Academia.eduAcademia.edu

Outline

Title
Abstract
Figures
Introduction
Al Related Techniques
Time-Series Data
Complex Data
Curriculum Design
How to Select a CL Method
Discussion and Conclusions
Trends and Future Developments
References
All Topics
Psychology
Cognitive Science

Human-in-the-loop machine learning: a state of the art

Profile image of Ángel Fernández LealÁngel Fernández Leal

Artificial Intelligence Review

https://doi.org/10.1007/S10462-022-10246-W
visibility

…

description

50 pages

link

1 file

Abstract

Researchers are defining new types of interactions between humans and machine learning algorithms generically called human-in-the-loop machine learning. Depending on who is in control of the learning process, we can identify: active learning, in which the system remains in control; interactive machine learning, in which there is a closer interaction between users and learning systems; and machine teaching, where human domain experts have control over the learning process. Aside from control, humans can also be involved in the learning process in other ways. In curriculum learning human domain experts try to impose some structure on the examples presented to improve the learning; in explainable AI the focus is on the ability of the model to explain to humans why a given solution was chosen. This collaboration between AI models and humans should not be limited only to the learning process; if we go further, we can see other terms that arise such as Usable and Useful AI. In this paper ...

Figures (10)
Fig. 1 Human-in-the-loop machine learning (HITL-ML) mind map
Fig. 1 Human-in-the-loop machine learning (HITL-ML) mind map
Fig. 2 Active learning (AL) mind map
Fig. 2 Active learning (AL) mind map
Fig. 4 Interactive machine learning (IML) mind map 3.1 IML definitions
Fig. 4 Interactive machine learning (IML) mind map 3.1 IML definitions
aa ae Ce ee ee +7 In Fig. 5 we can see a schematic representation of the IML process, we can see that it is a freer scheme than the one of AL represented in Fig. 3. In this diagram interactiv- ity becomes more important and there are different tasks that the human can do inter- actively (such as selecting examples, labeling cases, etc.) depending on which is more appropriate at each step (Suh et al. 2019). Sea Se ee Se Se ee a a ee a ee ee This leads to another difference between AL and IML systems: how to evaluate them Fiebrink et al. (2011) already stated that in an IML system the evaluation of the model should go beyond their accuracy and should include subjective judgments of propertie such as cost, confidence, complexity, etc. AL focuses on building better models in a algorithm-centered evaluation, but in IML systems we have to take into account humat factors, so there is also a human-centered evaluation, focusing on the utility and effec tiveness of the application for end-users. Boukhelifa et al. (2018) stated that couplin; both evaluations in IML systems can bring forth insights that can play an important rol in addressing the “black-box” effect of machine learning algorithms. In Fig. 5 we can see a schematic representation of the IML process, we can see tha
aa ae Ce ee ee +7 In Fig. 5 we can see a schematic representation of the IML process, we can see that it is a freer scheme than the one of AL represented in Fig. 3. In this diagram interactiv- ity becomes more important and there are different tasks that the human can do inter- actively (such as selecting examples, labeling cases, etc.) depending on which is more appropriate at each step (Suh et al. 2019). Sea Se ee Se Se ee a a ee a ee ee This leads to another difference between AL and IML systems: how to evaluate them Fiebrink et al. (2011) already stated that in an IML system the evaluation of the model should go beyond their accuracy and should include subjective judgments of propertie such as cost, confidence, complexity, etc. AL focuses on building better models in a algorithm-centered evaluation, but in IML systems we have to take into account humat factors, so there is also a human-centered evaluation, focusing on the utility and effec tiveness of the application for end-users. Boukhelifa et al. (2018) stated that couplin; both evaluations in IML systems can bring forth insights that can play an important rol in addressing the “black-box” effect of machine learning algorithms. In Fig. 5 we can see a schematic representation of the IML process, we can see tha
4.1 Antecedents of machine teaching
4.1 Antecedents of machine teaching
ultimately, available for non-machine learning experts. This set of tools seeks to automate the decision on what learning algorithms to use, what hyper-parameters to select or which features are more relevant for a certain model. Furthermore, they provide a means of model evaluation and optimization. Many cloud platforms (e.g., AWS, Azure, Google) are includ- ing AutoML tools as part of their ML stacks. The problem with the AutoML approach is that it produces black-box models that are inadequate for scenarios requiring transparency, and also can be impractical to use in cases where there are not significant labeled data available.
ultimately, available for non-machine learning experts. This set of tools seeks to automate the decision on what learning algorithms to use, what hyper-parameters to select or which features are more relevant for a certain model. Furthermore, they provide a means of model evaluation and optimization. Many cloud platforms (e.g., AWS, Azure, Google) are includ- ing AutoML tools as part of their ML stacks. The problem with the AutoML approach is that it produces black-box models that are inadequate for scenarios requiring transparency, and also can be impractical to use in cases where there are not significant labeled data available.
Human-in-the-loop machine learning: a state of the art U7). Two are the motivations for applying CL, that is, to guide, regularizing the training towards better regions in parameter space, as from the perspective of the optimizatior problem; and to denoise, focusing on high-confidence easier area to mitigate the inter. ference of noisy data as from the perspective of data distribution. In this sense, most o: the applications of CL can be classified into these two groups (Wang et al. 2021). The guide motivation group involves difficult tasks where direct training on them results i poor performance or slow convergence. CL strategies guide the training from easier task: to the target ones, being examples sparse-reward Reinforcement Learning (Florensa et al 2017) or training Generative Adversarial Networks (Soviany et al. 2020). It also concerns tasks where the target distribution is quite different from the training distribution, and < good curriculum helps to guide the training to adapt to the target distribution. In this case domain adaptation setting (Zhang et al. 2019b) and imbalanced classification (Wang et al 2019) are representative settings. The denoise motivation group involves noisy or hetero: geneous training data sets and CL strategies could help making the training faster, more robust, and more generalizable, being a popular application the neural machine translatior field (Kumar et al. 2019). The way curriculum strategies have been defined leaves a lot of work to the teacher. Tc
Human-in-the-loop machine learning: a state of the art U7). Two are the motivations for applying CL, that is, to guide, regularizing the training towards better regions in parameter space, as from the perspective of the optimizatior problem; and to denoise, focusing on high-confidence easier area to mitigate the inter. ference of noisy data as from the perspective of data distribution. In this sense, most o: the applications of CL can be classified into these two groups (Wang et al. 2021). The guide motivation group involves difficult tasks where direct training on them results i poor performance or slow convergence. CL strategies guide the training from easier task: to the target ones, being examples sparse-reward Reinforcement Learning (Florensa et al 2017) or training Generative Adversarial Networks (Soviany et al. 2020). It also concerns tasks where the target distribution is quite different from the training distribution, and < good curriculum helps to guide the training to adapt to the target distribution. In this case domain adaptation setting (Zhang et al. 2019b) and imbalanced classification (Wang et al 2019) are representative settings. The denoise motivation group involves noisy or hetero: geneous training data sets and CL strategies could help making the training faster, more robust, and more generalizable, being a popular application the neural machine translatior field (Kumar et al. 2019). The way curriculum strategies have been defined leaves a lot of work to the teacher. Tc
The most commonly used nomenclature used in XAI communities includes terms such as understandability, comprehensibility, interpretability, explainability and transparency. In all the above terms, understandability comes up as the most important concept in XAI and it has to be considered on the one hand as model understandability and on the other, human understandability. This is the reason why the definition of XAI refers to the concept of audi- ence, defined as the users of the model, as the cognitive skills and pursued goal of those users have to be taken into account jointly with the intelligibility and comprehensibility
The most commonly used nomenclature used in XAI communities includes terms such as understandability, comprehensibility, interpretability, explainability and transparency. In all the above terms, understandability comes up as the most important concept in XAI and it has to be considered on the one hand as model understandability and on the other, human understandability. This is the reason why the definition of XAI refers to the concept of audi- ence, defined as the users of the model, as the cognitive skills and pursued goal of those users have to be taken into account jointly with the intelligibility and comprehensibility
Fig. 10 Human-in-the-loop machine learning (HITL-ML-relations) mind map . ° << so ee The first trend is that interactivity has increasing importance in the development f ML models, because as we move towards more human control we also move towards © more interactivity. For example, when in AL a learner requests a human oracle to label -xamples, the questions must be presented in a way that the human understands them. This mplies that the characteristics of usability in the learning process (clarity, consistency, effi- iency, etc.) is particularly relevant. As interactivity increases, these aspects become more mportant to the point that, in the case of MT, the measurement of metrics, such as produc- ivity, interpretability, robustness, and scaling is considered essential to check whether a system is successful.
Fig. 10 Human-in-the-loop machine learning (HITL-ML-relations) mind map . ° << so ee The first trend is that interactivity has increasing importance in the development f ML models, because as we move towards more human control we also move towards © more interactivity. For example, when in AL a learner requests a human oracle to label -xamples, the questions must be presented in a way that the human understands them. This mplies that the characteristics of usability in the learning process (clarity, consistency, effi- iency, etc.) is particularly relevant. As interactivity increases, these aspects become more mportant to the point that, in the case of MT, the measurement of metrics, such as produc- ivity, interpretability, robustness, and scaling is considered essential to check whether a system is successful.

