Visual Analytics for Explainable Deep Learning

How Does Attention Work in Vision Transformers? A Visual Analytics Attempt

Vision transformer (ViT) expands the success of transformer models from sequential data to images. The model decomposes an image into many smaller patches and arranges them into a sequence. Multi-head self-attentions are then applied to the sequence to learn the attention between patches. Despite many successful interpretations of transformers on sequential data, little effort has been devoted to the interpretation of ViTs, and many questions remain unanswered. For example, among the numerous attention heads, which one is more important? How strong are individual patches attending to their spatial neighbors in different heads? What attention patterns have individual heads learned? In this work, we answer these questions through a visual analytics approach. Specifically, we first identify what heads are more important in ViTs by introducing multiple pruning-based metrics. Then, we profile the spatial distribution of attention strengths between patches inside individual heads, as well as the trend of attention strengths across attention layers. Third, using an autoencoder-based learning solution, we summarize all possible attention patterns that individual heads could learn. Examining the attention strengths and patterns of the important heads, we answer why they are important. Through concrete case studies with experienced deep learning experts on multiple ViTs, we validate the effectiveness of our solution that deepens the understanding of ViTs from head importance, head attention strength, and head attention pattern.

Visual Analytics of Neuron Vulnerability to Adversarial Attacks on Convolutional Neural Networks

Adversarial attacks on a convolutional neural network (CNN)—injecting human-imperceptible perturbations into an input image—could fool a high-performance CNN into making incorrect predictions. The success of adversarial attacks raises serious concerns about the robustness of CNNs, and prevents them from being used in safety-critical applications, such as medical diagnosis and autonomous driving. Our work introduces a visual analytics approach to understanding adversarial attacks by answering two questions: (1) which neurons are more vulnerable to attacks and (2) which image features do these vulnerable neurons capture during the prediction? For the first question, we introduce multiple perturbation-based measures to break down the attacking magnitude into individual CNN neurons and rank the neurons by their vulnerability levels. For the second, we identify image features (e.g., cat ears) that highly stimulate a user-selected neuron to augment and validate the neuron’s responsibility. Furthermore, we support an interactive exploration of a large number of neurons by aiding with hierarchical clustering based on the neurons’ roles in the prediction. To this end, a visual analytics system is designed to incorporate visual reasoning for interpreting adversarial attacks. We validate the effectiveness of our system through multiple case studies as well as feedback from domain experts.

Visualization and Visual Analytics Approaches for Image and Video Datasets: A Survey

Image and video data analysis has become an increasingly important research area with applications in different domains such as security surveillance, healthcare, augmented and virtual reality, video and image editing, activity analysis and recognition, synthetic content generation, distance education, telepresence, remote sensing, sports analytics, art, non-photorealistic rendering, search engines, and social media. Recent advances in Artificial Intelligence (AI) and particularly deep learning have sparked new research challenges and led to significant advancements, especially in image and video analysis. These advancements have also resulted in significant research and development in other areas such as visualization and visual analytics, and have created new opportunities for future lines of research. In this survey paper, we present the current state of the art at the intersection of visualization and visual analytics, and image and video data analysis. We categorize the visualization papers included in our survey based on different taxonomies used in visualization and visual analytics research. We review these papers in terms of task requirements, tools, datasets, and application areas. We also discuss insights based on our survey results, trends and patterns, the current focus of visualization research, and opportunities for future research.

Visual Analytics for RNN-Based Deep Reinforcement Learning

Deep reinforcement learning (DRL) targets to train an autonomous agent to interact with a pre-defined environment and strives to achieve specific goals through deep neural networks (DNN). Recurrent neural network (RNN) based DRL has demonstrated superior performance, as RNNs can effectively capture the temporal evolution of the environment and respond with proper agent actions. However, apart from the outstanding performance, little is known about how RNNs understand the environment internally and what has been memorized over time. Revealing these details is extremely important for deep learning experts to understand and improve DRLs, which in contrast, is also challenging due to the complicated data transformations inside these models. In this article, we propose Deep Reinforcement Learning Interactive Visual Explorer (DRLIVE), a visual analytics system to effectively explore, interpret, and diagnose RNN-based DRLs. Having focused on DRL agents trained for different Atari games, DRLIVE accomplishes three tasks: game episode exploration, RNN hidden/cell state examination, and interactive model perturbation. Using the system, one can flexibly explore a DRL agent through interactive visualizations, discover interpretable RNN cells by prioritizing RNN hidden/cell states with a set of metrics, and further diagnose the DRL model by interactively perturbing its inputs. Through concrete studies with multiple deep learning experts, we validated the efficacy of DRLIVE.

Survey on Visual Analysis of Event Sequence Data

Event sequence data record series of discrete events in the time order of occurrence. They are commonly observed in a variety of applications ranging from electronic health records to network logs, with the characteristics of large-scale, high-dimensional and heterogeneous. This high complexity of event sequence data makes it difficult for analysts to manually explore and find patterns, resulting in ever-increasing needs for computational and perceptual aids from visual analytics techniques to extract and communicate insights from event sequence datasets. In this paper, we review the state-of-the-art visual analytics approaches, characterize them with our proposed design space, and categorize them based on analytical tasks and applications. From our review of relevant literature, we have also identified several remaining research challenges and future research opportunities.

