An Analysis of Machine- and Human-Analytics in Classification

The Switch, the Ladder, and the Matrix: Models for Classifying AI Systems

Organisations that design and deploy artificial intelligence (AI) systems increasingly commit themselves to high-level, ethical principles. However, there still exists a gap between principles and practices in AI ethics. One major obstacle organisations face when attempting to operationalise AI Ethics is the lack of a well-defined material scope. Put differently, the question to which systems and processes AI ethics principles ought to apply remains unanswered. Of course, there exists no universally accepted definition of AI, and different systems pose different ethical challenges. Nevertheless, pragmatic problem-solving demands that things should be sorted so that their grouping will promote successful actions for some specific end. In this article, we review and compare previous attempts to classify AI systems for the purpose of implementing AI governance in practice. We find that attempts to classify AI systems proposed in previous literature use one of three mental models: the Switch, i.e., a binary approach according to which systems either are or are not considered AI systems depending on their characteristics; the Ladder, i.e., a risk-based approach that classifies systems according to the ethical risks they pose; and the Matrix, i.e., a multi-dimensional classification of systems that take various aspects into account, such as context, input data, and decision-model. Each of these models for classifying AI systems comes with its own set of strengths and weaknesses. By conceptualising different ways of classifying AI systems into simple mental models, we hope to provide organisations that design, deploy, or regulate AI systems with the vocabulary needed to demarcate the material scope of their AI governance frameworks.

Visualizing Ensemble Predictions of Music Mood

Music mood classification has been a challenging problem in comparison with other music classification problems (e.g., genre, composer, or period). One solution for addressing this challenge is to use an ensemble of machine learning models. In this paper, we show that visualization techniques can effectively convey the popular prediction as well as uncertainty at different music sections along the temporal axis while enabling the analysis of individual ML models in conjunction with their application to different musical data. In addition to the traditional visual designs, such as stacked line graph, ThemeRiver, and pixel-based visualization, we introduce a new variant of ThemeRiver, called "dual-flux ThemeRiver", which allows viewers to observe and measure the most popular prediction more easily than stacked line graph and ThemeRiver. Together with pixel-based visualization, dual-flux ThemeRiver plots can also assist in model-development workflows, in addition to annotating music using ensemble model predictions.

Visual Exploration of Machine Learning Model Behavior with Hierarchical Surrogate Rule Sets

One of the potential solutions for model interpretation is to train a surrogate model: a more transparent model that approximates the behavior of the model to be explained. Typically, classification rules or decision trees are used due to their logic-based expressions. However, decision trees can grow too deep, and rule sets can become too large to approximate a complex model. Unlike paths on a decision tree that must share ancestor nodes (conditions), rules are more flexible. However, the unstructured visual representation of rules makes it hard to make inferences across rules. In this paper, we focus on tabular data and present novel algorithmic and interactive solutions to address these issues. First, we present Hierarchical Surrogate Rules (HSR), an algorithm that generates hierarchical rules based on user-defined parameters. We also contribute SuRE, a visual analytics (VA) system that integrates HSR and an interactive surrogate rule visualization, the Feature-Aligned Tree, which depicts rules as trees while aligning features for easier comparison. We evaluate the algorithm in terms of parameter sensitivity, time performance, and comparison with surrogate decision trees and find that it scales reasonably well and overcomes the shortcomings of surrogate decision trees. We evaluate the visualization and the system through a usability study and an observational study with domain experts. Our investigation shows that the participants can use feature-aligned trees to perform non-trivial tasks with very high accuracy. We also discuss many interesting findings, including a rule analysis task characterization, that can be used for visualization design and future research.

Constructing Explainable Classifiers from the Start—Enabling Human-in-the Loop Machine Learning

Interactive machine learning (IML) enables the incorporation of human expertise because the human participates in the construction of the learned model. Moreover, with human-in-the-loop machine learning (HITL-ML), the human experts drive the learning, and they can steer the learning objective not only for accuracy but perhaps for characterisation and discrimination rules, where separating one class from others is the primary objective. Moreover, this interaction enables humans to explore and gain insights into the dataset as well as validate the learned models. Validation requires transparency and interpretable classifiers. The huge relevance of understandable classification has been recently emphasised for many applications under the banner of explainable artificial intelligence (XAI). We use parallel coordinates to deploy an IML system that enables the visualisation of decision tree classifiers but also the generation of interpretable splits beyond parallel axis splits. Moreover, we show that characterisation and discrimination rules are also well communicated using parallel coordinates. In particular, we report results from the largest usability study of a IML system, confirming the merits of our approach.