Related papers

Explainable Active Learning (XAL): An Empirical Study of How Local Explanations Impact Annotator Experience
Vera Liao

2020

The wide adoption of Machine Learning technologies has created a rapidly growing demand for people who can train ML models. Some advocated the term &quot;machine teacher&quot; to refer to the role of people who inject domain knowledge into ML models. One promising learning paradigm is Active Learning (AL), by which the model intelligently selects instances to query the machine teacher for labels. However, in current AL settings, the human-AI interface remains minimal and opaque. We begin considering AI explanations as a core element of the human-AI interface for teaching machines. When a human student learns, it is a common pattern to present one&#39;s own reasoning and solicit feedback from the teacher. When a ML model learns and still makes mistakes, the human teacher should be able to understand the reasoning underlying the mistakes. When the model matures, the machine teacher should be able to recognize its progress in order to trust and feel confident about their teaching outco...

View PDFchevron_right
Explainable machine learning methods as a tool for higher education improvement
Vladislav Miškovic

ZBORNIK RADOVA UNIVERZITETA SINERGIJA

Group work and communication problems that occurred during the pandemic have significantly increased the use of new technologies, especially in higher education, where methods and systems of artificial intelligence, primarily machine learning, are increasingly used in the automation of various aspects of the teaching process. An application of these methods introduces completely new ethical and legal problems, related to the nature and manner of use of automatically created knowledge in various social processes. Confidence in automated generated knowledge and the issue of responsibility for the results of decisions that are made on the basis of that knowledge are trying to be solved by using the so-called Explainable Machine Learning methods. The paper discusses the application of these methods in the automation of various aspects of higher education, and demonstrates practical examples of their use in predicting student performance and teacher evaluation. In the practical examples,...

View PDFchevron_right
Explainable AI in education: Fostering human oversight and shared responsibility An initiative of the
Lezel Roddeck

European Commission Publications Office, 2025

This report explores how explainable artificial intelligence (XAI) can foster trust, transparency, and human agency in educational environments. It provides a comprehensive introduction to the technical foundations of XAI, discusses the ethical and practical importance of human-centred explanations, and addresses the critical role of XAI in supporting informed and responsible learning practices. My contribution focussed on navigating compliance with the evolving European legal framework, particularly the AI Act, the GDPR, and related digital laws. It examines how educational AI systems must balance innovation with transparency, fairness, accountability, and protection of fundamental rights. The report clarifies roles and responsibilities for providers, deployers, and developers of AI systems in education, emphasising explainability obligations and human oversight requirements. Through fictional use cases and technical recommendations, it illustrates how institutions and developers can operationalise legal compliance while maintaining pedagogical effectiveness and user trust. In addition, the report analyses stakeholder perspectives on XAI, including learners, educators, and policymakers, and outlines core competencies educators need to critically engage with AI tools. Practical examples, competence frameworks, and strategic guidelines aim to support the responsible adoption of AI, ensuring that AI-enhanced educational environments are ethical, transparent, and human-centred. DOI:10.13140/RG.2.2.36045.96481

View PDFchevron_right
The Roles and Modes of Human Interactions with Automated Machine Learning Systems
Khuat Thanh Tung

2022

As automated machine learning (AutoML) systems continue to progress in both sophistication and performance, it becomes important to understand the 'how' and 'why' of human-computer interaction (HCI) within these frameworks, both current and expected. Such a discussion is necessary for optimal system design, leveraging advanced data-processing capabilities to support decision-making involving humans, but it is also key to identifying the opportunities and risks presented by ever-increasing levels of machine autonomy. Within this context, we focus on the following questions: (i) How does HCI currently look like for state-of-the-art AutoML algorithms, especially during the stages of development, deployment, and maintenance? (ii) Do the expectations of HCI within AutoML frameworks vary for different types of users and stakeholders? (iii) How can HCI be managed so that AutoML solutions acquire human trust and broad acceptance? (iv) As AutoML systems become more autonomous and capable of learning from complex open-ended environments, will the fundamental nature of HCI evolve? To consider these questions, we project existing literature in HCI into the space of AutoML; this connection has, to date, largely been unexplored. In so doing, we review topics including user-interface design, human-bias mitigation, and trust in artificial intelligence (AI). Additionally, to rigorously gauge the future of HCI, we contemplate how AutoML may manifest in effectively open-ended environments. This discussion necessarily reviews projected developmental pathways for AutoML, such as the incorporation of reasoning, although the focus remains on how and why HCI may occur in such a framework rather than on any implementational details. Ultimately, this review serves to identify key research directions aimed at better facilitating the roles and modes of human interactions with both current and future AutoML systems.

View PDFchevron_right
Explainable AI
Britta Wrede

KI - Künstliche Intelligenz

View PDFchevron_right
The Role of Human Knowledge in Explainable AI
Marco Brambilla

Data

As the performance and complexity of machine learning models have grown significantly over the last years, there has been an increasing need to develop methodologies to describe their behaviour. Such a need has mainly arisen due to the widespread use of black-box models, i.e., high-performing models whose internal logic is challenging to describe and understand. Therefore, the machine learning and AI field is facing a new challenge: making models more explainable through appropriate techniques. The final goal of an explainability method is to faithfully describe the behaviour of a (black-box) model to users who can get a better understanding of its logic, thus increasing the trust and acceptance of the system. Unfortunately, state-of-the-art explainability approaches may not be enough to guarantee the full understandability of explanations from a human perspective. For this reason, human-in-the-loop methods have been widely employed to enhance and/or evaluate explanations of machine...

View PDFchevron_right
Explanatory learner models: Why machine learning (alone) is not the answer
Kenneth Koedinger

British Journal of Educational Technology, 2019

Using data to understand learning and improve education has great promise. However, the promise will not be achieved simply by AI and Machine Learning researchers developing innovative models that more accurately predict labeled data. As AI advances, modeling techniques and the models they produce are getting increasingly complex, often involving tens of thousands of parameters or more. Though strides towards interpretation of complex models are being made in core machine learning communities, it remains true in these cases of “black box” modeling that research teams may have little possibility to peer inside to try understand how, why, or even whether such models will work when applied beyond the data on which they were built. Rather than relying on AI expertise alone, we suggest that learning engineering teams bring interdisciplinary expertise to bear to develop explanatory learner models that provide interpretable and actionable insights in addition to accurate prediction. We des...

View PDFchevron_right
Expl(AI)n It to Me – Explainable AI and Information Systems Research
Christof Weinhardt

Business &amp; Information Systems Engineering, 2021

View PDFchevron_right
Thematic Working Group 4 - State of the Art in Thinking About Machine Learning: Implications for Education
Jason Zagami

2019

Introduction Examples of machine learning (ML) that many people will be familiar with are self-driving cars, online recommendations from Amazon or Netflix, voice controlled digital assistants on mobile phones and spam filters. More broadly, applications of machine learning are widespread and increasing across most areas of human endeavour including agriculture (Liakos, Busato et al. 2018), the energy industry (Cheng and Yu 2019), e-commerce (Zhang, Yang et al. 2018) , fault detection and diagnosis across most types of machinery (Zhao, Yan et al. 2019) and healthcare (Faust, Hagiwara et al. 2018). Likewise, in education, machine learning is becoming more widespread and has been used, for example, for improving curriculum design (Ball, Duhadway et al. 2019), predicting students’ grades (Livieris, Drakopoulou et al. 2019), recommending higher education courses to students (Obeid, Lahoud et al. 2018); student modelling for intelligent tutoring systems (Conati, PorayskaPomsta et al. 2018).

View PDFchevron_right
A Rationale-Centric Framework for Human-in-the-loop Machine Learning
Yue Zhang, Jinghui Lu

Proceedings of the 60th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)

We present a novel rationale-centric framework with human-in-the-loop-Rationales-centric Double-robustness Learning (RDL)-to boost model out-of-distribution performance in few-shot learning scenarios. By using static semi-factual generation and dynamic humanintervened correction, RDL exploits rationales (i.e. phrases that cause the prediction), human interventions and semi-factual augmentations to decouple spurious associations and bias models towards generally applicable underlying distributions, which enables fast and accurate generalisation. Experimental results show that RDL leads to significant prediction benefits on both in-distribution and out-of-distribution tests compared to many state-of-the-art benchmarks-especially for few-shot learning scenarios. We also perform extensive ablation studies to support in-depth analyses of each component in our framework.

