scholarly journals Explaining Deep Learning using examples: Optimal feature weighting methods for twin systems using post-hoc, explanation-by-example in XAI

2021 ◽  
Vol 233 ◽  
pp. 107530
Author(s):  
Eoin M. Kenny ◽  
Mark T. Keane
Keyword(s):  
Deep Learning ◽  
Feature Weighting ◽  
Post Hoc ◽  
Optimal Feature ◽  
Weighting Methods

scholarly journals Deep learning for visualization and novelty detection in large X-ray diffraction datasets

2021 ◽  
Vol 7 (1) ◽  
Author(s):  
Lars Banko ◽  
Phillip M. Maffettone ◽  
Dennis Naujoks ◽  
Daniel Olds ◽  
Alfred Ludwig
Keyword(s):  
Thin Film ◽  
Crystal Structure ◽  
Artificial Intelligence ◽  
Deep Learning ◽  
Novelty Detection ◽  
Structural Similarity ◽  
X Ray Diffraction ◽  
X Ray ◽  
Diffraction Patterns ◽  
Post Hoc

AbstractWe apply variational autoencoders (VAE) to X-ray diffraction (XRD) data analysis on both simulated and experimental thin-film data. We show that crystal structure representations learned by a VAE reveal latent information, such as the structural similarity of textured diffraction patterns. While other artificial intelligence (AI) agents are effective at classifying XRD data into known phases, a similarly conditioned VAE is uniquely effective at knowing what it doesn’t know: it can rapidly identify data outside the distribution it was trained on, such as novel phases and mixtures. These capabilities demonstrate that a VAE is a valuable AI agent for aiding materials discovery and understanding XRD measurements both ‘on-the-fly’ and during post hoc analysis.

Enhancement of email spam detection using improved deep learning algorithms for cyber security

2021 ◽  
pp. 1-34
Author(s):  
Kadam Vikas Samarthrao ◽  
Vandana M. Rohokale
Keyword(s):  
Feature Selection ◽  
Deep Learning ◽  
Visual Features ◽  
Spam Detection ◽  
Learning Approaches ◽  
Learning Technique ◽  
Text Features ◽  
Optimal Feature Selection ◽  
Optimal Feature ◽  
Email Spam

Email has sustained to be an essential part of our lives and as a means for better communication on the internet. The challenge pertains to the spam emails residing a large amount of space and bandwidth. The defect of state-of-the-art spam filtering methods like misclassification of genuine emails as spam (false positives) is the rising challenge to the internet world. Depending on the classification techniques, literature provides various algorithms for the classification of email spam. This paper tactics to develop a novel spam detection model for improved cybersecurity. The proposed model involves several phases like dataset acquisition, feature extraction, optimal feature selection, and detection. Initially, the benchmark dataset of email is collected that involves both text and image datasets. Next, the feature extraction is performed using two sets of features like text features and visual features. In the text features, Term Frequency-Inverse Document Frequency (TF-IDF) is extracted. For the visual features, color correlogram and Gray-Level Co-occurrence Matrix (GLCM) are determined. Since the length of the extracted feature vector seems to the long, the optimal feature selection process is done. The optimal feature selection is performed by a new meta-heuristic algorithm called Fitness Oriented Levy Improvement-based Dragonfly Algorithm (FLI-DA). Once the optimal features are selected, the detection is performed by the hybrid learning technique that is composed of two deep learning approaches named Recurrent Neural Network (RNN) and Convolutional Neural Network (CNN). For improving the performance of existing deep learning approaches, the number of hidden neurons of RNN and CNN is optimized by the same FLI-DA. Finally, the optimized hybrid learning technique having CNN and RNN classifies the data into spam and ham. The experimental outcomes show the ability of the proposed method to perform the spam email classification based on improved deep learning.

scholarly journals A Stacked Fully Convolutional Networks with Feature Alignment Framework for Multi-Label Land-cover Segmentation

2019 ◽  
Vol 11 (9) ◽  
pp. 1051 ◽  
Author(s):  
Guangming Wu ◽  
Yimin Guo ◽  
Xiaoya Song ◽  
Zhiling Guo ◽  
Haoran Zhang ◽  
...  
Keyword(s):  
Deep Learning ◽  
Land Cover ◽  
Superior Performance ◽  
Learning Methods ◽  
Convolutional Networks ◽  
Fully Convolutional Networks ◽  
Optimal Feature ◽  
Imaging Conditions ◽  
Very High ◽  
Feature Alignment

