The currently dominating artificial intelligence and machine learning technology, neural networks, builds on inductive statistical learning processes. Being void of knowledge that can be used deductively these systems cannot distinguish exemplars part of the target domain from those not part of it. This ability is critical when the aim is to build human trust in real-world settings and essential to avoid usage in domains wherein a system cannot be trusted. In the work presented here, we conduct two qualitative contextual user studies and one controlled experiment to uncover research paths and design openings for the sought distinction. Through our experiments, we find a need to refocus from average case metrics and benchmarking datasets toward systems that can be falsified. The work uncovers and lays bare the need to incorporate and internalise a domain ontology in the systems and/or present evidence for a decision in a fashion that allows a human to use our unique knowledge and reasoning capability. Additional material and code to reproduce our experiments can be found at https://github.com/k3larra/ood.

Artificial intelligence and machine learning (ML) in particular increasingly impact human life by creating value from collected data. This assetisation affects all aspectsof human life, from choosing a significant other to recommending a product for us to consume. This type of ML-based system thrives because it predicts human behaviour based on average case performance metrics (like accuracy). However, its usefulnessis more limited when it comes to being transparent about its internal knowledge representations for singular decisions, for example, it is not good at explaining why ithas suggested a particular decision in a specific context.The goal of this work is to let end users be in command of how ML systems are used and thereby combine the strengths of humans and machines – machines which can propose transparent decisions. Artificial neural networks are an interesting candidate for a setting of this type, given that this technology has been successful in building knowledge representations from raw data. A neural network can be trained by exposing it to data from the target domain. It can then internalise knowledge representations from the domain and perform contextual tasks. In these situations, the fragment of the actual world internalised in an ML system has to be contextualised by a human to beuseful and trustworthy in non-static settings.This setting is explored through the overarching research question: What challenges and opportunities can emerge when an end user uses neural networks in context to support singular decision-making? To address this question, Research through Design is used as the central methodology, as this research approach matches the openness of the research question. Through six design experiments, I explore and expand on challenges and opportunities in settings where singular contextual decisions matter. The initial design experiments focus on opportunities in settings that augment human cognitive abilities. Thereafter, the experiments explore challenges related to settings where neural networks can enhance human cognitive abilities. This part concerns approaches intended to explain promoted decisions.This work contributes in three ways: 1) exploring learning related to neural networks in context to put forward a core terminology for contextual decision-making using ML systems, wherein the terminology includes the generative notions of true-to-the-domain, concept, out-of-distribution and generalisation; 2) presenting a number of design guidelines; and 3) showing the need to align internal knowledge representations with concepts if neural networks are to produce explainable decisions. I also argue that training neural networks to generalise basic concepts like shapes and colours, concepts easily understandable by humans, is a path forward. This research direction leads towards neural network-based systems that can produce more complex explanations that build on basic generalisable concepts.

Concepts are powerful human mental representations used to explain, reason and understand. In this work, we use theories on concepts as an analytical lens to compare internal knowledge representations in neural networks to human concepts. In two image classification studies we find an unclear alignment between these, but more pronounced, we find the need to further develop explanation methods that incorporate concept ontologies. 

Machine Learning (ML) and Artificial Intelligence (AI) impact many aspects of human life, from recommending a significant other to assist the search for extraterrestrial life. The area develops rapidly and exiting unexplored design spaces are constantly laid bare. The focus in this work is one of these areas; ML systems where decisions concerning ML model training, usage and selection of target domain lay in the hands of domain experts. 

This work is then on ML systems that function as a tool that augments and/or enhance human capabilities. The approach presented is denoted Human In Command ML (HIC-ML) systems. To enquire into this research domain design experiments of varying fidelity were used. Two of these experiments focus on augmenting human capabilities and targets the domains commuting and sorting batteries. One experiment focuses on enhancing human capabilities by identifying similar hand-painted plates. The experiments are used as illustrative examples to explore settings where domain experts potentially can: independently train an ML model and in an iterative fashion, interact with it and interpret and understand its decisions. 

HIC-ML should be seen as a governance principle that focuses on adding value and meaning to users. In this work, concrete application areas are presented and discussed. To open up for designing ML-based products for the area an abstract model for HIC-ML is constructed and design guidelines are proposed. In addition, terminology and abstractions useful when designing for explicability are presented by imposing structure and rigidity derived from scientific explanations. Together, this opens up for a contextual shift in ML and makes new application areas probable, areas that naturally couples the usage of AI technology to human virtues and potentially, as a consequence, can result in a democratisation of the usage and knowledge concerning this powerful technology.

In this paper, we study explanations in a setting where human capabilities are in parity with Machine Learning (ML) capabilities. If an ML system is to be trusted in this situation, limitations in the trained ML model’s abilities have to be exposed to the end-user. A majority of current approaches focus on the task of creating explanations for a proposed decision, but less attention is given to the equally important task of exposing limitations in the ML model’s capabilities, limitations that in turn affect the validity of created explanations. Using a small-scale design experiment we compare human explanations with explanations created by an ML system. This paper explores and presents how the structure and terminology of scientific explanations can expose limitations in the ML models knowledge and be used as an approach for research and design in the area of explainable artificial intelligence.

Contemporary Machine Learning (ML) often focuseson large existing and labeled datasets and metrics aroundaccuracy and performance. In pervasive online systems, conditionschange constantly and there is a need for systems thatcan adapt. In Machine Teaching (MT) a human domain expertis responsible for the knowledge transfer and can thus addressthis. In my work, I focus on domain experts and the importanceof, for the ML system, available features and the space they span.This space confines the, to the ML systems, observable fragmentof the physical world. My investigation of the feature space isgrounded in a conducted study and related theories. The resultof this work is applicable when designing systems where domainexperts have a key role as teachers.