View PDFchevron_right

References (169)

  1. Abdul A, Vermeulen J, Wang D et al (2018) Trends and trajectories for explainable, accountable and intel- ligible systems: an hci research agenda. In: Proceedings of the 2018 CHI conference on human factors in computing systems. Association for Computing Machinery, New York, NY, USA, CHI '18, pp 1-18, https:// doi. org/ 10. 1145/ 31735 74. 31741 56
  2. Abiteboul S, Buneman P, Suciu D (2000) Data on the web: from relations to semistructured data and XML. Morgan Kaufmann, Data Management Systems Series
  3. Adadi A, Berrada M (2018) Peeking inside the black-box: a survey on explainable artificial intelligence (XAI). IEEE Access 6:52,138-52,160. https:// doi. org/ 10. 1109/ ACCESS. 2018. 28700 52
  4. Aggarwal CC, Kong X, Gu Q et al (2014) Active learning: a survey. Data classification: algorithms and applications. Chapman and Hall/CRC, Boca Raton, pp 599-634
  5. Amazon (2022) Amazon mechanical turk. https:// www. mturk. com/. Accessed on 23 Mar 2022
  6. Amershi S, Cakmak M, Knox WB et al (2014) Power to the people: the role of humans in interactive machine learning. AI Magazine 35(4):105-120. https:// doi. org/ 10. 1609/ aimag. v35i4. 2513
  7. Angluin D (1987) Learning regular sets from queries and counterexamples. Inf Comput 75(2):87-106. https:// doi. org/ 10. 1016/ 0890-5401(87) 90052-6
  8. Arras GL, Montavon, Müller KR, Samek W (2017) Explaining recurrent neural network predictions in sen- timent analysis. In: EMNLP'17 workshop on computational approaches to subjectivity, sentiment and social media analysis, https:// doi. org/ 10. 18653/ v1/ W17-5221
  9. Barakat NH, Bradley AP (2007) Rule extraction from support vector machines: a sequential covering approach. IEEE Trans Knowl Data Eng 19(6):729-741. https:// doi. org/ 10. 1109/ TKDE. 2007. 190610
  10. Barredo Arrieta A, Díaz-Rodríguez N, Del Ser J et al (2020) Explainable artificial intelligence (xai): con- cepts, taxonomies, opportunities and challenges toward responsible AI. Inf Fusion 58:82-115. https:// doi. org/ 10. 1016/j. inffus. 2019. 12. 012
  11. Bau D, Zhou B, Khosla A et al (2017) Network dissection: Quantifying interpretability of deep visual rep- resentations. In: 2017 IEEE conference on computer vision and pattern recognition (CVPR), 3319- 3327, https:// doi. org/ 10. 1109/ CVPR. 2017. 354, https:// ieeex plore. ieee. org/ docum ent/ 80998 37
  12. Begeja L, Renger B, Gibbon D et al (2004) Interactive machine learning techniques for improving SLU models. In: Proceedings of the HLT-NAACL 2004 workshop on spoken language understanding for conversational systems and higher level linguistic information for speech processing. Association for Computational Linguistics, Boston, Massachusetts, USA, 10-16, https:// aclan tholo gy. org/ W04-3003
  13. Bengio Y, Louradour J, Collobert R et al (2009) Curriculum learning. In: Proceedings of the 26th annual international conference on machine learning. Association for Computing Machinery, New York, NY, USA, ICML '09, 41-48, https:// doi. org/ 10. 1145/ 15533 74. 15533 80
  14. Bennetot A, Laurent JL, Chatila R et al (2019) Towards explainable neural-symbolic visual reasoning. arxiv: 1909. 09065
  15. Berg S, Kutra D, Kroeger T et al (2019) Ilastik: interactive machine learning for (bio)image analysis. Nat Methods 16(12):1226-1232. https:// doi. org/ 10. 1038/ s41592-019-0582-9
  16. Berghel H (1997) Cyberspace 2000: dealing with information overload. Commun ACM 40(2):19-24. https:// doi. org/ 10. 1145/ 253671. 253680
  17. Blumberg R, Atre S (2003) The problem with unstructured data. DM Rev 13(42-49):62
  18. Bonwell CC, Eison JA (1991) Active learning: creating excitement in the classroom. 1991 ASHE-ERIC higher education reports. ERIC Clearinghouse on Higher Education, The George Washington Univer- sity, One Dupont Circle, Suite 630, Washington, DC 20036-1183
  19. Boukhelifa N, Bezerianos A, Lutton E (2018) Evaluation of interactive machine learning systems. In: Zhou J, Chen F (eds) Human and machine learning: visible, explainable, trustworthy and transparent. Springer, Cham, pp 341-360
  20. Budd S, Robinson EC, Kainz B (2021) A survey on active learning and human-in-the-loop deep learning for medical image analysis. Med Image Anal 71(102):062. https:// doi. org/ 10. 1016/j. media. 2021. 102062
  21. Carlson G (2015) What eactly is complex data? https:// www. ayasdi. com/ exact ly-compl ex-data/. Accessed on 04 Mar 2021
  22. Castle E (2017) 7 signs you're dealing with complex data. https:// www. sisen se. com/ blog/7-signs-youre- deali ng-with-compl ex-data/. Accessed on 04 Mar 2022
  23. Che Z, Purushotham S, Khemani R et al (2015) Distilling knowledge from deep networks with applica- tions to healthcare domain. arXiv e-prints arxiv: 1512. 03542 [stat.ML]
  24. Che Z, Purushotham S, Khemani R et al (2017) Interpretable deep models for ICU outcome prediction. In: AMIA annual symposium proceedings, 371-380, https:// pubmed. ncbi. nlm. nih. gov/ 28269 832/ Chen Z, Li J, Wei L (2007) A multiple kernel support vector machine scheme for feature selection and rule extraction from gene expression data of cancer tissue. Artif Intell Med 41(2):161-175. https:// doi. org/ 10. 1016/j. artmed. 2007. 07. 008
  25. Chen Y, Singla A, Aodha OM et al (2018) Understanding the role of adaptivity in machine teaching: The case of version space learners. In: Proceedings of the 32nd international conference on neural information processing systems. Curran Associates Inc., Red Hook, NY, USA, NIPS'18, 1483- 1493, https:// dl. acm. org/ doi/ abs/ 10. 5555/ 33269 43. 33270 79
  26. Choi E, Bahadori T, Schuetz A et al (2016) Retain: Interpretable predictive model in healthcare using reverse time attention mechanism. In: Proceedings of the 30th international conference on neural information processing systems. Curran Associates Inc., Red Hook, NY, USA, NIPS'16, 3512-3520
  27. Churchill EF, van Allen P, Kuniavsky M (2018) Designing AI. Interactions 25(6):34-37. https:// doi. org/ 10. 1145/ 32817 64
  28. Cirik V, Hovy E, Morency LP (2016) Visualizing and understanding curriculum learning for long short- term memory networks. arXiv e-prints arxiv: 1611. 06204 [cs.CL]
  29. Cohn D, Atlas L, Ladner R (1994) Improving generalization with active learning. Mach Learn 15(2):201-221. https:// doi. org/ 10. 1007/ BF009 93277
  30. Cohn DA, Ghahramani Z, Jordan MI (1996) Active learning with statistical models. J Artif Intell Res 4(1):129-145. https:// doi. org/ 10. 5555/ 16227 37. 16227 44
  31. d'Avila Garcez A, Gori M, Lamb LC et al (2019) Neural-symbolic computing: an effective methodology for principled integration of machine learning and reasoning. arXiv e-prints arxiv: 1905. 06088 [cs.AI]
  32. Das A, Rad P (2020) Opportunities and challenges in explainable artificial intelligence (XAI): a survey. arXiv e-prints arxiv: 2006. 11371 [cs.CV]
  33. De Angeli K, Gao S, Alawad M et al (2021) Deep active learning for classifying cancer pathology reports. BMC Bioinform 22(1):1-25
  34. Devidze R, Mansouri F, Haug L et al (2020) Understanding the power and limitations of teaching with imperfect knowledge. In: Bessiere C (ed) Proceedings of the twenty-ninth international joint con- ference on artificial intelligence, IJCAI-20. International Joint Conferences on Artificial Intelli- gence Organization, 2647-2654, https:// doi. org/ 10. 24963/ ijcai. 2020/ 367
  35. Diamant E (2009) Machine learning: When and where the horses went astray? In: Zhang Y (ed) Machine learning. InTech, London, pp 1-18. https:// doi. org/ 10. 5772/ 9156
  36. Diamant E (2006) Learning to understand image content: Machine learning versus machine teaching alternative. In: 2006 International conference on information technology: research and education, 26-29, https:// doi. org/ 10. 1109/ ITRE. 2006. 381526
  37. Donadello I, Kessler F, Dragoni M et al (2019) Persuasive explanation of reasoning inferences on die- tary data. In: Joint proceedings of the 6th international workshop on dataset profilling and search and the 1st workshop on semantic explainability co-located with the 18th international semantic web conference (ISWC 2019)
