Observation-Level Interaction with Clustering and Dimension Reduction Algorithms

Wenskovitch, John; North, Chris

doi:10.1145/3077257.3077259

Cited by 26 publications

(16 citation statements)

References 22 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…However, because of the design of the OLI clustering tasks in this study, the participants only wanted to consider the points specifically mentioned in the tasks. New interaction designs are needed that encompass both of these types of OLI tasks, such as the use of a second view for moved points or including explicit representations of cluster boundaries (Wenskovitch and North 2017).…”

Section: :29mentioning

confidence: 99%

Observation-Level and Parametric Interaction for High-Dimensional Data Analysis

Self

Dowling

Wenskovitch

et al. 2018

ACM Trans. Interact. Intell. Syst.

Self Cite

View full text Add to dashboard Cite

Exploring high-dimensional data is challenging. Dimension reduction algorithms, such as weighted multidimensional scaling, support data exploration by projecting datasets to two dimensions for visualization. These projections can be explored through parametric interaction, tweaking underlying parameterizations, and observation-level interaction, directly interacting with the points within the projection. In this article, we present the results of a controlled usability study determining the differences, advantages, and drawbacks among parametric interaction, observation-level interaction, and their combination. The study assesses both interaction technique effects on domain-specific high-dimensional data analyses performed by non-experts of statistical algorithms. This study is performed using Andromeda, a tool that enables both parametric and observation-level interaction to provide in-depth data exploration. The results indicate that the two forms of interaction serve different, but complementary, purposes in gaining insight through steerable dimension reduction algorithms. CCS Concepts: • Human-centered computing → Empirical studies in visualization;

show abstract

Section: :29mentioning

confidence: 99%

Observation-Level and Parametric Interaction for High-Dimensional Data Analysis

Self

Dowling

Wenskovitch

et al. 2018

ACM Trans. Interact. Intell. Syst.

Self Cite

View full text Add to dashboard Cite

show abstract

“…In total, for the systems we reviewed, there were nine papers with case studies and usage scenarios [5,17,20,21,22,25,32,34,48], ten user studies [1,2,6,10,11,19,22,26,30,32] and two observational studies [37,38], in addition to surveys, questionnaires and interviews (seven papers). Although a number of papers included some form of a controlled user study, it was however acknowledged that this type of evaluation is generally difficult to conduct due to the various potential confounding factors such as previous knowledge [32].…”

Section: Evaluation Methods and Metricsmentioning

confidence: 99%

Evaluation of Interactive Machine Learning Systems

Boukhelifa

Bezerianos

Lutton

2018

Human–Computer Interaction Series

View full text Add to dashboard Cite

The evaluation of interactive machine learning systems remains a difficult task. These systems learn from and adapt to the human, but at the same time, the human receives feedback and adapts to the system. Getting a clear understanding of these subtle mechanisms of co-operation and co-adaptation is challenging. In this chapter, we report on our experience in designing and evaluating various interactive machine learning applications from different domains. We argue for coupling two types of validation: algorithm-centered analysis, to study the computational behaviour of the system; and human-centered evaluation, to observe the utility and effectiveness of the application for end-users. We use a visual analytics application for guided search, built using an interactive evolutionary approach, as an exemplar of our work. We argue that human-centered design and evaluation complement algorithmic analysis, and can play an important role in addressing the "black-box" effect of machine learning. Finally, we discuss research opportunities that require human-computer interaction methodologies, in order to support both the visible and hidden roles that humans play in interactive machine learning.

show abstract

“…Observation-Level Interaction (OLI) [5] is one of the major types of HITL machine learning. There are two types of OLI: exploratory and expressive interactions [1].…”

Section: Prior Workmentioning

confidence: 99%

Hierarchical Data Analysis Using Discernibility Relation

So¹

2020

IJETT

View full text Add to dashboard Cite

Human-in-the-Loop (HITL) machine learning uses human feedback to improve performance of machine learning models. One of the focuses in HITL machine learning research is to explore ways to capture human feedback and transform feedback to useful information that can inform the learning process. This paper outlines a clustering method that is based on discernibility relation in rough set theory. This clustering method presents intermediate clustering results using indiscernibility definition graph. Human users can provide feedback by manipulating the cluster representatives that are presented in an indiscernibility definition graph. Discernibility relation offers a more intuitive understanding of clustering results when compared to distance-based relationship in terms of providing useful feedback to inform the clustering algorithm about its performance.

show abstract

Observation-Level Interaction with Clustering and Dimension Reduction Algorithms

Cited by 26 publications

References 22 publications

Observation-Level and Parametric Interaction for High-Dimensional Data Analysis

Observation-Level and Parametric Interaction for High-Dimensional Data Analysis

Evaluation of Interactive Machine Learning Systems

Hierarchical Data Analysis Using Discernibility Relation

Contact Info

Product

Resources

About