Towards rapid interactive machine learning

Arendt, Dustin; Saldanha, Emily; Wesslen, Ryan; Volkova, Svitlana; Dou, Wenwen

doi:10.1145/3301275.3302280

Cited by 13 publications

(7 citation statements)

References 37 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Users can not process more than possibly a few dozen without being stuck in a tedious labeling process. Following the central idea of (visual) active learning [3,15,17], we choose representatives based on the trade-off between exploitation and exploration. On the exploitation side, PSEUDo includes the top-5 candidates with the highest similarity in the samples.…”

Section: Relevance Feedbackmentioning

confidence: 99%

PSEUDo: Interactive Pattern Search in Multivariate Time Series with Locality-Sensitive Hashing and Relevance Feedback

Kruyff²,

Jiao

et al. 2022

IEEE Trans. Visual. Comput. Graphics

View full text Add to dashboard Cite

show abstract

Section: Relevance Feedbackmentioning

confidence: 99%

PSEUDo: Interactive Pattern Search in Multivariate Time Series with Locality-Sensitive Hashing and Relevance Feedback

Kruyff²,

Jiao

et al. 2022

IEEE Trans. Visual. Comput. Graphics

View full text Add to dashboard Cite

show abstract

“…Dudley et al [17] describe a general approach to interface design for IML providing an overview of challenges and common guiding principles. Arendt et al [2] present an IML interface with model feedback after every interaction by updating the items shown for each class. The users can drag misplace data items to the appropriate class and, if needed, create a new one.…”

Section: Visual Active and Interactive Machine Learningmentioning

confidence: 99%

“…Classification of Data Items A classification of a given data item is performed recursively, similar to a decision tree. (1) Find the most similar neuron in the root SOM; (2) If the node has a child, perform the same action recursively on the child SOM; (3) If the SOM node has no child, classify the item as the predominant label of the respective cell, i.e., relevant or irrelevant; (4) If no label information is available for this node, use the next most similar cell with label information in that specific SOM.…”

Section: Som As Visual Classifiermentioning

confidence: 99%

FDive: Learning Relevance Models Using Pattern-based Similarity Measures

Dennig

Polk

Lin

et al. 2019

2019 IEEE Conference on Visual Analytics Science and Technology (VAST)

View full text Add to dashboard Cite

Figure 1: FDIVE learns to distinguish relevant from irrelevant data through an iteratively improving classification model by learning the best-fitting feature descriptor and distance function. (1) Users express their notion of relevance by labeling a set of query items, in this case, images. (2) These labels are used to rank all similarity measures by their ability to distinguish relevant from irrelevant data. (3) The system applies the selected similarity measure to learn a Self-Organizing Map (SOM)-based relevance model. Users explore and refine the model by supplying relevance labels in uncertain data regions, especially near the decision boundaries. ABSTRACTThe detection of interesting patterns in large high-dimensional datasets is difficult because of their dimensionality and pattern complexity. Therefore, analysts require automated support for the extraction of relevant patterns. In this paper, we present FDIVE, a visual active learning system that helps to create visually explorable relevance models, assisted by learning a pattern-based similarity. We use a small set of user-provided labels to rank similarity measures, consisting of feature descriptor and distance function combinations, by their ability to distinguish relevant from irrelevant data. Based on the best-ranked similarity measure, the system calculates an interactive Self-Organizing Map-based relevance model, which classifies data according to the cluster affiliation. It also automatically prompts further relevance feedback to improve its accuracy. Uncertain areas, especially near the decision boundaries, are highlighted and can be refined by the user. We evaluate our approach by comparison to state-of-the-art feature selection techniques and demonstrate the usefulness of our approach by a case study classifying electron microscopy images of brain cells. The results show that FDIVE enhances both the quality and understanding of relevance models and can thus lead to new insights for brain research.

show abstract

“…In most previous work, IML systems are designed based on classification algorithms [5], [18], [30], [31], [53]. Classification can be seen as the most straightforward ML formulation, and users can define target category labels and add their corresponding training samples.…”

Section: Introductionmentioning

confidence: 99%

Image-to-Text Translation for Interactive Image Recognition: A Comparative User Study with Non-expert Users

Kawabe,

Sugano

2024

Journal of Information Processing

View full text Add to dashboard Cite

Interactive machine learning (IML) allows users to build their custom machine learning models without expert knowledge. While most existing IML systems are designed with classification algorithms, they sometimes oversimplify the capabilities of machine learning algorithms and restrict the user's task definition. On the other hand, as recent large-scale language models have shown, natural language representation has the potential to enable more flexible and generic task descriptions. Models that take images as input and output text have the potential to represent a variety of tasks by providing appropriate text labels for training. However, the effect of introducing text labels to IML system design has never been investigated. In this work, we aim to investigate the difference between image-to-text translation and image classification for IML systems. Using our prototype systems, we conducted a comparative user study with non-expert users, where participants solved various tasks. Our results demonstrate the underlying difficulty for users in properly defining image recognition tasks while highlighting the potential and challenges of interactive image-to-text translation systems.

show abstract

Towards rapid interactive machine learning

Cited by 13 publications

References 37 publications

PSEUDo: Interactive Pattern Search in Multivariate Time Series with Locality-Sensitive Hashing and Relevance Feedback

PSEUDo: Interactive Pattern Search in Multivariate Time Series with Locality-Sensitive Hashing and Relevance Feedback

FDive: Learning Relevance Models Using Pattern-based Similarity Measures

Image-to-Text Translation for Interactive Image Recognition: A Comparative User Study with Non-expert Users

Contact Info

Product

Resources

About