Explaining Semi-Supervised Text Alignment Through Visualization

The analysis of variance in complex text traditions is an arduous task when carried out manually. Text alignment algorithms provide domain experts with a robust alternative to such repetitive tasks. Existing white-box approaches allow the digital humanities to establish syntax-based metrics taking into account the spelling, morphology and order of words. However, they produce limited results, as semantic meanings are typically not taken into account. Our interdisciplinary collaboration between visualization and digital humanities combined a semi-supervised text alignment approach based on word embeddings that take not only syntactic but also semantic text features into account, thereby improving the overall quality of the alignment. In our collaboration, we developed different visual interfaces that communicate the word distribution in high-dimensional vector space generated by the underlying neural network for increased transparency, assessment of the tool's reliability and overall improved hypothesis generation. We further offer visual means to enable the expert reader to feed domain knowledge into the system at multiple levels with the aim of improving both the product and the process of text alignment. This ultimately illustrates how visualization can engage with and augment complex modes of reading in the humanities.

A walk in the black-box: 3D visualization of large neural networks in virtual reality

Within the last decade Deep Learning has become a tool for solving challenging problems like image recognition. Still, Convolutional Neural Networks (CNNs) are considered black-boxes, which are difficult to understand by humans. Hence, there is an urge to visualize CNN architectures, their internal processes and what they actually learn. Previously, virtual realityhas been successfully applied to display small CNNs in immersive 3D environments. In this work, we address the problem how to feasibly render large-scale CNNs, thereby enabling the visualization of popular architectures with ten thousands of feature maps and branches in the computational graph in 3D. Our software ”DeepVisionVR” enables the user to freely walk through the layered network, pick up and place images, move/scale layers for better readability, perform feature visualization and export the results. We also provide a novel Pytorch module to dynamically link PyTorch with Unity, which gives developers and researchers a convenient interface to visualize their own architectures. The visualization is directly created from the PyTorch class that defines the Pytorch model used for training and testing. This approach allows full access to the network’s internals and direct control over what exactly is visualized. In a use-case study, we apply the module to analyze models with different generalization abilities in order to understand how networks memorize images. We train two recent architectures, CovidResNet and CovidDenseNet on the Caltech101 and the SARS-CoV-2 datasets and find that bad generalization is driven by high-frequency features and the susceptibility to specific pixel arrangements, leading to implications for the practical application of CNNs. The code is available on Github https://github.com/Criscraft/DeepVisionVR.

EEG-Based Emotion Classification Using Improved Cross-Connected Convolutional Neural Network

The use of electroencephalography to recognize human emotions is a key technology for advancing human–computer interactions. This study proposes an improved deep convolutional neural network model for emotion classification using a non-end-to-end training method that combines bottom-, middle-, and top-layer convolution features. Four sets of experiments using 4500 samples were conducted to verify model performance. Simultaneously, feature visualization technology was used to extract the three-layer features obtained by the model, and a scatterplot analysis was performed. The proposed model achieved a very high accuracy of 93.7%, and the extracted features exhibited the best separability among the tested models. We found that adding redundant layers did not improve model performance, and removing the data of specific channels did not significantly reduce the classification effect of the model. These results indicate that the proposed model allows for emotion recognition with a higher accuracy and speed than the previously reported models. We believe that our approach can be implemented in various applications that require the quick and accurate identification of human emotions.

Visualization and Analysis Model of Industrial Economy Status and Development Based on Knowledge Graph and Deep Neural Network

This paper adopts knowledge mapping combined with a deep neural network algorithm to conduct in-depth research and analysis on the current situation and development of the industrial economy and designs a visual analysis model of economic development based on knowledge mapping combined with a deep neural network algorithm. Cultivate the concept of coordinated development and legal system of the subject, improve the awareness of network security and integrity self-discipline of the subject, improve the level of network hardware equipment manufacturing, improve the level of network platform construction, build a network security technology prevention system, improve the repair system of network information alienation, set up a specialized agency setting for the coordinated development of network ecology and industrial economy, and increase the capital investment in network infrastructure and network information technology research and development. A framework of breadth and depth recommendation ranking based on a knowledge graph is proposed and implemented. This paper provides a visual analysis method to sort and classify multivariate data. The method first determines users' preferences through their interactive operations, calculates the weights of each attribute according to the users' preference model, then uses the obtained attribute weight sets to sort the whole data set, and finally completes the category classification according to the sorting results and the users' markings on some data. The visual display allows users to intuitively perform data sorting and classification operations and quickly understand the characteristics and category features of the data. The framework achieves modeling and integration of knowledge graph neighborhood information from breadth dimension and depth dimension to realize personalized recommendation sorting and improves the F1 metrics by 8.59%, 14.36%, and 15.22% on the public datasets Amazon-book, Yelp2018, and ILast-FM compared with the previous optimal model.

TorchEsegeta: Framework for Interpretability and Explainability of Image-Based Deep Learning Models

Clinicians are often very sceptical about applying automatic image processing approaches, especially deep learning-based methods, in practice. One main reason for this is the black-box nature of these approaches and the inherent problem of missing insights of the automatically derived decisions. In order to increase trust in these methods, this paper presents approaches that help to interpret and explain the results of deep learning algorithms by depicting the anatomical areas that influence the decision of the algorithm most. Moreover, this research presents a unified framework, TorchEsegeta, for applying various interpretability and explainability techniques for deep learning models and generates visual interpretations and explanations for clinicians to corroborate their clinical findings. In addition, this will aid in gaining confidence in such methods. The framework builds on existing interpretability and explainability techniques that are currently focusing on classification models, extending them to segmentation tasks. In addition, these methods have been adapted to 3D models for volumetric analysis. The proposed framework provides methods to quantitatively compare visual explanations using infidelity and sensitivity metrics. This framework can be used by data scientists to perform post hoc interpretations and explanations of their models, develop more explainable tools, and present the findings to clinicians to increase their faith in such models. The proposed framework was evaluated based on a use case scenario of vessel segmentation models trained on Time-of-Flight (TOF) Magnetic Resonance Angiogram (MRA) images of the human brain. Quantitative and qualitative results of a comparative study of different models and interpretability methods are presented. Furthermore, this paper provides an extensive overview of several existing interpretability and explainability methods.