Machine learning research today is dominated by atechnocentric perspective and in many cases disconnected fromthe users of the technology. The machine teaching paradigm insteadshifts the focus from machine learning experts towards thedomain experts and users of machine learning technology. Thisshift opens up for new perspectives on the current use of machinelearning as well as new usage areas to explore. In this study,we apply and map existing machine teaching principles ontoa contextual machine teaching implementation in a commutingsetting. The aim is to highlight areas in machine teaching theorythat requires more attention. The main contribution of this workis an increased focus on available features, the features space andthe potential to transfer some of the domain expert’s explanatorypowers to the machine learning system.

Building interpretable machine learning agents is a challenge that needs to be addressed to make the agents trustworthy and align the usage of the technology with human values. In this work, we focus on how to evaluate interpretability in a machine teaching setting, a settingthat involves a human domain expert as a teacher in relation to a machine learning agent. By using a prototype in a study, we discuss theinterpretability denition and show how interpretability can be evaluatedon a functional-, human- and application level. We end the paperby discussing open questions and suggestions on how our results can be transferable to other domains.

Machine Learning (ML) has been a prominent area of research within Artificial Intelligence (AI). ML uses mathematical models to recognize patterns in large and complex data sets to aid decision making in different application areas, such as image and speech recognition, consumer recommendations, fraud detection and more. ML systems typically go through a training period in which the system encounters and learns about the data; further, this training often requires some degree of human intervention. Interactive machine learning (IML) refers to ML applications that depend on continuous user interaction. From an HCI perspective, how humans interact with and experience ML models in training is the main focus of this workshop proposal. In this workshop we focus on the user experience (UX) of Interactive Machine Learning, a topic with implications not only for usability but also for the long-term success of the IML systems themselves.

Mobile apps are an increasingly important part of public transport, and can be seen as part of the journey experience. Personalisation of the app is then one aspect of the experience that, for example, can give travellers a possibility to save favourite journeys for easy access. Such a list of journeys can be extensive and inaccurate if it doesn’t consider the traveller’s context. Making an app context aware and present upcoming journeys transforms the app experience in a personal direction, especially for commuters. By using historical personal contextual data, a travel app can present probable journeys or accurately predict and present an upcoming journey with departure times. The predictions can take place when the app is started or be used to remind a commuter when it is time to leave in order to catch a regularly travelled bus or train.

To address this research opportunity we have created an individually trained Machine Learning (ML) agent that we added to a publicly available commuter app. The added part of the app uses weekday, time, user activity and location to predict a user’s upcoming journey. Predictions are made when the app starts and departure times for the most probable transport are presented to the traveller. In our case a commuter only makes a few journey searches in the app every day which implies that, based on our contextual parameters, it will take at least some weeks to create journey patterns that can give acceptable accuracy for the predictions. In the work we present here, we focus on how to handle this cold start problem e.g. the situation when no or inaccurate historical data is available for the Machine Learning agent to train from. These situations will occur both initially when no data exists and due to concept drift originating from changes in travel patterns. In these situations, no predictions or only inaccurate predictions of upcoming journeys can be made.    

We present experiences and evaluate results gathered when designing the interactions needed for the MT session as well as design decisions for the ML pipeline and the ML agent. The user’s interaction with the ML agent during the teaching session is a crucial factor for the success. During the teaching session, information on what the agent already has learnt has to be presented to the user as well as possibilities to unlearn obsolete commute patterns and to teach new. We present a baseline that shows an idealised situation and the amount of training data that the user needs to add in a MT session to reach acceptable accuracy in predictions. Our main contribution is user evaluated design proposals for the MT session.

Using individually trained ML agents opens up opportunities to protect personal data and this approach can be used to create mobile applications that is independent of local transport providers and thus act on open data on a global scale.

The success of neural networks largely builds on their ability to create internal knowledge representations from real-world high-dimensional data, such as images, sound, or text. Approaches to extract and present these representations, in order to explain a neural network’s decision, is an active and multifaceted research field. To gain a deeper understanding of a central aspect of this field, we performed a targeted literature review focusing on research that aims to associate internal representations with human understandable concepts. By using deductive nomological explanations combined with causality theories as an analytical lens, we analyse nine carefully selected research papers. We find our analytical lens, the explanation structure and causality, useful to understand what can be expected, and not expected, from explanations inferred from neural networks. The analysis additionally uncovers an ambiguity in the reviewed literature related to the goal: is it (a) understanding the ML model, (b) the training data or (c) actionable explanations that are true-to-the-domain?

Central to deep learning is an ability to generalise within a target domain consistent with human beliefs within the same domain. A label inferred by the neural network then maps to a human mental representation of a, to the label, corresponding concept. If an explanation concerning why a specific decision is promoted it is important that we move from average case performance metrics towards interpretable explanations that build on human understandable concepts connected to the promoted label. In this work, we use Explainable Artificial Intelligence (XAI) methods to investigate if internal knowledge representations in trained neural networks are aligned and generalise in correspondence to human mental representations. Our findings indicate an, in neural networks, epistemic misalignment between machine and human knowledge representations. Consequently, if the goal is classifications explainable for en users we can question the usefulness of neural networks trained without considering concept alignment.