Understanding the Relationship between Interactions and Outcomes in Human-in-the-Loop Machine Learning

Cui, Yuchen; Koppol, Pallavi; Admoni, Henny; Niekum, Scott; Simmons, Reid; Steinfeld, Aaron; Fitzgerald, Tesca

doi:10.24963/ijcai.2021/599

Cited by 19 publications

(15 citation statements)

References 50 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…1 Supervised learning covers a broad range of model classes, ranging from vision transformers, large language models, and impactful application areas, like medical diagnostics [69] and criminal justice [99]. We consider other objectives, which may include reinforcement learning [23,30] or unsupervised learning [24,55], to be out-of-scope for this paper. We focus on methods more commonly deployed in practice, and omit the Bayesian analog for our pipeline, where experts can express preferences over the distribution of functions [56,97].…”

Section: Feedback-update Taxonomymentioning

confidence: 99%

“…Before deploying a machine learning (ML) model in high-stakes use cases, practitioners, who are responsible for developing and maintaining models, may solicit and incorporate feedback from experts [4,30,44]. Prior work has largely focused on incorporating feedback of technical experts (e.g., from ML engineers, data scientists) into models [2,90,92,107,114,116,124].…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Perspectives on Incorporating Expert Feedback into Model Updates

Chen¹,

Bhatt²,

Heidari³

et al. 2022

Preprint

View full text Add to dashboard Cite

Machine learning (ML) practitioners are increasingly tasked with developing models that are aligned with non-technical experts' values and goals. However, there has been insufficient consideration of how practitioners should translate domain expertise into ML updates. In this paper, we consider how to capture interactions between practitioners and experts systematically. We devise a taxonomy to match expert feedback types with practitioner updates. A practitioner may receive feedback from an expert at the observationor domain-level, and convert this feedback into updates to the dataset, loss function, or parameter space. We review existing work from ML and human-computer interaction to describe this feedback-update taxonomy, and highlight the insufficient consideration given to incorporating feedback from non-technical experts. We end with a set of open questions that naturally arise from our proposed taxonomy and subsequent survey.

show abstract

Section: Feedback-update Taxonomymentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Perspectives on Incorporating Expert Feedback into Model Updates

Chen¹,

Bhatt²,

Heidari³

et al. 2022

Preprint

View full text Add to dashboard Cite

show abstract

“…Human Feedback: This describes the feedback that the teacher will provide to the machine. [Cui et al, 2021] defined four broad types of feedback i.e. showing, categorizing, sorting and evaluating.…”

Section: Teaching Interfacementioning

confidence: 99%

“…Human-in-the-loop (HITL) ML: [Cui et al, 2021] surveyed HITL ML and how design choices affect interactive learning. In their framework, they define four types of interaction: Showing, Categorizing, Sorting and Evaluating.…”

Section: Ablation Studiesmentioning

confidence: 99%

A Framework for Interactive Knowledge-Aided Machine Teaching

Taneja¹,

Sikka²,

Goel³

2022

Preprint

View full text Add to dashboard Cite

Machine Teaching (MT) is an interactive process where humans train a machine learning model by playing the role of a teacher. The process of designing an MT system involves decisions that can impact both efficiency of human teachers and performance of machine learners. Previous research has proposed and evaluated specific MT systems but there is limited discussion on a general framework for designing them. We propose a framework for designing MT systems and also detail a system for the text classification problem as a specific instance. Our framework focuses on three components i.e. teaching interface, machine learner, and knowledge base; and their relations describe how each component can benefit the others. Our preliminary experiments show how MT systems can reduce both human teaching time and machine learner error rate.

show abstract

“…Finally, this paper has explored one method of interaction to enable a human teacher to provide corrections to the robot. However, in human-in-the-loop learning problems, the ideal interaction type is dependent on the teacher's role in the learning system, and the context in which the robot is used (Cui et al, 2021). For example, the teacher may not have time to correct every step of the robot's action, or may instead prefer to provide corrections only after the robot has tried and failed to complete a task.…”

Section: Open Questionsmentioning

confidence: 99%

Modeling and Learning Constraints for Creative Tool Use

2021

Self Cite

View full text Add to dashboard Cite

Improvisation is a hallmark of human creativity and serves a functional purpose in completing everyday tasks with novel resources. This is particularly exhibited in tool-using tasks: When the expected tool for a task is unavailable, humans often are able to replace the expected tool with an atypical one. As robots become more commonplace in human society, we will also expect them to become more skilled at using tools in order to accommodate unexpected variations of tool-using tasks. In order for robots to creatively adapt their use of tools to task variations in a manner similar to humans, they must identify tools that fulfill a set of task constraints that are essential to completing the task successfully yet are initially unknown to the robot. In this paper, we present a high-level process for tool improvisation (tool identification, evaluation, and adaptation), highlight the importance of tooltips in considering tool-task pairings, and describe a method of learning by correction in which the robot learns the constraints from feedback from a human teacher. We demonstrate the efficacy of the learning by correction method for both within-task and across-task transfer on a physical robot.

show abstract

Understanding the Relationship between Interactions and Outcomes in Human-in-the-Loop Machine Learning

Cited by 19 publications

References 50 publications

Perspectives on Incorporating Expert Feedback into Model Updates

Perspectives on Incorporating Expert Feedback into Model Updates

A Framework for Interactive Knowledge-Aided Machine Teaching

Modeling and Learning Constraints for Creative Tool Use

Contact Info

Product

Resources

About