Class imbalance learning with Bayesian optimization applied in drug discovery

Machine intelligence (MI), including machine learning and deep learning, have been regarded as promising methods to reduce the prohibitively high cost of drug development. However, a dilemma within MI has limited its wide application: machine learning models are easier to interpret but yield worse predictive performance than deep learning models. Therefore, we propose a pipeline called Class Imbalance Learning with Bayesian Optimization (CILBO) to improve the performance of machine learning models in drug discovery. To demonstrate the efficacy of the CILBO pipeline, we developed an example model to predict antibacterial candidates. Comparison of the antibacterial prediction performance between our model and a well-known deep learning model published by Stokes et al. suggests that our model can perform as well as the deep learning model in drug activity prediction. The CILBO pipeline we propose provides a simple, alternative approach to accelerate preliminary screenings and decrease the cost of drug discovery.

Applications of artificial intelligence in the thorax: a narrative review focusing on thoracic radiology

Objective

This review will focus on how AI—and, specifically, deep learning—can be applied to complement aspects of the current healthcare system. We describe how AI-based tools can augment existing clinical workflows by discussing the applications of AI to worklist prioritization and patient triage, the performance-boosting effects of AI as a second reader, and the use of AI to facilitate complex quantifications. We also introduce prominent examples of recent AI applications, such as tuberculosis screening in resource-constrained environments, the detection of lung cancer with screening CT, and the diagnosis of COVID-19. We also provide examples of prognostic predictions and new discoveries beyond existing clinical practices.

Background

Artificial intelligence (AI) has shown promising performance for thoracic diseases, particularly in the field of thoracic radiology. However, it has not yet been established how AI-based image analysis systems can help physicians in clinical practice.

Methods

This review included peer-reviewed research articles on AI in the thorax published in English between 2015 and 2021.

Conclusions

With advances in technology and appropriate preparation of physicians, AI could address various clinical problems that have not been solved due to a lack of clinical resources or technological limitations.

Keywords

Artificial intelligence (AI); deep learning (DL); computer aided diagnosis (CAD); thoracic radiology; pulmonary medicine

Visual Analytics: A Method to Explore Natural Histories of Oral Epithelial Dysplasia

Risk assessment and follow-up of oral potentially malignant disorders in patients with mild or moderate oral epithelial dysplasia is an ongoing challenge for improved oral cancer prevention. Part of the challenge is a lack of understanding of how observable features of such dysplasia, gathered as data by clinicians during follow-up, relate to underlying biological processes driving progression. Current research is at an exploratory phase where the precise questions to ask are not known. While traditional statistical and the newer machine learning and artificial intelligence methods are effective in well-defined problem spaces with large datasets, these are not the circumstances we face currently. We argue that the field is in need of exploratory methods that can better integrate clinical and scientific knowledge into analysis to iteratively generate viable hypotheses. In this perspective, we propose that visual analytics presents a set of methods well-suited to these needs. We illustrate how visual analytics excels at generating viable research hypotheses by describing our experiences using visual analytics to explore temporal shifts in the clinical presentation of epithelial dysplasia. Visual analytics complements existing methods and fulfills a critical and at-present neglected need in the formative stages of inquiry we are facing.

Towards Visual Explainable Active Learning for Zero-Shot Classification

Zero-shot classification is a promising paradigm to solve an applicable problem when the training classes and test classes are disjoint. Achieving this usually needs experts to externalize their domain knowledge by manually specifying a class-attribute matrix to define which classes have which attributes. Designing a suitable class-attribute matrix is the key to the subsequent procedure, but this design process is tedious and trial-and-error with no guidance. This paper proposes a visual explainable active learning approach with its design and implementation called semantic navigator to solve the above problems. This approach promotes human-AI teaming with four actions (ask, explain, recommend, respond) in each interaction loop. The machine asks contrastive questions to guide humans in the thinking process of attributes. A novel visualization called semantic map explains the current status of the machine. Therefore analysts can better understand why the machine misclassifies objects. Moreover, the machine recommends the labels of classes for each attribute to ease the labeling burden. Finally, humans can steer the model by modifying the labels interactively, and the machine adjusts its recommendations. The visual explainable active learning approach improves humans' efficiency of building zero-shot classification models interactively, compared with the method without guidance. We justify our results with user studies using the standard benchmarks for zero-shot classification.