Applying deep-learning methods, especially fully convolutional networks (FCNs), has become a popular option for land-cover classification or segmentation in remote sensing. Compared with traditional solutions, these approaches have shown promising generalization capabilities and precision levels in various datasets of different scales, resolutions, and imaging conditions. To achieve superior performance, a lot of research has focused on constructing more complex or deeper networks. However, using an ensemble of different fully convolutional models to achieve better generalization and to prevent overfitting has long been ignored. In this research, we design four stacked fully convolutional networks (SFCNs), and a feature alignment framework for multi-label land-cover segmentation. The proposed feature alignment framework introduces an alignment loss of features extracted from basic models to balance their similarity and variety. Experiments on a very high resolution(VHR) image dataset with six categories of land-covers indicates that the proposed SFCNs can gain better performance when compared to existing deep learning methods. In the 2nd variant of SFCN, the optimal feature alignment gains increments of 4.2% (0.772 vs. 0.741), 6.8% (0.629 vs. 0.589), and 5.5% (0.727 vs. 0.689) for its f1-score, jaccard index, and kappa coefficient, respectively.

scholarly journals Modular deep learning enables automated identification of monoclonal cell lines

2020 ◽  
Author(s):  
Brodie Fischbacher ◽  
Sarita Hedaya ◽  
Brigham J. Hartley ◽  
Zhongwei Wang ◽  
Gregory Lallos ◽  
...  
Keyword(s):  
Deep Learning ◽  
Cell Lines ◽  
Gene Editing ◽  
Algorithm Design ◽  
Automated Identification ◽  
Technical Variability ◽  
A Cell ◽  
Or Gene ◽  
Induced Pluripotent ◽  
Post Hoc

AbstractMonoclonalization refers to the isolation and expansion of a single cell derived from a cultured population. This is a valuable step in cell culture so as to minimize a cell line’s technical variability downstream of cell-altering events, such as reprogramming or gene editing, as well as for processes such as monoclonal antibody development. However, traditional methods for verifying clonality do not scale well, posing a critical obstacle to studies involving large cohorts. Without automated, standardized methods for assessing clonality post-hoc, methods involving monoclonalization cannot be reliably upscaled without exacerbating the technical variability of cell lines. We report the design of a deep learning workflow that automatically detects colony presence and identifies clonality from cellular imaging. The workflow, termed Monoqlo, integrates multiple convolutional neural networks and, critically, leverages the chronological directionality of the cell culturing process. Our algorithm design provides a fully scalable, highly interpretable framework, capable of analyzing industrial data volumes in under an hour using commodity hardware. In the present study, we focus on monoclonalization of human induced pluripotent stem cells (HiPSCs) as a case example. Monoqlo standardizes the monoclonalization process, enabling colony selection protocols to be infinitely upscaled while minimizing technical variability.

scholarly journals Local Post-hoc Explainable Methods for Adversarial Text Attacks

2021 ◽  
Author(s):  
Yidong Chai ◽  
Ruicheng Liang ◽  
Hongyi Zhu ◽  
Sagar Samtani ◽  
Meng Wang ◽  
...  
Keyword(s):  
Deep Learning ◽  
Natural Language Processing ◽  
Language Processing ◽  
Black Box ◽  
Learning Models ◽  
Two Phase ◽  
Sensitivity Estimation ◽  
Execution Phase ◽  
Adversarial Examples ◽  
Post Hoc