  38. Donmez P, Carbonell JG (2008) Proactive learning: Cost-sensitive active learning with multiple imper- fect oracles. In: Proceedings of the 17th ACM conference on information and knowledge manage- ment. Association for Computing Machinery, New York, NY, USA, CIKM '08, 619-628, https:// doi. org/ 10. 1145/ 14580 82. 14581 65
  39. Dudley JJ, Kristensson PO (2018) A review of user interface design for interactive machine learning. ACM Trans Interact Intell Syst. https:// doi. org/ 10. 1145/ 31855 17
  40. El-Hasnony IM, Elzeki OM, Alshehri A et al (2022) Multi-label active learning-based machine learning model for heart disease prediction. Sensors. https:// doi. org/ 10. 3390/ s2203 1184
  41. Elman JL (1993) Learning and development in neural networks: the importance of starting small. Cogni- tion 48(1):71-99. https:// doi. org/ 10. 1016/ 0010-0277(93) 90058-4
  42. Fadhil A, Wang Y (2018) Towards automatic & personalised mobile health interventions: an interactive machine learning perspective. arXiv e-prints arxiv: 1803. 01842 [cs.CY]
  43. Fails JA, Olsen DR (2003) Interactive machine learning. In: Proceedings of the 8th international confer- ence on intelligent user interfaces. Association for Computing Machinery, New York, NY, USA, IUI '03, 39-45, https:// doi. org/ 10. 1145/ 604045. 604056
  44. Fiebrink RA (2011) Real-time human interaction with supervised learning algorithms for music compo- sition and performance. PhD thesis, Computer Science Dept. Princeton University, Princeton, NJ, USA, https:// dl. acm. org/ doi/ book/ 10. 5555/ 21257 76
  45. Fiebrink R, Cook PR (2010) The wekinator: a system for real-time, interactive machine learning in music. In: Proceedings of The Eleventh International Society for Music Information Retrieval Conference (ISMIR 2010), Utrecht
  46. Fiebrink R, Cook PR, Trueman D (2011) Human model evaluation in interactive supervised learning. In: Proceedings of the SIGCHI conference on human factors in computing systems. Association for Computing Machinery, New York, NY, USA, CHI '11, 147-156, https:// doi. org/ 10. 1145/ 19789 42. 19789 65
  47. Florensa C, Held D, Wulfmeier M et al (2017) Reverse curriculum generation for reinforcement learn- ing. In: Levine S, Vanhoucke V, Goldberg K (eds) Proceedings of the 1st annual conference on robot learning, proceedings of machine learning research, vol 78. PMLR, 482-495, http:// proce edings. mlr. press/ v78/ flore nsa17a. html
  48. Fogarty J, Tan D, Kapoor A et al (2008) Cueflik: Interactive concept learning in image search. In: Pro- ceedings of the SIGCHI conference on human factors in computing systems. Association for Com- puting Machinery, New York, NY, USA, CHI '08, 29-38, https:// doi. org/ 10. 1145/ 13570 54. 13570 61
  49. Gaonkar B, Shinohara TR, Davatzikos C (2015) Interpreting support vector machine models for multi- variate group wise analysis in neuroimaging. Med Image Anal 24(1):190-204. https:// doi. org/ 10. 1016/j. media. 2015. 06. 008
  50. Goodman B, Flaxman S (2017) European union regulations on algorithmic decision-making and a "right to explanation." AI Mag 38(3):50-57. https:// doi. org/ 10. 1609/ aimag. v38i3. 2741
  51. Gunning D (2017) Explainable artificial intelligence (xAI). Tech. rep., Defense Advanced Research Pro- jects Agency (DARPA), https:// www. darpa. mil/ progr am/ expla inable-artifi cial-intel ligen ce
  52. Hacohen G, Weinshall D (2019) On the power of curriculum learning in training deep networks. In: Chaudhuri K, Salakhutdinov R (eds) Proceedings of the 36th international conference on machine learning, ICML 2019, 9-15 June 2019, Long Beach, California, USA, Proceedings of Machine Learning Research, vol 97. PMLR, 2535-2544, http:// proce edings. mlr. press/ v97/ hacoh en19a. html
  53. Hara S, Hayashi K (2018) Making tree ensembles interpretable: A bayesian model selection approach. In: Storkey A, Perez-Cruz F (eds) Proceedings of the twenty-first international conference on arti- ficial intelligence and statistics, proceedings of machine learning research, vol 84. PMLR, 77-85, https:// proce edings. mlr. press/ v84/ hara1 8a. html
  54. He X, Zhao K, Chu X (2021) AutoML: a survey of the state-of-the-art. Knowl Based Syst 212(106):622. https:// doi. org/ 10. 1016/j. knosys. 2020. 106622
  55. Heimerl F, Koch S, Bosch H et al (2012) Visual classifier training for text document retrieval. IEEE Trans Vis Comput Graphics 18(12):2839-2848. https:// doi. org/ 10. 1109/ TVCG. 2012. 277
  56. Hills TT, Todd PM, Lazer D et al (2015) Exploration versus exploitation in space, mind, and society. Trends Cogn Sci 19(1):46-54. https:// doi. org/ 10. 1016/j. tics. 2014. 10. 004
  57. Hipke K, Toomim M, Fiebrink R et al (2014) Beatbox: End-user interactive definition and training of recognizers for percussive vocalizations. In: Proceedings of the 2014 international working con- ference on advanced visual interfaces. Association for Computing Machinery, New York, NY, USA, AVI '14, 121-124, https:// doi. org/ 10. 1145/ 25981 53. 25981 89
  58. Hoi SCH, Jin R, Zhu J et al (2006) Batch mode active learning and its application to medical image classification. In: Proceedings of the 23rd international conference on machine learning. Associa- tion for Computing Machinery, New York, NY, USA, ICML '06, 417-424, https:// doi. org/ 10. 1145/ 11438 44. 11438 97
  59. Holmberg L, Davidsson P, Linde P (2020) A feature space focus in machine teaching. In: 2020 IEEE international conference on pervasive computing and communications workshops (PerCom Work- shops), 1-2, https:// doi. org/ 10. 1109/ PerCo mWork shops 48775. 2020. 91561 75
  60. Holzinger A (2016) Interactive machine learning for health informatics: when do we need the human-in- the-loop? Brain Inform 3(2):119-131. https:// doi. org/ 10. 1007/ s40708-016-0042-6
  61. Holzinger A, Jurisica I (2014) Knowledge discovery and data mining in biomedical informatics: the future is in integrative, interactive machine learning solutions. In: Holzinger A, Jurisica I (eds) Interactive knowledge discovery and data mining in biomedical informatics: state-of- the-art and future challenges. Springer, Berlin, Heidelberg, pp 1-18. https:// doi. org/ 10. 1007/ 978-3-662-43968-5_1
  62. Holzinger A, Plass M, Kickmeier-Rust M et al (2019) Interactive machine learning: experimental evi- dence for the human in the algorithmic loop. Appl Intell 49(7):2401-2414. https:// doi. org/ 10. 1007/ s10489-018-1361-5
  63. Holzinger A, Biemann C, Pattichis CS, et al (2017) What do we need to build explainable AI systems for the medical domain? arXiv e-prints arxiv: 1712. 09923 [cs.AI]
  64. Ionescu RT, Alexe B, Leordeanu M et al (2016) How hard can it be? estimating the difficulty of visual search in an image. In: 2016 IEEE conference on computer vision and pattern recognition (CVPR), pp 2157-2166, https:// doi. org/ 10. 1109/ CVPR. 2016. 237
  65. Ishibashi T, Nakao Y, Sugano Y (2020) Investigating audio data visualization for interactive sound recogni- tion. In: Proceedings of the 25th international conference on intelligent user interfaces. Association for Computing Machinery, New York, NY, USA, IUI '20, 67-77, https:// doi. org/ 10. 1145/ 33773 25. 33774 83
  66. Jamieson KG, Jain L, Fernandez C et al (2015) Next: a system for real-world development, evaluation, and application of active learning. In: Cortes C, Lawrence N, Lee D et al (eds) Advances in neural infor- mation processing systems, vol 28. Curran Associates Inc, Red Hook Jiang L, Meng D, Zhao Q et al (2015) Self-paced curriculum learning. In: Proceedings of the twenty-ninth AAAI conference on artificial intelligence. AAAI Press, AAAI'15, 2694-2700, https:// doi. org/ 10. 5555/ 28865 21. 28866 96
  67. Jiang L, Liu S, Chen C (2019) Recent research advances on interactive machine learning. J Vis 22(2):401- 417. https:// doi. org/ 10. 1007/ s12650-018-0531-1
  68. Johns E, Mac Aodha O, Brostow GJ (2015) Becoming the expert-interactive multi-class machine teach- ing. In: Proceedings of the IEEE conference on computer vision and pattern recognition, 2616-2624, https:// doi. org/ 10. 1109/ CVPR. 2015. 72988 77