Learn, Generate, Rank, Explain: A Case Study of Visual Explanation by Generative Machine Learning

While the computer vision problem of searching for activities in videos is usually addressed by using discriminative models, their decisions tend to be opaque and difficult for people to understand. We propose a case study of a novel machine learning approach for generative searching and ranking of motion capture activities with visual explanation. Instead of directly ranking videos in the database given a text query, our approach uses a variant of Generative Adversarial Networks (GANs) to generate exemplars based on the query and uses them to search for the activity of interest in a large database. Our model is able to achieve comparable results to its discriminative counterpart, while being able to dynamically generate visual explanations. In addition to our searching and ranking method, we present an explanation interface that enables the user to successfully explore the model’s explanations and its confidence by revealing query-based, model-generated motion capture clips that contributed to the model’s decision. Finally, we conducted a user study with 44 participants to show that by using our model and interface, participants benefit from a deeper understanding of the model’s conceptualization of the search query. We discovered that the XAI system yielded a comparable level of efficiency, accuracy, and user-machine synchronization as its black-box counterpart, if the user exhibited a high level of trust for AI explanation.

A Multidisciplinary Survey and Framework for Design and Evaluation of Explainable AI Systems

The need for interpretable and accountable intelligent systems grows along with the prevalence of artificial intelligence ( AI ) applications used in everyday life. Explainable AI ( XAI ) systems are intended to self-explain the reasoning behind system decisions and predictions. Researchers from different disciplines work together to define, design, and evaluate explainable systems. However, scholars from different disciplines focus on different objectives and fairly independent topics of XAI research, which poses challenges for identifying appropriate design and evaluation methodology and consolidating knowledge across efforts. To this end, this article presents a survey and framework intended to share knowledge and experiences of XAI design and evaluation methods across multiple disciplines. Aiming to support diverse design goals and evaluation methods in XAI research, after a thorough review of XAI related papers in the fields of machine learning, visualization, and human-computer interaction, we present a categorization of XAI design goals and evaluation methods. Our categorization presents the mapping between design goals for different XAI user groups and their evaluation methods. From our findings, we develop a framework with step-by-step design guidelines paired with evaluation methods to close the iterative design and evaluation cycles in multidisciplinary XAI teams. Further, we provide summarized ready-to-use tables of evaluation methods and recommendations for different goals in XAI research.

Visual analytics tool for the interpretation of hidden states in recurrent neural networks

In this paper, we introduce a visual analytics approach aimed at helping machine learning experts analyze the hidden states of layers in recurrent neural networks. Our technique allows the user to interactively inspect how hidden states store and process information throughout the feeding of an input sequence into the network. The technique can help answer questions, such as which parts of the input data have a higher impact on the prediction and how the model correlates each hidden state configuration with a certain output. Our visual analytics approach comprises several components: First, our input visualization shows the input sequence and how it relates to the output (using color coding). In addition, hidden states are visualized through a nonlinear projection into a 2-D visualization space using t-distributed stochastic neighbor embedding to understand the shape of the space of the hidden states. Trajectories are also employed to show the details of the evolution of the hidden state configurations. Finally, a time-multi-class heatmap matrix visualizes the evolution of the expected predictions for multi-class classifiers, and a histogram indicates the distances between the hidden states within the original space. The different visualizations are shown simultaneously in multiple views and support brushing-and-linking to facilitate the analysis of the classifications and debugging for misclassified input sequences. To demonstrate the capability of our approach, we discuss two typical use cases for long short-term memory models applied to two widely used natural language processing datasets.

Explainable Artificial Intelligence for Human Decision Support System in the Medical Domain

In this paper, we present the potential of Explainable Artificial Intelligence methods for decision support in medical image analysis scenarios. Using three types of explainable methods applied to the same medical image data set, we aimed to improve the comprehensibility of the decisions provided by the Convolutional Neural Network (CNN). In vivo gastral images obtained by a video capsule endoscopy (VCE) were the subject of visual explanations, with the goal of increasing health professionals’ trust in black-box predictions. We implemented two post hoc interpretable machine learning methods, called Local Interpretable Model-Agnostic Explanations (LIME) and SHapley Additive exPlanations (SHAP), and an alternative explanation approach, the Contextual Importance and Utility (CIU) method. The produced explanations were assessed by human evaluation. We conducted three user studies based on explanations provided by LIME, SHAP and CIU. Users from different non-medical backgrounds carried out a series of tests in a web-based survey setting and stated their experience and understanding of the given explanations. Three user groups (n = 20, 20, 20) with three distinct forms of explanations were quantitatively analyzed. We found that, as hypothesized, the CIU-explainable method performed better than both LIME and SHAP methods in terms of improving support for human decision-making and being more transparent and thus understandable to users. Additionally, CIU outperformed LIME and SHAP by generating explanations more rapidly. Our findings suggest that there are notable differences in human decision-making between various explanation support settings. In line with that, we present three potential explainable methods that, with future improvements in implementation, can be generalized to different medical data sets and can provide effective decision support to medical experts.

Clinician-driven artificial intelligence in ophthalmology: resources enabling democratization

PURPOSE OF REVIEW

This article aims to discuss the current state of resources enabling the democratization of artificial intelligence (AI) in ophthalmology.

RECENT FINDINGS

Open datasets, efficient labeling techniques, code-free automated machine learning (AutoML) and cloud-based platforms for deployment are resources that enable clinicians with scarce resources to drive their own AI projects.

SUMMARY

Clinicians are the use-case experts who are best suited to drive AI projects tackling patient-relevant outcome measures. Taken together, open datasets, efficient labeling techniques, code-free AutoML and cloud platforms break the barriers for clinician-driven AI. As AI becomes increasingly democratized through such tools, clinicians and patients stand to benefit greatly.