Task-Based Visual Interactive Modeling: Decision Trees and Rule-Based Classifiers

Visual analytics enables the coupling of machine learning models and humans in a tightly integrated workflow, addressing various analysis tasks. Each task poses distinct demands to analysts and decision-makers. In this survey, we focus on one canonical technique for rule-based classification, namely decision tree classifiers. We provide an overview of available visualizations for decision trees with a focus on how visualizations differ with respect to 16 tasks. Further, we investigate the types of visual designs employed, and the quality measures presented. We find that (i) interactive visual analytics systems for classifier development offer a variety of visual designs, (ii) utilization tasks are sparsely covered, (iii) beyond classifier development, node-link diagrams are omnipresent, (iv) even systems designed for machine learning experts rarely feature visual representations of quality measures other than accuracy. In conclusion, we see a potential for integrating algorithmic techniques, mathematical quality measures, and tailored interactive visualizations to enable human experts to utilize their knowledge more effectively.

A Bounded Measure for Estimating the Benefit of Visualization (Part II): Case Studies and Empirical Evaluation

Many visual representations, such as volume-rendered images and metro maps, feature a noticeable amount of information loss due to a variety of many-to-one mappings. At a glance, there seem to be numerous opportunities for viewers to misinterpret the data being visualized, hence, undermining the benefits of these visual representations. In practice, there is little doubt that these visual representations are useful. The recently-proposed information-theoretic measure for analyzing the cost–benefit ratio of visualization processes can explain such usefulness experienced in practice and postulate that the viewers’ knowledge can reduce the potential distortion (e.g., misinterpretation) due to information loss. This suggests that viewers’ knowledge can be estimated by comparing the potential distortion without any knowledge and the actual distortion with some knowledge. However, the existing cost–benefit measure consists of an unbounded divergence term, making the numerical measurements difficult to interpret. This is the second part of a two-part paper, which aims to improve the existing cost–benefit measure. Part I of the paper provided a theoretical discourse about the problem of unboundedness, reported a conceptual analysis of nine candidate divergence measures for resolving the problem, and eliminated three from further consideration. In this Part II, we describe two groups of case studies for evaluating the remaining six candidate measures empirically. In particular, we obtained instance data for (i) supporting the evaluation of the remaining candidate measures and (ii) demonstrating their applicability in practical scenarios for estimating the cost–benefit of visualization processes as well as the impact of human knowledge in the processes. The real world data about visualization provides practical evidence for evaluating the usability and intuitiveness of the candidate measures. The combination of the conceptual analysis in Part I and the empirical evaluation in this part allows us to select the most appropriate bounded divergence measure for improving the existing cost–benefit measure.

A Bounded Measure for Estimating the Benefit of Visualization (Part I): Theoretical Discourse and Conceptual Evaluation

Information theory can be used to analyze the cost–benefit of visualization processes. However, the current measure of benefit contains an unbounded term that is neither easy to estimate nor intuitive to interpret. In this work, we propose to revise the existing cost–benefit measure by replacing the unbounded term with a bounded one. We examine a number of bounded measures that include the Jenson–Shannon divergence, its square root, and a new divergence measure formulated as part of this work. We describe the rationale for proposing a new divergence measure. In the first part of this paper, we focus on the conceptual analysis of the mathematical properties of these candidate measures. We use visualization to support the multi-criteria comparison, narrowing the search down to several options with better mathematical properties. The theoretical discourse and conceptual evaluation in this part provides the basis for further data-driven evaluation based on synthetic and experimental case studies that are reported in the second part of this paper.