  69. Kabra M, Robie AA, Rivera-Alba M et al (2013) Jaaba: interactive machine learning for automatic annota- tion of animal behavior. Nat Methods 10(1):64-67. https:// doi. org/ 10. 1038/ nmeth. 2281
  70. Kapoor A, Lee B, Tan D et al (2010) Interactive optimization for steering machine classification. In: Pro- ceedings of the SIGCHI conference on human factors in computing systems. Association for Comput- ing Machinery, New York, NY, USA, CHI '10, 1343-1352, https:// doi. org/ 10. 1145/ 17533 26. 17535 29
  71. Kellenberger B, Tuia D, Morris D (2020) Aide: accelerating image-based ecological surveys with inter- active machine learning. Methods Ecol Evol 11(12):1716-1727. https:// doi. org/ 10. 1111/ 2041-210X. 13489
  72. Kim B, Patel K, Rostamizadeh A et al (2015) Scalable and interpretable data representation for high-dimen- sional, complex data. In: Proceedings of the twenty-ninth AAAI conference on artificial intelligence, Association for the Advancement of Artificial Intelligence (AAAI), Austin, Texas USA, 1763-1769, https:// ojs. aaai. org/ index. php/ AAAI/ artic le/ view/ 9474
  73. Koesten L, Simperl E (2021) Ux of data: making data available doesn't make it usable. Interactions 28(2):97-99. https:// doi. org/ 10. 1145/ 34488 88
  74. Kosmyna N, Tarpin-Bernard F, Rivet B (2015) Adding human learning in brain-computer interfaces (bcis): towards a practical control modality. ACM Trans Comput-Hum Interact. https:// doi. org/ 10. 1145/ 27231 62
  75. Krakovna V, Doshi-Velez F (2016) Increasing the interpretability of recurrent neural networks using hidden markov models. arXiv e-prints arxiv: 1606. 0532 [cond-mat.soft]
  76. Kumar M, Packer B, Koller D (2010) Self-paced learning for latent variable models. In: Lafferty J, Williams C, Shawe-Taylor J et al (eds) Advances in neural information processing systems. Curran Associates Inc, Red Hook, pp 1189-1197
  77. Kumar G, Foster G, Cherry C et al (2019) Reinforcement learning based curriculum optimization for neu- ral machine translation. In: Proceedings of the 2019 conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, Volume 1 (Long and Short Papers). Association for Computational Linguistics, Minneapolis, Minnesota, 2054-2061, https:// doi. org/ 10. 18653/ v1/ N19-1208, https:// www. aclweb. org/ antho logy/ N19-1208
  78. Laws F, Scheible C, Schütze H (2011) Active Learning with Amazon Mechanical Turk. In: Proceedings of the conference on empirical methods in natural language processing. Association for Computational Linguistics, USA, EMNLP '11, 1546-1556, https:// doi. org/ 10. 5555/ 21454 32. 21455 97
  79. LeCun Y, Bengio Y, Hinton G (2015) Deep learning. Nature 521(7553):436-444. https:// doi. org/ 10. 1038/ natur e14539
  80. Leslie D (2019) Understanding artificial intelligence ethics and safety: A guide for the responsible design and implementation of ai systems in the public sector. 10.5281/zenodo.3240529
  81. Lindvall M, Molin J, Löwgren J (2018) From machine learning to machine teaching: the importance of UX. Interactions 25(6):52-57. https:// doi. org/ 10. 1145/ 32828 60
  82. Lipton ZC (2018) The mythos of model interpretability: in machine learning, the concept of interpret- ability is both important and slippery. Queue 16(3):31-57. https:// doi. org/ 10. 1145/ 32363 86. 32413 40
  83. Liu W, Dai B, Humayun A et al (2017) Iterative machine teaching. In: Precup D, Teh YW (eds) Proceed- ings of the 34th international conference on machine learning, proceedings of machine learning research, vol 70. PMLR, 2149-2158, https:// proce edings. mlr. press/ v70/ liu17b. html
  84. Liu C, He S, Liu K et al (2018a) Curriculum learning for natural answer generation. In: Proceedings of the 27th international joint conference on artificial intelligence. AAAI Press, IJCAI'18, 4223- 4229, https:// doi. org/ 10. 24963/ ijcai. 2018/ 587
  85. Liu J, Lichtenberg T, Hoadley KA et al (2018b) An integrated TCGA pan-cancer clinical data resource to drive high-quality survival outcome analytics. Cell 173(2):400-416.e11. https:// doi. org/ 10. 1016/j. cell. 2018. 02. 052
  86. Liu W, Dai B, Li X et al (2018c) Towards black-box iterative machine teaching. In: Dy J, Krause A (eds) Proceedings of the 35th international conference on machine learning, proceedings of machine learning research, vol 80. PMLR, 3141-3149, https:// proce edings. mlr. press/ v80/ liu18b. html
  87. Liu Z, Feng X, Wang Y et al (2021) Self-paced learning enhanced neural matrix factorization for noise- aware recommendation. Knowl Based Syst 213(106):660. https:// doi. org/ 10. 1016/j. knosys. 2020. 106660
  88. Lopes M, Melo F, Montesano L (2009) Active learning for reward estimation in inverse reinforcement learning. Joint European conference on machine learning and knowledge discovery in databases. Springer, Berlin Heidelberg, pp 31-46
  89. Loyola-González O (2019) Black-box vs. white-box: understanding their advantages and weaknesses from a practical point of view. IEEE Access 7:154096-154113. https:// doi. org/ 10. 1109/ ACCESS. 2019. 29492 86
  90. Luo T, Kramer K, Samson S et al (2004) Active learning to recognize multiple types of plankton. In: Proceedings of the 17th international conference on pattern recognition, 2004. ICPR 2004., 478- 481 Vol.3, https:// doi. org/ 10. 1109/ ICPR. 2004. 13345 70
  91. Mahendran A, Vedaldi A (2015) Understanding deep image representations by inverting them. In: 2015 IEEE conference on computer vision and pattern recognition (CVPR), 5188-5196, https:// doi. org/ 10. 1109/ CVPR. 2015. 72991 55, https:// ieeex plore. ieee. org/ docum ent/ 72991 55
  92. Matiisen T, Oliver A, Cohen T et al (2020) Teacher-student curriculum learning. IEEE Trans Neural Netw Learn Syst 31(9):3732-3740. https:// doi. org/ 10. 1109/ TNNLS. 2019. 29349 06
  93. Mei S, Zhu X (2015) Using machine teaching to identify optimal training-set attacks on machine learn- ers. In: Proc. of the 29th AAAI conference on artificial intelligence, 2871-2877, https:// ojs. aaai. org/ index. php/ AAAI/ artic le/ view/ 9569
  94. Meske C, Bunde E, Schneider J et al (2022) Explainable artificial intelligence: objectives, stakeholders, and future research opportunities. Inf Syst Manag 39(1):53-63. https:// doi. org/ 10. 1080/ 10580 530. 2020. 18494 65
  95. Meza Martínez MA, Nadj M, Maedche A (2019) Towards an integrative theoretical framework of inter- active machine learning systems. In: Proceedings of the 27th European conference on information systems (ECIS), Stockholm & Uppsala, Sweden, https:// aisel. aisnet. org/ ecis2 019_ rp/ 172
  96. Michael CJ, Acklin D, Scheuerman J (2020) On interactive machine learning and the potential of cogni- tive feedback. arXiv e-prints arxiv: 2003. 10365 [cs.HC]
  97. Microsoft (2022) Qna maker. https:// www. qnama ker. ai/. Accessed on 23 Mar 2022
  98. Minh D, Wang HX, Li YF et al (2021) Explainable artificial intelligence: a comprehensive review. Artif Intell Rev. https:// doi. org/ 10. 1007/ s10462-021-10088-y
  99. Montavon G, Lapuschkin S, Binder A et al (2017) Explaining nonlinear classification decisions with deep Taylor decomposition. Pattern Recognit 65:211-222. https:// doi. org/ 10. 1016/j. patcog. 2016. 11. 008
  100. Mosqueira-Rey E, Alonso-Ríos D, Baamonde-Lozano A (2021) Integrating iterative machine teaching and active learning into the machine learning loop. Procedia Comput Sci 192:553-562. https:// doi. org/ 10. 1016/j. procs. 2021. 08. 057
  101. Mosqueira-Rey E, Hernández-Pereira E, Alonso-Ríos D et al (2022) A classification and review of tools for developing and interacting with machine learning systems. In: Proceedings of the 37th annual ACM symposium on applied computing. Association for Computing Machinery, New York, NY, USA, 1083-1092, https:// doi. org/ 10. 1145/ 34773 14. 35073 10
  102. Muir BM (1987) Trust between humans and machines, and the design of decision aids. Int J Man-Mach Stud 27(5):527-539. https:// doi. org/ 10. 1016/ S0020-7373(87) 80013-5
  103. Munro R (2020) Human-in-the-loop machine learning. Manning Publications, Shelter Island Nguyen DHM, Patrick JD (2014) Supervised machine learning and active learning in classification of radi- ology reports. J Am Med Inform Assoc 21(5):893-901. https:// doi. org/ 10. 1136/ amiaj nl-2013-002516