Bio-inspired computation for big data fusion, storage, processing, learning and visualization: state of the art and future directions

This overview gravitates on research achievements that have recently emerged from the confluence between Big Data technologies and bio-inspired computation. A manifold of reasons can be identified for the profitable synergy between these two paradigms, all rooted on the adaptability, intelligence and robustness that biologically inspired principles can provide to technologies aimed to manage, retrieve, fuse and process Big Data efficiently. We delve into this research field by first analyzing in depth the existing literature, with a focus on advances reported in the last few years. This prior literature analysis is complemented by an identification of the new trends and open challenges in Big Data that remain unsolved to date, and that can be effectively addressed by bio-inspired algorithms. As a second contribution, this work elaborates on how bio-inspired algorithms need to be adapted for their use in a Big Data context, in which data fusion becomes crucial as a previous step to allow processing and mining several and potentially heterogeneous data sources. This analysis allows exploring and comparing the scope and efficiency of existing approaches across different problems and domains, with the purpose of identifying new potential applications and research niches. Finally, this survey highlights open issues that remain unsolved to date in this research avenue, alongside a prescription of recommendations for future research.

Analyzing the Noise Robustness of Deep Neural Networks

Adversarial examples, generated by adding small but intentionally imperceptible perturbations to normal examples, can mislead deep neural networks (DNNs) to make incorrect predictions. Although much work has been done on both adversarial attack and defense, a fine-grained understanding of adversarial examples is still lacking. To address this issue, we present a visual analysis method to explain why adversarial examples are misclassified. The key is to compare and analyze the datapaths of both the adversarial and normal examples. A datapath is a group of critical neurons along with their connections. We formulate the datapath extraction as a subset selection problem and solve it by constructing and training a neural network. A multi-level visualization consisting of a network-level visualization of data flows, a layer-level visualization of feature maps, and a neuron-level visualization of learned features, has been designed to help investigate how datapaths of adversarial and normal examples diverge and merge in the prediction process. A quantitative evaluation and a case study were conducted to demonstrate the promise of our method to explain the misclassification of adversarial examples.

A deep learning-based algorithm for detection of cortical arousal during sleep

STUDY OBJECTIVES

The frequency of cortical arousals is an indicator of sleep quality. Additionally, cortical arousals are used to identify hypopneic events. However, it is inconvenient to record electroencephalogram (EEG) data during home sleep testing. Fortunately, most cortical arousal events are associated with autonomic nervous system activity that could be observed on an electrocardiography (ECG) signal. ECG data have lower noise and are easier to record at home than EEG. In this study, we developed a deep learning-based cortical arousal detection algorithm that uses a single-lead ECG to detect arousal during sleep.

METHODS

This study included 1,547 polysomnography records that met study inclusion criteria and were selected from the Multi-Ethnic Study of Atherosclerosis database. We developed an end-to-end deep learning model consisting of convolutional neural networks and recurrent neural networks which: (1) accepted varying length physiological data; (2) directly extracted features from the raw ECG signal; (3) captured long-range dependencies in the physiological data; and (4) produced arousal probability in 1-s resolution.

RESULTS

We evaluated the model on a test set (n = 311). The model achieved a gross area under precision-recall curve score of 0.62 and a gross area under receiver operating characteristic curve score of 0.93.

CONCLUSION

This study demonstrated the end-to-end deep learning approach with a single-lead ECG has the potential to be used to accurately detect arousals in home sleep tests.

Argumentation and explainable artificial intelligence: a survey

Argumentation and eXplainable Artificial Intelligence (XAI) are closely related, as in the recent years, Argumentation has been used for providing Explainability to AI. Argumentation can show step by step how an AI System reaches a decision; it can provide reasoning over uncertainty and can find solutions when conflicting information is faced. In this survey, we elaborate over the topics of Argumentation and XAI combined, by reviewing all the important methods and studies, as well as implementations that use Argumentation to provide Explainability in AI. More specifically, we show how Argumentation can enable Explainability for solving various types of problems in decision-making, justification of an opinion, and dialogues. Subsequently, we elaborate on how Argumentation can help in constructing explainable systems in various applications domains, such as in Medical Informatics, Law, the Semantic Web, Security, Robotics, and some general purpose systems. Finally, we present approaches that combine Machine Learning and Argumentation Theory, toward more interpretable predictive models.

Code-free deep learning for multi-modality medical image classification

A number of large technology companies have created code-free cloud-based platforms that allow researchers and clinicians without coding experience to create deep learning algorithms. In this study, we comprehensively analyse the performance and featureset of six platforms, using four representative cross-sectional and en-face medical imaging datasets to create image classification models. The mean (s.d.) F1 scores across platforms for all model–dataset pairs were as follows: Amazon, 93.9 (5.4); Apple, 72.0 (13.6); Clarifai, 74.2 (7.1); Google, 92.0 (5.4); MedicMind, 90.7 (9.6); Microsoft, 88.6 (5.3). The platforms demonstrated uniformly higher classification performance with the optical coherence tomography modality. Potential use cases given proper validation include research dataset curation, mobile ‘edge models’ for regions without internet access, and baseline models against which to compare and iterate bespoke deep learning approaches.