Deep learning models have significantly advanced various natural language processing tasks. However, they are strikingly vulnerable to adversarial text attacks, even in the black-box setting where no model knowledge is accessible to hackers. Such attacks are conducted with a two-phase framework: 1) a sensitivity estimation phase to evaluate each element’s sensitivity to the target model’s prediction, and 2) a perturbation execution phase to craft the adversarial examples based on estimated element sensitivity. This study explored the connections between the local post-hoc explainable methods for deep learning and black-box adversarial text attacks and proposed a novel eXplanation-based method for crafting Adversarial Text Attacks (XATA). XATA leverages local post-hoc explainable methods (e.g., LIME or SHAP) to measure input elements’ sensitivity and adopts the word replacement perturbation strategy to craft adversarial examples. We evaluated the attack performance of the proposed XATA on three commonly used text-based datasets: IMDB Movie Review, Yelp Reviews-Polarity, and Amazon Reviews-Polarity. The proposed XATA outperformed existing baselines in various target models, including LSTM, GRU, CNN, and BERT. Moreover, we found that improved local post-hoc explainable methods (e.g., SHAP) lead to more effective adversarial attacks. These findings showed that when researchers constantly advance the explainability of deep learning models with local post-hoc methods, they also provide hackers with weapons to craft more targeted and dangerous adversarial attacks.

scholarly journals Quantifying the Utility of Low Resolution Brain Imaging:: Treatment of Infant Hydrocephalus

2021 ◽  
Author(s):  
Joshua Harper ◽  
Venkateswararao Cherukuri ◽  
Tom O'Riley ◽  
Mingzhao Yu ◽  
Edith Mbabazi-Kabachelor ◽  
...  
Keyword(s):  
Deep Learning ◽  
Image Quality ◽  
Treatment Planning ◽  
Image Resolution ◽  
Learning Enhancement ◽  
Low Contrast ◽  
Substantial Risk ◽  
Noise Ratio ◽  
Hydrocephalus Treatment ◽  
Post Hoc

As low-field MRI technology is being disseminated into clinical settings, it is important to assess the image quality required to properly diagnose and treat a given disease. In this post-hoc analysis of an ongoing randomized clinical trial, we assessed the diagnostic utility of reduced-quality and deep learning enhanced images for hydrocephalus treatment planning. Images were degraded in terms of resolution, noise, and contrast between brain and CSF and enhanced using deep learning algorithms. Both degraded and enhanced images were presented to three experienced pediatric neurosurgeons accustomed to working in LMIC for assessment of clinical utility in treatment planning for hydrocephalus. Results indicate that image resolution and contrast-to-noise ratio between brain and CSF predict the likelihood of a useful image for hydrocephalus treatment planning. For images with 128x128 resolution, a contrast-to-noise ratio of 2.5 has a high probability of being useful (91%, 95% CI 73% to 96%; P=2e-16). Deep learning enhancement of a 128x128 image with very low contrast-to-noise (1.5) and low probability of being useful (23%, 95% CI 14% to 36%; P=2e-16) increases CNR improving the apparent likelihood of being useful, but carries substantial risk of structural errors leading to misleading clinical interpretation (CNR after enhancement = 5; risk of misleading results = 21%, 95% CI 3% to 32%; P=7e-11). Lower quality images not customarily considered acceptable by clinicians can be useful in planning hydrocephalus treatment. We find substantial risk of misleading structural errors when using deep learning enhancement of low quality images. These findings advocate for new standards in assessing acceptable image quality for clinical use.

scholarly journals FLEX: Faithful Linguistic Explanations for Neural Net Based Model Decisions

2019 ◽  
Vol 33 ◽  
pp. 2539-2546
Author(s):  
Sandareka Wickramanayake ◽  
Wynne Hsu ◽  
Mong Li Lee
Keyword(s):  
Neural Network ◽  
Deep Learning ◽  
State Of The Art ◽  
Neural Net ◽  
End User ◽  
Learning Network ◽  
Benchmark Datasets ◽  
Deep Learning Network ◽  
Post Hoc ◽  
User Trust

Explaining the decisions of a Deep Learning Network is imperative to safeguard end-user trust. Such explanations must be intuitive, descriptive, and faithfully explain why a model makes its decisions. In this work, we propose a framework called FLEX (Faithful Linguistic EXplanations) that generates post-hoc linguistic justifications to rationalize the decision of a Convolutional Neural Network. FLEX explains a model’s decision in terms of features that are responsible for the decision. We derive a novel way to associate such features to words, and introduce a new decision-relevance metric that measures the faithfulness of an explanation to a model’s reasoning. Experiment results on two benchmark datasets demonstrate that the proposed framework can generate discriminative and faithful explanations compared to state-of-the-art explanation generators. We also show how FLEX can generate explanations for images of unseen classes as well as automatically annotate objects in images.

Sign in / Sign up

Export Citation Format

Share Document