A Taxonomy of Property Measures to Unify Active Learning and Human-centered Approaches to Data Labeling

Strategies for selecting the next data instance to label, in service of generating labeled data for machine learning, have been considered separately in the machine learning literature on active learning and in the visual analytics literature on human-centered approaches. We propose a unified design space for instance selection strategies to support detailed and fine-grained analysis covering both of these perspectives. We identify a concise set of 15 properties, namely measureable characteristics of datasets or of machine learning models applied to them, that cover most of the strategies in these literatures. To quantify these properties, we introduce Property Measures (PM) as fine-grained building blocks that can be used to formalize instance selection strategies. In addition, we present a taxonomy of PMs to support the description, evaluation, and generation of PMs across four dimensions: machine learning (ML) Model Output , Instance Relations , Measure Functionality , and Measure Valence . We also create computational infrastructure to support qualitative visual data analysis: a visual analytics explainer for PMs built around an implementation of PMs using cascades of eight atomic functions. It supports eight analysis tasks, covering the analysis of datasets and ML models using visual comparison within and between PMs and groups of PMs, and over time during the interactive labeling process. We iteratively refined the PM taxonomy, the explainer, and the task abstraction in parallel with each other during a two-year formative process, and show evidence of their utility through a summative evaluation with the same infrastructure. This research builds a formal baseline for the better understanding of the commonalities and differences of instance selection strategies, which can serve as the stepping stone for the synthesis of novel strategies in future work.

Why Visualize? Untangling a Large Network of Arguments

Visualization has been deemed a useful technique by researchers and practitioners, alike, leaving a trail of arguments behind that reason why visualization works. In addition, examples of misleading usages of visualizations in information communication have occasionally been pointed out. Thus, to contribute to the fundamental understanding of our discipline, we require a comprehensive collection of arguments on "why visualize?" (or "why not?"), untangling the rationale behind positive and negative viewpoints. In this paper, we report a theoretical study to understand the underlying reasons of various arguments; their relationships (e.g., built-on, and conflict); and their respective dependencies on tasks, users, and data. We curated an argumentative network based on a collection of arguments from various fields, including information visualization, cognitive science, psychology, statistics, philosophy, and others. Our work proposes several categorizations for the arguments, and makes their relations explicit. We contribute the first comprehensive and systematic theoretical study of the arguments on visualization. Thereby, we provide a roadmap towards building a foundation for visualization theory and empirical research as well as for practical application in the critique and design of visualizations. In addition, we provide our argumentation network and argument collection online at https://whyvis.dbvis.de, supported by an interactive visualization.

HyperTendril: Visual Analytics for User-Driven Hyperparameter Optimization of Deep Neural Networks

To mitigate the pain of manually tuning hyperparameters of deep neural networks, automated machine learning (AutoML) methods have been developed to search for an optimal set of hyperparameters in large combinatorial search spaces. However, the search results of AutoML methods significantly depend on initial configurations, making it a non-trivial task to find a proper configuration. Therefore, human intervention via a visual analytic approach bears huge potential in this task. In response, we propose HyperTendril, a web-based visual analytics system that supports user-driven hyperparameter tuning processes in a model-agnostic environment. HyperTendril takes a novel approach to effectively steering hyperparameter optimization through an iterative, interactive tuning procedure that allows users to refine the search spaces and the configuration of the AutoML method based on their own insights from given results. Using HyperTendril, users can obtain insights into the complex behaviors of various hyperparameter search algorithms and diagnose their configurations. In addition, HyperTendril supports variable importance analysis to help the users refine their search spaces based on the analysis of relative importance of different hyperparameters and their interaction effects. We present the evaluation demonstrating how HyperTendril helps users steer their tuning processes via a longitudinal user study based on the analysis of interaction logs and in-depth interviews while we deploy our system in a professional industrial environment.

HypoML: Visual Analysis for Hypothesis-based Evaluation of Machine Learning Models

In this paper, we present a visual analytics tool for enabling hypothesis-based evaluation of machine learning (ML) models. We describe a novel ML-testing framework that combines the traditional statistical hypothesis testing (commonly used in empirical research) with logical reasoning about the conclusions of multiple hypotheses. The framework defines a controlled configuration for testing a number of hypotheses as to whether and how some extra information about a "concept" or "feature" may benefit or hinder an ML model. Because reasoning multiple hypotheses is not always straightforward, we provide HypoML as a visual analysis tool, with which, the multi-thread testing results are first transformed to analytical results using statistical and logical inferences, and then to a visual representation for rapid observation of the conclusions and the logical flow between the testing results and hypotheses. We have applied HypoML to a number of hypothesized concepts, demonstrating the intuitive and explainable nature of the visual analysis.