Improving efficiency of training a virtual treatment planner network via knowledge-guided deep reinforcement learning for intelligent automatic treatment planning of radiotherapy

PURPOSE

We previously proposed an intelligent automatic treatment planning framework for radiotherapy, in which a virtual treatment planner network (VTPN) is built using deep reinforcement learning (DRL) to operate a treatment planning system (TPS) by adjusting treatment planning parameters in it to generate high-quality plans. We demonstrated the potential feasibility of this idea in prostate cancer intensity-modulated radiation therapy (IMRT). Despite the success, the process to train a VTPN via the standard DRL approach with an ϵ-greedy algorithm was time-consuming. The required training time was expected to grow with the complexity of the treatment planning problem, preventing the development of VTPN for more complicated but clinically relevant scenarios. In this study, we proposed a novel knowledge-guided DRL (KgDRL) approach that incorporated knowledge from human planners to guide the training process to improve the efficiency of training a VTPN.

METHOD

Using prostate cancer IMRT as a test bed, we first summarized a number of rules in the actions of adjusting treatment planning parameters of our in-house TPS. During the training process of VTPN, in addition to randomly navigating the large state-action space, as in the standard DRL approach using the ϵ-greedy algorithm, we also sampled actions defined by the rules. The priority of sampling actions from rules decreased over the training process to encourage VTPN to explore new policy on parameter adjustment that were not covered by the rules. To test this idea, we trained a VTPN using KgDRL and compared its performance with another VTPN trained using the standard DRL approach. Both networks were trained using 10 training patient cases and five additional cases for validation, while another 59 cases were employed for the evaluation purpose.

RESULTS

It was found that both VTPNs trained via KgDRL and standard DRL spontaneously learned how to operate the in-house TPS to generate high-quality plans, achieving plan quality scores of 8.82 (±0.29) and 8.43 (±0.48), respectively. Both VTPNs outperformed treatment planning purely based on the rules, which had a plan score of 7.81 (±1.59). VTPN trained with eight episodes using KgDRL was able to perform similar to that trained using DRL with 100 epochs. The training time was reduced from more than a week to ~13 hrs.

CONCLUSION

The proposed KgDRL framework was effective in accelerating the training process of a VTPN by incorporating human knowledge, which will facilitate the development of VTPN for more complicated treatment planning scenarios.

Interactive Visual Study of Multiple Attributes Learning Model of X-Ray Scattering Images

Existing interactive visualization tools for deep learning are mostly applied to the training, debugging, and refinement of neural network models working on natural images. However, visual analytics tools are lacking for the specific application of x-ray image classification with multiple structural attributes. In this paper, we present an interactive system for domain scientists to visually study the multiple attributes learning models applied to x-ray scattering images. It allows domain scientists to interactively explore this important type of scientific images in embedded spaces that are defined on the model prediction output, the actual labels, and the discovered feature space of neural networks. Users are allowed to flexibly select instance images, their clusters, and compare them regarding the specified visual representation of attributes. The exploration is guided by the manifestation of model performance related to mutual relationships among attributes, which often affect the learning accuracy and effectiveness. The system thus supports domain scientists to improve the training dataset and model, find questionable attributes labels, and identify outlier images or spurious data clusters. Case studies and scientists feedback demonstrate its functionalities and usefulness.

Visual Neural Decomposition to Explain Multivariate Data Sets

Investigating relationships between variables in multi-dimensional data sets is a common task for data analysts and engineers. More specifically, it is often valuable to understand which ranges of which input variables lead to particular values of a given target variable. Unfortunately, with an increasing number of independent variables, this process may become cumbersome and time-consuming due to the many possible combinations that have to be explored. In this paper, we propose a novel approach to visualize correlations between input variables and a target output variable that scales to hundreds of variables. We developed a visual model based on neural networks that can be explored in a guided way to help analysts find and understand such correlations. First, we train a neural network to predict the target from the input variables. Then, we visualize the inner workings of the resulting model to help understand relations within the data set. We further introduce a new regularization term for the backpropagation algorithm that encourages the neural network to learn representations that are easier to interpret visually. We apply our method to artificial and real-world data sets to show its utility.

CNNPruner: Pruning Convolutional Neural Networks with Visual Analytics

Convolutional neural networks (CNNs) have demonstrated extraordinarily good performance in many computer vision tasks. The increasing size of CNN models, however, prevents them from being widely deployed to devices with limited computational resources, e.g., mobile/embedded devices. The emerging topic of model pruning strives to address this problem by removing less important neurons and fine-tuning the pruned networks to minimize the accuracy loss. Nevertheless, existing automated pruning solutions often rely on a numerical threshold of the pruning criteria, lacking the flexibility to optimally balance the trade-off between efficiency and accuracy. Moreover, the complicated interplay between the stages of neuron pruning and model fine-tuning makes this process opaque, and therefore becomes difficult to optimize. In this paper, we address these challenges through a visual analytics approach, named CNNPruner. It considers the importance of convolutional filters through both instability and sensitivity, and allows users to interactively create pruning plans according to a desired goal on model size or accuracy. Also, CNNPruner integrates state-of-the-art filter visualization techniques to help users understand the roles that different filters played and refine their pruning plans. Through comprehensive case studies on CNNs with real-world sizes, we validate the effectiveness of CNNPruner.