  104. Nguyen A, Dosovitskiy A, Yosinski J et al (2016) Synthesizing the preferred inputs for neurons in neural networks via deep generator networks. In: Proceedings of the 30th international conference on neural information processing systems. Curran Associates Inc., Red Hook, NY, USA, NIPS'16, 3395-3403, https:// doi. org/ 10. 5555/ 31573 82. 31574 77
  105. Nwana HS (1990) Intelligent tutoring systems: an overview. Artif Intell Rev 4(4):251-277. https:// doi. org/ 10. 1007/ BF001 68958
  106. Olsson F (2009) A literature survey of active machine learning in the context of natural language process- ing. Tech. rep., Swedish Institute of Computer Science, http:// urn. kb. se/ resol ve? urn= urn: nbn: se: ri: diva-23510
  107. O'Malley J (2018) Captcha if you can: how you've been training ai for years without realising it. https:// www. techr adar. com/ news/ captc ha-if-you-can-how-youve-been-train ing-ai-for-years-witho ut-reali sing-it Peng B, Li C, Li J et al (2021) Soloist: building task bots at scale with transfer learning and machine teach- ing. Trans Assoc Comput Linguist 9:807-824. https:// doi. org/ 10. 1162/ tacl_a_ 00399
  108. Penha G, Hauff C (2020) Curriculum learning strategies for IR. In: Jose JM, Yilmaz E, Magalhães J et al (eds) European conference on information retrieval: advances in information retrieval. Springer, Cham, pp 699-713. https:// doi. org/ 10. 1007/ 978-3-030-45439-5_ 46
  109. Platanios EA, Stretcu O, Neubig G et al (2019) Competence-based curriculum learning for neural machine translation. In: Proceedings of the 2019 conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, Volume 1 (Long and Short Papers). Association for Computational Linguistics, Minneapolis, Minnesota, pp 1162-1172, https:// doi. org/ 10. 18653/ v1/ N19-1119, https:// www. aclweb. org/ antho logy/ N19-1119
  110. Porter R, Theiler J, Hush D (2013) Interactive machine learning in data exploitation. Comput Sci Eng 15(5):12-20. https:// doi. org/ 10. 1109/ MCSE. 2013. 74
  111. Ramos G, Meek C, Simard P et al (2020) Interactive machine teaching: a human-centered approach to building machine-learned models. Hum Comput Interact 35(5-6):413-451. https:// doi. org/ 10. 1080/ 07370 024. 2020. 17349 31
  112. Reyes O, Pérez E, del Carmen Rodrıguez-Hernández M et al (2016) Jclal: a java framework for active learn- ing. J Mach Learn Res 17:1-5
  113. Ribeiro M, Grolinger K, Capretz MA (2015) MLaaS: Machine learning as a service. In: 2015 IEEE 14th international conference on machine learning and applications (ICMLA), 896-902, https:// doi. org/ 10. 1109/ ICMLA. 2015. 152
  114. Rubens N, Elahi M, Sugiyama M et al (2015) Active learning in recommender systems. In: Ricci F, Rokach L, Shapira B (eds) Recommender systems handbook. Springer, Boston, pp 809-846
  115. Rusu O, Halcu I, Grigoriu O et al (2013) Converting unstructured and semi-structured data into knowledge. In: 013 11th RoEduNet international conference, 1-4, https:// doi. org/ 10. 1109/ RoEdu Net. 2013. 65117 36
  116. Sammut C, Banerji RB (1986) Learning concepts by asking questions. In: Michalski RS, Carbonell J, Mitchell T (eds) Machine learning: an artificial intelligence approach, vol 2. Morgan Kaufmann, Burlington, pp 167-192
  117. Šavelka J, Trivedi G, Ashley KD (2015) Applying an interactive machine learning approach to statutory analysis. In: Rotolo A (ed) Legal knowledge and information systems, frontiers in artificial intel- ligence and applications, vol 279. IOS Press, Amsterdam, pp 101-110. https:// doi. org/ 10. 3233/ 978-1-61499-609-5-101
  118. Selvaraju RR, Cogswell M, Das A et al (2017) Grad-cam: Visual explanations from deep networks via gra- dient-based localization. In: 2017 IEEE international conference on computer vision (ICCV), 618- 626, https:// doi. org/ 10. 1109/ ICCV. 2017. 74, https:// ieeex plore. ieee. org/ docum ent/ 82373 36
  119. Sena A, Howard M (2020) Quantifying teaching behavior in robot learning from demonstration. Int J Robot Res 39(1):54-72. https:// doi. org/ 10. 1177/ 02783 64919 884623
  120. Sena A, Zhao Y, Howard MJ (2018) Teaching human teachers to teach robot learners. In: 2018 IEEE inter- national conference on robotics and automation (ICRA), 5675-5681, https:// doi. org/ 10. 1109/ ICRA. 2018. 84611 94
  121. Settles B (2009) Active learning literature survey. Tech. rep., University of Wisconsin-Madison. Department of Computer Sciences, https:// minds. wisco nsin. edu/ handle/ 1793/ 60660
  122. Settles B (2011) From theories to queries: Active learning in practice. In: Guyon I, Cawley G, Dror G et al (eds) Active learning and experimental design workshop In conjunction with AISTATS 2010, pro- ceedings of machine learning research, vol 16. JMLR workshop and conference proceedings, Sar- dinia, Italy, 1-18, http:// proce edings. mlr. press/ v16/ settl es11a. html
  123. Shneiderman B (2020) Human-centered artificial intelligence: reliable, safe & trustworthy. Int J Hum Com- put Interact 36(6):495-504. https:// doi. org/ 10. 1080/ 10447 318. 2020. 17411 18
  124. Simard PY, Amershi S, Chickering DM et al (2017) Machine teaching: A new paradigm for building machine learning systems. arXiv e-prints arxiv: 1707. 06742
  125. Singla A, Bogunovic I, Bartók G et al (2014) Near-optimally teaching the crowd to classify. In: Xing EP, Jebara T (eds) Proceedings of the 31st international conference on machine learning. PMLR, Bejing, China, proceedings of machine learning research, 154-162, http:// proce edings. mlr. press/ v32/ singl a14. pdf
  126. Sint R, Schaffert S, Stroka S et al (2009) Combining unstructured, fully structured and semi-structured information in semantic wikis. In: 4th semantic wiki workshop (SemWiki 2009) at the 6th Euro- pean semantic web conference (ESWC 2009), Hersonissos, Greece, 73-87, http:// ceur-ws. org/ Vol- 464/ paper-14. pdf
  127. Smith JS, Nebgen B, Lubbers N et al (2018) Less is more: sampling chemical space with active learning. J Chem Phys 148(24):241,733
  128. Soviany P, Ardei C, Ionescu RT et al (2020) Image difficulty curriculum for generative adversarial net- works (cugan). In: 2020 IEEE winter conference on applications of computer vision (WACV), 3452-3461, https:// doi. org/ 10. 1109/ WACV4 5572. 2020. 90934 08
  129. Spitkovsky VI, Alshawi H, Jurafsky D (2010) From baby steps to leapfrog: How "less is more" in unsu- pervised dependency parsing. In: Human language technologies: the 2010 annual conference of the North American Chapter of the Association for Computational Linguistics. Association for Computational Linguistics, Los Angeles, California, 751-759, https:// www. aclweb. org/ antho logy/ N10-1116
  130. Suh J, Ghorashi S, Ramos G et al (2019) Anchorviz: facilitating semantic data exploration and con- cept discovery for interactive machine learning. ACM Trans Interact Intell Syst. https:// doi. org/ 10. 1145/ 32413 79
  131. Talbot J, Lee B, Kapoor A et al (2009) Ensemblematrix: Interactive visualization to support machine learning with multiple classifiers. In: Proceedings of the SIGCHI conference on human factors in computing systems. Association for Computing Machinery, New York, NY, USA, CHI '09, 1283- 1292, https:// doi. org/ 10. 1145/ 15187 01. 15188 95
  132. Tang YP, Li GX, Huang SJ (2019) ALiPy: Active learning in python. Tech. rep., Nanjing University of Aeronautics and Astronautics, https:// github. com/ NUAA-AL/ ALiPy, available as arXiv preprint arxiv: 1901. 03802
  133. Teso S, Kersting K (2019) Explanatory interactive machine learning. In: Proceedings of the 2019 AAAI/ ACM Conference on AI, Ethics, and Society. Association for Computing Machinery, New York, NY, USA, AIES '19, 239-245, https:// doi. org/ 10. 1145/ 33066 18. 33142 93
  134. Tolls V (2018) An event-based approach to modeling complex data in critical care. PhD thesis, Queen's University (Canada), https:// qspace. libra ry. queen su. ca/ bitst ream/ handle/ 1974/ 24489/ Tolls_ Victo ria_J_ 201809_ MSC. pdf
  135. Tomczak K, Czerwińska P, Wiznerowicz M (2015) The cancer genome atlas (TCGA): an immeasurable source of knowledge. Contemp Oncol 19(1A):68-77. https:// doi. org/ 10. 5114/ wo. 2014. 47136