Bridging Text Visualization and Mining: A Task-Driven Survey

Visual text analytics has recently emerged as one of the most prominent topics in both academic research and the commercial world. To provide an overview of the relevant techniques and analysis tasks, as well as the relationships between them, we comprehensively analyzed 263 visualization papers and 4,346 mining papers published between 1992-2017 in two fields: visualization and text mining. From the analysis, we derived around 300 concepts (visualization techniques, mining techniques, and analysis tasks) and built a taxonomy for each type of concept. The co-occurrence relationships between the concepts were also extracted. Our research can be used as a stepping-stone for other researchers to 1) understand a common set of concepts used in this research topic; 2) facilitate the exploration of the relationships between visualization techniques, mining techniques, and analysis tasks; 3) understand the current practice in developing visual text analytics tools; 4) seek potential research opportunities by narrowing the gulf between visualization and mining techniques based on the analysis tasks; and 5) analyze other interdisciplinary research areas in a similar way. We have also contributed a web-based visualization tool for analyzing and understanding research trends and opportunities in visual text analytics.

Manifold: A Model-Agnostic Framework for Interpretation and Diagnosis of Machine Learning Models

Interpretation and diagnosis of machine learning models have gained renewed interest in recent years with breakthroughs in new approaches. We present Manifold, a framework that utilizes visual analysis techniques to support interpretation, debugging, and comparison of machine learning models in a more transparent and interactive manner. Conventional techniques usually focus on visualizing the internal logic of a specific model type (i.e., deep neural networks), lacking the ability to extend to a more complex scenario where different model types are integrated. To this end, Manifold is designed as a generic framework that does not rely on or access the internal logic of the model and solely observes the input (i.e., instances or features) and the output (i.e., the predicted result and probability distribution). We describe the workflow of Manifold as an iterative process consisting of three major phases that are commonly involved in the model development and diagnosis process: inspection (hypothesis), explanation (reasoning), and refinement (verification). The visual components supporting these tasks include a scatterplot-based visual summary that overviews the models' outcome and a customizable tabular view that reveals feature discrimination. We demonstrate current applications of the framework on the classification and regression tasks and discuss other potential machine learning use scenarios where Manifold can be applied.

An Interactive Method to Improve Crowdsourced Annotations

In order to effectively infer correct labels from noisy crowdsourced annotations, learning-from-crowds models have introduced expert validation. However, little research has been done on facilitating the validation procedure. In this paper, we propose an interactive method to assist experts in verifying uncertain instance labels and unreliable workers. Given the instance labels and worker reliability inferred from a learning-from-crowds model, candidate instances and workers are selected for expert validation. The influence of verified results is propagated to relevant instances and workers through the learning-from-crowds model. To facilitate the validation of annotations, we have developed a confusion visualization to indicate the confusing classes for further exploration, a constrained projection method to show the uncertain labels in context, and a scatter-plot-based visualization to illustrate worker reliability. The three visualizations are tightly integrated with the learning-from-crowds model to provide an iterative and progressive environment for data validation. Two case studies were conducted that demonstrate our approach offers an efficient method for validating and improving crowdsourced annotations.

An Information-Theoretic Approach to the Cost-benefit Analysis of Visualization in Virtual Environments

Visualization and virtual environments (VEs) have been two interconnected parallel strands in visual computing for decades. Some VEs have been purposely developed for visualization applications, while many visualization applications are exemplary showcases in general-purpose VEs. Because of the development and operation costs of VEs, the majority of visualization applications in practice have yet to benefit from the capacity of VEs. In this paper, we examine this status quo from an information-theoretic perspective. Our objectives are to conduct cost-benefit analysis on typical VE systems (including augmented and mixed reality, theater-based systems, and large powerwalls), to explain why some visualization applications benefit more from VEs than others, and to sketch out pathways for the future development of visualization applications in VEs. We support our theoretical propositions and analysis using theories and discoveries in the literature of cognitive sciences and the practical evidence reported in the literatures of visualization and VEs.

