document's title	Towards Explainable Deep Neural Networks (xDNN)
resource representing a document's body	covid:5660c5ec41ab8e76b21f3ffafa5f4f663914ac7f#body_text
is schema:about of	named entity 'CLASSIFIER' named entity 'CLASSES' named entity 'FORM' named entity 'ACTUAL' covid:arg/5660c5ec41ab8e76b21f3ffafa5f4f663914ac7f named entity 'data sets' named entity 'closed form' named entity 'classifier' named entity 'generative model' named entity 'Caltech' named entity 'data points' named entity 'pdf' named entity 'pre-trained' named entity 'iterative optimization' named entity 'synthetic data' named entity 'deep neural network' named entity 'state-of-the-art' named entity 'image classification' named entity 'Softmax' named entity 'kNN' named entity 'decision-making' named entity 'vector' named entity 'black-box' named entity 'algorithm' named entity 'data samples' named entity 'pre-trained' named entity 'prototype' named entity 'deep learning' named entity 'empirical data' named entity 'data sets' named entity 'HEN' named entity 'GPUs' named entity 'closed form' named entity 'decision-making' named entity 'prototype' named entity 'labeled data' named entity 'deep neural networks' named entity 'fully connected' named entity 'disjunction' named entity 'prototype' named entity 'fully connected' named entity 'Caltech' named entity 'data samples' named entity 'state-of-the-art' named entity 'feature extraction' named entity 'pdf' named entity 'high-level features' named entity 'vector' named entity '1, 2' named entity 'image classification' named entity 'explainability' named entity 'SVM' named entity 'MacOS' named entity 'pdf' named entity 'GPU' named entity 'deep learning' named entity 'decision-making' named entity 'complex problems' named entity 'image recognition' named entity 'data points' named entity 'Voronoi tesselation' named entity 'explainability' named entity 'classifiers' named entity 'vector' named entity 'probability distribution function' named entity 'prototype-based' named entity 'closed form' named entity 'variance' named entity 'deep learning' named entity 'prototype' named entity 'local optimization' named entity 'analytical form' named entity 'explainability' named entity 'Future research' named entity 'point of view' named entity 'Deep learning' named entity 'algorithm' named entity 'algorithm' named entity 'data points' named entity 'classifier' named entity 'data samples' named entity 'convolutional neural network' named entity 'prototype' named entity 'autonomous cars' named entity 'pre-trained' named entity 'Data samples'

Faceted Search & Find service v1.13.91 as of Mar 24 2020

Alternative Linked Data Documents: Sponger | ODE Content Formats:

RDF

ODATA

Microdata

About

OpenLink Virtuoso version 07.20.3229 as of Jul 10 2020, on Linux (x86_64-pc-linux-gnu), Single-Server Edition (94 GB total memory)
Data on this page belongs to its respective rights holders.
Virtuoso Faceted Browser Copyright © 2009-2024 OpenLink Software