  136. Treveil M, Omont N, Stenac C et al (2020) Introducing MLOps. O'Reilly Media, Sebastopol Traoré R, Caselles-Dupré ea (2019) Discorl: continual reinforcement learning via policy distillation. arXiv e-prints arxiv: 1907. 05855 [cs.LG]
  137. Valiant LG (1984) A theory of the learnable. Commun ACM 27(11):1134-1142. https:// doi. org/ 10. 1145/ 1968. 1972
  138. van Allen P (2018) Prototyping ways of prototyping AI. Interactions 25(6):46-51. https:// doi. org/ 10. 1145/ 32745 66
  139. von Ahn L, Dabbish L (2004) Labeling images with a computer game. In: Proceedings of the SIGCHI conference on human factors in computing systems. ACM, New York, NY, USA, CHI '04, 319- 326, https:// doi. org/ 10. 1145/ 985692. 985733
  140. von Ahn L, Blum M, Langford J (2004) Telling humans and computers apart automatically. Commun ACM 47(2):56-60. https:// doi. org/ 10. 1145/ 966389. 966390
  141. Visi FG, Tanaka A (2021) Interactive machine learning of musical gesture. In: Miranda ER (ed) Hand- book of artificial intelligence for music: foundations, advanced approaches, and developments for creativity. Springer, Cham, pp 771-798. https:// doi. org/ 10. 1007/ 978-3-030-72116-9_ 27
  142. Wall E, Ghorashi S, Ramos G (2019) Using expert patterns in assisted interactive machine learning: a study in machine teaching. In: Lamas D, Loizides F, Nacke L et al (eds) Human-computer interaction- INTERACT 2019. Springer, Berlin, pp 578-599. https:// doi. org/ 10. 1007/ 978-3-030-29387-1_ 34
  143. Wallace BC, Small K, Brodley CE et al (2012) Deploying an interactive machine learning system in an evidence-based practice center: Abstrackr. In: Proceedings of the 2nd ACM SIGHIT International Health Informatics Symposium. Association for Computing Machinery, New York, NY, USA, IHI '12, 819-824, https:// doi. org/ 10. 1145/ 21103 63. 21104 64
  144. Wang Y, Gan W, Yang J et al (2019) Dynamic curriculum learning for imbalanced data classification. In: 2019 IEEE/CVF international conference on computer vision (ICCV), 5016-5025, https:// doi. org/ 10. 1109/ ICCV. 2019. 00512
  145. Wang X, Pham H, Michel P et al (2020) Optimizing data usage via differentiable rewards. In: III HD, Singh A (eds) Proceedings of the 37th international conference on machine learning, proceedings of machine learning research, vol 119. PMLR, 9983-9995, https:// proce edings. mlr. press/ v119/ wang2 0p. html
  146. Wang X, Chen Y, Zhu W (2021) A survey on curriculum learning. IEEE Trans Pattern Anal Mach Intell. https:// doi. org/ 10. 1109/ TPAMI. 2021. 30699 08
  147. Ware M, Frank E, Holmes G et al (2001) Interactive machine learning: letting users build classifiers. Int J Hum Comput Stud 55(3):281-292. https:// doi. org/ 10. 1006/ ijhc. 2001. 0499
  148. Weimer M (2010) Machine teaching: a machine learning approach to technology enhanced learning. PhD thesis, Darmstadt University of Technology, http:// tupri nts. ulb. tu-darms tadt. de/ 2109/
  149. Weinshall D, Cohen G, Amir D (2018) Curriculum learning by transfer learning: Theory and experiments with deep networks. In: Proceedings of the 35th annual international conference on machine learning, 5235-5243, http:// proce edings. mlr. press/ v80/ weins hall1 8a. html
  150. Weiss K, Khoshgoftaar TM, Wang D (2016) A survey of transfer learning. J Big data 3(1):1-40. https:// doi. org/ 10. 1186/ s40537-016-0043-6
  151. Weitekamp D, Harpstead E, Koedinger KR (2020) An interaction design for machine teaching to develop AI tutors. In: Proceedings of the 2020 CHI conference on human factors in computing systems, 1-11, https:// doi. org/ 10. 1145/ 33138 31. 33762 26
  152. Wing JM (2021) Trustworthy AI. Commun ACM 64(10):64-71. https:// doi. org/ 10. 1145/ 34482 48
  153. Wong JS (2018) Design and fiction: imagining civic AI. Interactions 25(6):42-45. https:// doi. org/ 10. 1145/ 32745 68
  154. Xu W (2019) Toward human-centered AI: a perspective from human-computer interaction. Interactions 26(4):42-46. https:// doi. org/ 10. 1145/ 33284 85
  155. Xu B, Zhang L, Mao Z et al (2020) Curriculum learning for natural language understanding. In: Proceed- ings of the 58th annual meeting of the Association for Computational Linguistics. Association for Computational Linguistics, Online, 6095-6104, https:// doi. org/ 10. 18653/ v1/ 2020. acl-main. 542, https:// www. aclweb. org/ antho logy/ 2020. acl-main. 542
  156. Yang Q, Suh J, Chen NC et al (2018) Grounding interactive machine learning tool design in how non- experts actually build models. In: Proceedings of the 2018 designing interactive systems conference. Association for Computing Machinery, New York, NY, USA, DIS '18, 573-584, https:// doi. org/ 10. 1145/ 31967 09. 31967 29
  157. Yang SJ, Ogata H, Matsui T et al (2021) Human-centered artificial intelligence in education: seeing the invisible through the visible. Comput Educ 2(100):008. https:// doi. org/ 10. 1016/j. caeai. 2021. 100008
  158. Zbyszynski M, Tanaka A, Visi F (2020) Interactive machine learning: strategies for live performance using electromyography. In: Silva H (ed) Open source biomedical engineering. Springer, Berlin
  159. Zeiler MD, Fergus R (2014) Visualizing and understanding convolutional networks. In: Fleet D, Pajdla T, Schiele B et al (eds) European conference on computer vision. Springer, Cham, pp 818-833. https:// doi. org/ 10. 1007/ 978-3-319-10590-1_ 53
  160. Zeiler MD, Taylor GW, Fergus R (2011) Adaptive deconvolutional networks for mid and high level feature learning. In: 2011 International conference on computer vision, 2018-2025, https:// doi. org/ 10. 1109/ ICCV. 2011. 61264 74, https:// ieeex plore. ieee. org/ docum ent/ 61264 74
  161. Zhang X, Kumar G, Khayrallah H et al (2018) An empirical exploration of curriculum learning for neural machine translation. arXiv e-prints arxiv: 1811. 00739 [cs.CL]
  162. Zhang D, Han J, Guo G et al (2019a) Learning object detectors with semi-annotated weak labels. IEEE Trans Circuits Syst Video Technol 29(12):3622-3635. https:// doi. org/ 10. 1109/ TCSVT. 2018. 28841 73
  163. Zhang X, Shapiro P, Kumar G et al (2019b) Curriculum learning for domain adaptation in neural machine translation. In: Proceedings of the 2019 conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, Volume 1 (Long and Short Papers). Association for Computational Linguistics, Minneapolis, Minnesota, 1903-1915, https:// doi. org/ 10. 18653/ v1/ N19-1189, https:// www. aclweb. org/ antho logy/ N19-1189
  164. Zhao Y, Prosperi M, Lyu T et al (2020) Integrating crowdsourcing and active learning for classification of work-life events from tweets. In: Fujita H, Fournier-Viger P, Ali M et al (eds) Trends in artificial intel- ligence theory and applications. Artificial intelligence practices. Springer, Cham, pp 333-344. https:// doi. org/ 10. 1007/ 978-3-030-55789-8_ 30
  165. Zhou B, Khosla A, Lapedriza A et al (2016) Learning deep features for discriminative localization. In: 2016 IEEE conference on computer vision and pattern recognition (CVPR), 2921-2929, https:// doi. org/ 10. 1109/ CVPR. 2016. 319, https:// ieeex plore. ieee. org/ docum ent/ 77806 88
  166. Zhou Y, Yang B, Wong DF et al (2020) Uncertainty-aware curriculum learning for neural machine transla- tion. In: Proceedings of the 58th annual meeting of the Association for Computational Linguistics. Association for Computational Linguistics, Online, 6934-6944, https:// doi. org/ 10. 18653/ v1/ 2020. acl- main. 620, https:// www. aclweb. org/ antho logy/ 2020. acl-main. 620
  167. Zhu X (2015) Machine teaching: An inverse problem to machine learning and an approach toward opti- mal education. In: Proceedings of the twenty-ninth AAAI conference on artificial intelligence. AAAI Press, AAAI'15, 4083-4087, https:// ojs. aaai. org/ index. php/ AAAI/ artic le/ view/ 9761
  168. Zhu X, Singla A, Zilles S et al (2018) An overview of machine teaching. arXiv e-prints arxiv: 1801. 05927
  169. Zhuang F, Qi Z, Duan K et al (2021) A comprehensive survey on transfer learning. Proc IEEE 109(1):43- 76. https:// doi. org/ 10. 1109/ JPROC. 2020. 30045 55