VIS4ML: An Ontology for Visual Analytics Assisted Machine Learning

While many VA workflows make use of machine-learned models to support analytical tasks, VA workflows have become increasingly important in understanding and improving Machine Learning (ML) processes. In this paper, we propose an ontology (VIS4ML) for a subarea of VA, namely "VA-assisted ML". The purpose of VIS4ML is to describe and understand existing VA workflows used in ML as well as to detect gaps in ML processes and the potential of introducing advanced VA techniques to such processes. Ontologies have been widely used to map out the scope of a topic in biology, medicine, and many other disciplines. We adopt the scholarly methodologies for constructing VIS4ML, including the specification, conceptualization, formalization, implementation, and validation of ontologies. In particular, we reinterpret the traditional VA pipeline to encompass model-development workflows. We introduce necessary definitions, rules, syntaxes, and visual notations for formulating VIS4ML and make use of semantic web technologies for implementing it in the Web Ontology Language (OWL). VIS4ML captures the high-level knowledge about previous workflows where VA is used to assist in ML. It is consistent with the established VA concepts and will continue to evolve along with the future developments in VA and ML. While this ontology is an effort for building the theoretical foundation of VA, it can be used by practitioners in real-world applications to optimize model-development workflows by systematically examining the potential benefits that can be brought about by either machine or human capabilities. Meanwhile, VIS4ML is intended to be extensible and will continue to be updated to reflect future advancements in using VA for building high-quality data-analytical models or for building such models rapidly.

Visual Analytics in Deep Learning: An Interrogative Survey for the Next Frontiers

Deep learning has recently seen rapid development and received significant attention due to its state-of-the-art performance on previously-thought hard problems. However, because of the internal complexity and nonlinear structure of deep neural networks, the underlying decision making processes for why these models are achieving such performance are challenging and sometimes mystifying to interpret. As deep learning spreads across domains, it is of paramount importance that we equip users of deep learning with tools for understanding when a model works correctly, when it fails, and ultimately how to improve its performance. Standardized toolkits for building neural networks have helped democratize deep learning; visual analytics systems have now been developed to support model explanation, interpretation, debugging, and improvement. We present a survey of the role of visual analytics in deep learning research, which highlights its short yet impactful history and thoroughly summarizes the state-of-the-art using a human-centered interrogative framework, focusing on the Five W's and How (Why, Who, What, How, When, and Where). We conclude by highlighting research directions and open research problems. This survey helps researchers and practitioners in both visual analytics and deep learning to quickly learn key aspects of this young and rapidly growing body of research, whose impact spans a diverse range of domains.

Analyzing the Training Processes of Deep Generative Models

Among the many types of deep models, deep generative models (DGMs) provide a solution to the important problem of unsupervised and semi-supervised learning. However, training DGMs requires more skill, experience, and know-how because their training is more complex than other types of deep models such as convolutional neural networks (CNNs). We develop a visual analytics approach for better understanding and diagnosing the training process of a DGM. To help experts understand the overall training process, we first extract a large amount of time series data that represents training dynamics (e.g., activation changes over time). A blue-noise polyline sampling scheme is then introduced to select time series samples, which can both preserve outliers and reduce visual clutter. To further investigate the root cause of a failed training process, we propose a credit assignment algorithm that indicates how other neurons contribute to the output of the neuron causing the training failure. Two case studies are conducted with machine learning experts to demonstrate how our approach helps understand and diagnose the training processes of DGMs. We also show how our approach can be directly used to analyze other types of deep models, such as CNNs.

Voila: Visual Anomaly Detection and Monitoring with Streaming Spatiotemporal Data

The increasing availability of spatiotemporal data continuously collected from various sources provides new opportunities for a timely understanding of the data in their spatial and temporal context. Finding abnormal patterns in such data poses significant challenges. Given that there is often no clear boundary between normal and abnormal patterns, existing solutions are limited in their capacity of identifying anomalies in large, dynamic and heterogeneous data, interpreting anomalies in their multifaceted, spatiotemporal context, and allowing users to provide feedback in the analysis loop. In this work, we introduce a unified visual interactive system and framework, Voila, for interactively detecting anomalies in spatiotemporal data collected from a streaming data source. The system is designed to meet two requirements in real-world applications, i.e., online monitoring and interactivity. We propose a novel tensor-based anomaly analysis algorithm with visualization and interaction design that dynamically produces contextualized, interpretable data summaries and allows for interactively ranking anomalous patterns based on user input. Using the "smart city" as an example scenario, we demonstrate the effectiveness of the proposed framework through quantitative evaluation and qualitative case studies.