VATLD: A Visual Analytics System to Assess, Understand and Improve Traffic Light Detection

Traffic light detection is crucial for environment perception and decision-making in autonomous driving. State-of-the-art detectors are built upon deep Convolutional Neural Networks (CNNs) and have exhibited promising performance. However, one looming concern with CNN based detectors is how to thoroughly evaluate the performance of accuracy and robustness before they can be deployed to autonomous vehicles. In this work, we propose a visual analytics system, VATLD, equipped with a disentangled representation learning and semantic adversarial learning, to assess, understand, and improve the accuracy and robustness of traffic light detectors in autonomous driving applications. The disentangled representation learning extracts data semantics to augment human cognition with human-friendly visual summarization, and the semantic adversarial learning efficiently exposes interpretable robustness risks and enables minimal human interaction for actionable insights. We also demonstrate the effectiveness of various performance improvement strategies derived from actionable insights with our visual analytics system, VATLD, and illustrate some practical implications for safety-critical applications in autonomous driving.

In Search of Trustworthy and Transparent Intelligent Systems With Human-Like Cognitive and Reasoning Capabilities

At present we are witnessing a tremendous interest in Artificial Intelligence (AI), particularly in Deep Learning (DL)/Deep Neural Networks (DNNs). One of the reasons appears to be the unmatched performance achieved by such systems. This has resulted in an enormous hope on such techniques and often these are viewed as all—cure solutions. But most of these systems cannot explain why a particular decision is made (black box) and sometimes miserably fail in cases where other systems would not. Consequently, in critical applications such as healthcare and defense practitioners do not like to trust such systems. Although an AI system is often designed taking inspiration from the brain, there is not much attempt to exploit cues from the brain in true sense. In our opinion, to realize intelligent systems with human like reasoning ability, we need to exploit knowledge from the brain science. Here we discuss a few findings in brain science that may help designing intelligent systems. We explain the relevance of transparency, explainability, learning from a few examples, and the trustworthiness of an AI system. We also discuss a few ways that may help to achieve these attributes in a learning system.

A Comprehensive Survey on Machine Learning-Based Big Data Analytics for IoT-Enabled Smart Healthcare System

The outbreak of chronic diseases such as COVID-19 has made a renewed call for providing urgent healthcare facilities to the citizens across the globe. The recent pandemic exposes the shortcomings of traditional healthcare system, i.e., hospitals and clinics alone are not capable to cope with this situation. One of the major technology that aids contemporary healthcare solutions is the smart and connected wearables. The advancement in Internet of Things (IoT) has enabled these wearables to collect data on an unprecedented scale. These wearables gather context-oriented information related to our physical, behavioural and psychological health. The big data generated by wearables and other healthcare devices of IoT is a challenging task to manage that can negatively affect the inference process at the decision centres. Applying big data analytics for mining information, extracting knowledge and making predictions/inferences has recently attracted significant attention. Machine learning is another area of research that has successfully been applied to solve various networking problems such as routing, traffic engineering, resource allocation, and security. Recently, we have seen a surge in the application of ML-based techniques for the improvement of various IoT applications. Although, big data analytics and machine learning are extensively researched, there is a lack of study that exclusively focus on the evolution of ML-based techniques for big data analysis in the IoT healthcare sector. In this paper, we have presented a comprehensive review on the application of machine learning techniques for big data analysis in the healthcare sector. Furthermore, strength and weaknesses of existing techniques along with various research challenges are highlighted. Our study will provide an insight for healthcare practitioners and government agencies to keep themselves well-equipped with the latest trends in ML-based big data analytics for smart healthcare.

Artificial intelligence for the management of pancreatic diseases

Novel artificial intelligence techniques are emerging in all fields of healthcare, including gastroenterology. The aim of this review is to give an overview of artificial intelligence applications in the management of pancreatic diseases. We performed a systematic literature search in PubMed and Medline up to May 2020 to identify relevant articles. Our results showed that the development of machine-learning based applications is rapidly evolving in the management of pancreatic diseases, guiding precision medicine in clinical, endoscopic and radiologic settings. Before implementation into clinical practice, further research should focus on the external validation of novel techniques, clarifying the accuracy and robustness of these models.

Integrating Machine Learning with Human Knowledge

Machine learning has been heavily researched and widely used in many disciplines. However, achieving high accuracy requires a large amount of data that is sometimes difficult, expensive, or impractical to obtain. Integrating human knowledge into machine learning can significantly reduce data requirement, increase reliability and robustness of machine learning, and build explainable machine learning systems. This allows leveraging the vast amount of human knowledge and capability of machine learning to achieve functions and performance not available before and will facilitate the interaction between human beings and machine learning systems, making machine learning decisions understandable to humans. This paper gives an overview of the knowledge and its representations that can be integrated into machine learning and the methodology. We cover the fundamentals, current status, and recent progress of the methods, with a focus on popular and new topics. The perspectives on future directions are also discussed.