Related papers

Machine learning for human learners: opportunities, issues, tensions and threats
Jason Zagami

Educational Technology Research and Development

Machine learning systems are infiltrating our lives and are beginning to become important in our education systems. This article, developed from a synthesis and analysis of previous research, examines the implications of recent developments in machine learning for human learners and learning. In this article we first compare deep learning in computers and humans to examine their similarities and differences. Deep learning is identified as a sub-set of machine learning, which is itself a component of artificial intelligence. Deep learning often depends on backwards propagation in weighted neural networks, so is non-deterministic—the system adapts and changes through practical experience or training. This adaptive behaviour predicates the need for explainability and accountability in such systems. Accountability is the reverse of explainability. Explainability flows through the system from inputs to output (decision) whereas accountability flows backwards, from a decision to the perso...

View PDFchevron_right
Understanding the Relationship between Interactions and Outcomes in Human-in-the-Loop Machine Learning
Reid Simmons

2021

Human-in-the-loop Machine Learning (HIL-ML) is a widely adopted paradigm for instilling human knowledge in autonomous agents. Many design choices influence the efficiency and effectiveness of such interactive learning processes, particularly the interaction type through which the human teacher may provide feedback. While different interaction types (demonstrations, preferences, etc.) have been proposed and evaluated in the HIL-ML literature, there has been little discussion of how these compare or how they should be selected to best address a particular learning problem. In this survey, we propose an organizing principle for HIL-ML that provides a way to analyze the effects of interaction types on human performance and training data. We also identify open problems in understanding the effects of interaction types.

View PDFchevron_right
Applications, Challenges, and Future Directions of Human-in-the-Loop Learning
Deepanwita Datta

IEEE access, 2024

Machine learning (ML) has become a popular technique for various automation tasks in the era of Industry 4.0, such as the analysis and synthesis of visual data such as images and videos, natural language and speech, financial data, and biomedical applications. However, ML-based automation techniques are facing difficulties like decision-making, thus incorporating user expertise into the system might be advantageous. The goal of adding human domain expertise with ML-based automation is to provide more accurate prediction models. Human-in-the-loop (HITL) systems that integrate human expertise with ML algorithms are becoming more and more common in various industries. However, there are a number of methodological, technical, and ethical difficulties with the development and application of HITL systems. This paper aims to explore the methodologies, challenges, and opportunities associated with HITL systems implementations. We also discuss a number of issues that must be resolved for HITL systems to be effective, including data quality, bias, and user engagement. Besides, we also explored several approaches that can be utilized to enhance the performance of HITL systems, such as active learning (AL), iterative ML, and reinforcement learning, as well as the current state of the art in HITL systems. We also selectively highlighted the advantages of HITL systems, such as their potential to increase decision-making process accountability and transparency by utilizing human experience to improve ML decision-making capability. The paper will be very useful for researchers, practitioners, and policymakers. Recent studies have highlighted the importance of human

View PDFchevron_right
Explainable AI in Education: Fostering Human Oversight and Shared Responsibility
Nurkhamimi Zainuddin

European Digital Education Hub, 2025

Explainable artificial intelligence (XAI) is a sub-field of artificial intelligence (AI), which aims to provide explanations about the reasons why an AI-based system takes a decision or provides an output (TechDispatch, 2023). The search for meaningful explanations is not new in the field of AI, but it has been mainly a technical issue for developers who were looking for reliability in the results obtained by their AI systems, so they could be accepted by end users of specific areas (Ali et al, 2023). The great advance of AI technology in the last years has turned these systems into general-purpose digital tools, and new considerations have arisen in this realm.

View PDFchevron_right
Teaching the Machine to Explain Itself using Domain Knowledge
Pedro Bizarro

ArXiv, 2020

Machine Learning (ML) has been increasingly used to aid humans to make better and faster decisions. However, non-technical humans-in-the-loop struggle to comprehend the rationale behind model predictions, hindering trust in algorithmic decision-making systems. Considerable research work on AI explainability attempts to win back trust in AI systems by developing explanation methods but there is still no major breakthrough. At the same time, popular explanation methods (e.g., LIME, and SHAP) produce explanations that are very hard to understand for non-data scientist persona. To address this, we present JOEL, a neural network-based framework to jointly learn a decision-making task and associated explanations that convey domain knowledge. JOEL is tailored to human-in-the-loop domain experts that lack deep technical ML knowledge, providing high-level insights about the model&#39;s predictions that very much resemble the experts&#39; own reasoning. Moreover, we collect the domain feedbac...

View PDFchevron_right

Related topics

  • Cognitive Science
  • Artificial Intelligence

    • Cited by

    Explainable Image Classification: The Journey So Far and the Road Ahead
    Vidhya Kamakshi

    AI

    Explainable Artificial Intelligence (XAI) has emerged as a crucial research area to address the interpretability challenges posed by complex machine learning models. In this survey paper, we provide a comprehensive analysis of existing approaches in the field of XAI, focusing on the tradeoff between model accuracy and interpretability. Motivated by the need to address this tradeoff, we conduct an extensive review of the literature, presenting a multi-view taxonomy that offers a new perspective on XAI methodologies. We analyze various sub-categories of XAI methods, considering their strengths, weaknesses, and practical challenges. Moreover, we explore causal relationships in model explanations and discuss approaches dedicated to explaining cross-domain classifiers. The latter is particularly important in scenarios where training and test data are sampled from different distributions. Drawing insights from our analysis, we propose future research directions, including exploring explai...

    View PDFchevron_right
    Application of streaming analytics for Artificial Lift systems: a human-in-the-loop approach for analysing clustered time-series data from progressive cavity pumps
    Peter Behrenbruch

    Neural Computing and Applications, 2022

    Assessing real-time performance of Artificial Lift Pumps is a prevalent time-series problem to tackle for natural gas operators in Eastern Australia. Multiple physics, data-driven, and hybrid approaches have been investigated to analyse or predict pump performance. However, these methods present a challenge in running compute-heavy algorithms on streaming time-series data. As there is limited research on novel approaches to tackle multivariate time-series analytics for Artificial Lift systems, this paper introduces a human-in-the-loop approach, where petroleum engineers label clustered time-series data to aid in streaming analytics. We rely on our recently developed novel approach of converting streaming time-series data into heatmap images to assist with real-time pump performance analytics. During this study, we were able to automate the labelling of streaming time-series data, which helped petroleum and well surveillance engineers better manage Artificial Lift Pumps through machine learning supported exception-based surveillance. The streaming analytics system developed as part of this research used historical time-series data from three hundred and fifty-nine (359) coal seam gas wells. The developed method is currently used by two natural gas operators, where the operators can accurately detect ten (10) performance-related events and five (5) anomalous events. This paper serves a two-fold purpose; first, we describe a step-by-step methodology that readers can use to reproduce the clustering method for multivariate time-series data. Second, we demonstrate how a human-in-the-loop approach adds value to the proposed method and achieves real-world results.

    View PDFchevron_right
    580 California St., Suite 400
    San Francisco, CA, 94104
    © 2025 Academia. All rights reserved