Explainable Artificial Intelligence for Developing Smart Cities Solutions

Traditional Artificial Intelligence (AI) technologies used in developing smart cities solutions, Machine Learning (ML) and recently Deep Learning (DL), rely more on utilising best representative training datasets and features engineering and less on the available domain expertise. We argue that such an approach to solution development makes the outcome of solutions less explainable, i.e., it is often not possible to explain the results of the model. There is a growing concern among policymakers in cities with this lack of explainability of AI solutions, and this is considered a major hindrance in the wider acceptability and trust in such AI-based solutions. In this work, we survey the concept of ‘explainable deep learning’ as a subset of the ‘explainable AI’ problem and propose a new solution using Semantic Web technologies, demonstrated with a smart cities flood monitoring application in the context of a European Commission-funded project. Monitoring of gullies and drainage in crucial geographical areas susceptible to flooding issues is an important aspect of any flood monitoring solution. Typical solutions for this problem involve the use of cameras to capture images showing the affected areas in real-time with different objects such as leaves, plastic bottles etc., and building a DL-based classifier to detect such objects and classify blockages based on the presence and coverage of these objects in the images. In this work, we uniquely propose an Explainable AI solution using DL and Semantic Web technologies to build a hybrid classifier. In this hybrid classifier, the DL component detects object presence and coverage level and semantic rules designed with close consultation with experts carry out the classification. By using the expert knowledge in the flooding context, our hybrid classifier provides the flexibility on categorising the image using objects and their coverage relationships. The experimental results demonstrated with a real-world use case showed that this hybrid approach of image classification has on average 11% improvement (F-Measure) in image classification performance compared to DL-only classifier. It also has the distinct advantage of integrating experts’ knowledge on defining the decision-making rules to represent the complex circumstances and using such knowledge to explain the results.

A bird's-eye view of deep learning in bioimage analysis

Deep learning of artificial neural networks has become the de facto standard approach to solving data analysis problems in virtually all fields of science and engineering. Also in biology and medicine, deep learning technologies are fundamentally transforming how we acquire, process, analyze, and interpret data, with potentially far-reaching consequences for healthcare. In this mini-review, we take a bird's-eye view at the past, present, and future developments of deep learning, starting from science at large, to biomedical imaging, and bioimage analysis in particular.

An Interactive Visualization for Feature Localization in Deep Neural Networks

Deep artificial neural networks have become the go-to method for many machine learning tasks. In the field of computer vision, deep convolutional neural networks achieve state-of-the-art performance for tasks such as classification, object detection, or instance segmentation. As deep neural networks become more and more complex, their inner workings become more and more opaque, rendering them a "black box" whose decision making process is no longer comprehensible. In recent years, various methods have been presented that attempt to peek inside the black box and to visualize the inner workings of deep neural networks, with a focus on deep convolutional neural networks for computer vision. These methods can serve as a toolbox to facilitate the design and inspection of neural networks for computer vision and the interpretation of the decision making process of the network. Here, we present the new tool Interactive Feature Localization in Deep neural networks (IFeaLiD) which provides a novel visualization approach to convolutional neural network layers. The tool interprets neural network layers as multivariate feature maps and visualizes the similarity between the feature vectors of individual pixels of an input image in a heat map display. The similarity display can reveal how the input image is perceived by different layers of the network and how the perception of one particular image region compares to the perception of the remaining image. IFeaLiD runs interactively in a web browser and can process even high resolution feature maps in real time by using GPU acceleration with WebGL 2. We present examples from four computer vision datasets with feature maps from different layers of a pre-trained ResNet101. IFeaLiD is open source and available online at https://ifealid.cebitec.uni-bielefeld.de.

Discovery and validation of biomarkers to aid the development of safe and effective pain therapeutics: challenges and opportunities

Pain medication plays an important role in the treatment of acute and chronic pain conditions, but some drugs, opioids in particular, have been overprescribed or prescribed without adequate safeguards, leading to an alarming rise in medication-related overdose deaths. The NIH Helping to End Addiction Long-term (HEAL) Initiative is a trans-agency effort to provide scientific solutions to stem the opioid crisis. One component of the initiative is to support biomarker discovery and rigorous validation in collaboration with industry leaders to accelerate high-quality clinical research into neurotherapeutics and pain. The use of objective biomarkers and clinical trial end points throughout the drug discovery and development process is crucial to help define pathophysiological subsets of pain, evaluate target engagement of new drugs and predict the analgesic efficacy of new drugs. In 2018, the NIH-led Discovery and Validation of Biomarkers to Develop Non-Addictive Therapeutics for Pain workshop convened scientific leaders from academia, industry, government and patient advocacy groups to discuss progress, challenges, gaps and ideas to facilitate the development of biomarkers and end points for pain. The outcomes of this workshop are outlined in this Consensus Statement.

Clinical Implementation of Deep Learning in Thoracic Radiology: Potential Applications and Challenges

Chest X-ray radiography and computed tomography, the two mainstay modalities in thoracic radiology, are under active investigation with deep learning technology, which has shown promising performance in various tasks, including detection, classification, segmentation, and image synthesis, outperforming conventional methods and suggesting its potential for clinical implementation. However, the implementation of deep learning in daily clinical practice is in its infancy and facing several challenges, such as its limited ability to explain the output results, uncertain benefits regarding patient outcomes, and incomplete integration in daily workflow. In this review article, we will introduce the potential clinical applications of deep learning technology in thoracic radiology and discuss several challenges for its implementation in daily clinical practice.

Explainable Artificial Intelligence for Neuroscience: Behavioral Neurostimulation

The use of Artificial Intelligence and machine learning in basic research and clinical neuroscience is increasing. AI methods enable the interpretation of large multimodal datasets that can provide unbiased insights into the fundamental principles of brain function, potentially paving the way for earlier and more accurate detection of brain disorders and better informed intervention protocols. Despite AI's ability to create accurate predictions and classifications, in most cases it lacks the ability to provide a mechanistic understanding of how inputs and outputs relate to each other. Explainable Artificial Intelligence (XAI) is a new set of techniques that attempts to provide such an understanding, here we report on some of these practical approaches. We discuss the potential value of XAI to the field of neurostimulation for both basic scientific inquiry and therapeutic purposes, as well as, outstanding questions and obstacles to the success of the XAI approach.