2020
DOI: 10.1007/s13218-020-00632-3
|View full text |Cite
|
Sign up to set email alerts
|

eXplainable Cooperative Machine Learning with NOVA

Abstract: In the following article, we introduce a novel workflow, which we subsume under the term “explainable cooperative machine learning” and show its practical application in a data annotation and model training tool called NOVA. The main idea of our approach is to interactively incorporate the ‘human in the loop’ when training classification models from annotated data. In particular, NOVA offers a collaborative annotation backend where multiple annotators join their workforce. A main aspect is the possibility of a… Show more

Help me understand this report

Search citation statements

Order By: Relevance

Paper Sections

Select...
1
1
1
1

Citation Types

0
21
0

Year Published

2020
2020
2023
2023

Publication Types

Select...
5
2
1

Relationship

3
5

Authors

Journals

citations
Cited by 31 publications
(21 citation statements)
references
References 53 publications
0
21
0
Order By: Relevance
“…Tasks related to machine learning (ML) are handed over and executed by our open-source Social Signal Interpretation (SSI) framework [3]. Since SSI is primarily designed to build online recognition systems, a trained model can be directly used to detect social cues in real-time [4]. A typical ML pipeline starts by prepossessing data to input data for the learning algorithm, a step known as feature extraction.…”
Section: Nova Toolmentioning
confidence: 99%
See 1 more Smart Citation
“…Tasks related to machine learning (ML) are handed over and executed by our open-source Social Signal Interpretation (SSI) framework [3]. Since SSI is primarily designed to build online recognition systems, a trained model can be directly used to detect social cues in real-time [4]. A typical ML pipeline starts by prepossessing data to input data for the learning algorithm, a step known as feature extraction.…”
Section: Nova Toolmentioning
confidence: 99%
“…This way, the annotator can immediately see how well the prediction worked. To evaluate the efficiency of the integrated CML strategy, in our earlier work [4] we performed a simulation study on an audio-related labeling task. Following this approach, we were able to reduce the initial annotation labour of 9.4h to 5.9h, which is a reduction of 37.23%.…”
Section: Cooperative Machine Learningmentioning
confidence: 99%
“…Second, we plan to explore interaction approaches that involve the user in the process, e.g., by only showing local information when the user asks for it as we did in the context of cooperative annotation [13]. This could reduce cognitive load while increasing the user's attention to the local information when it is needed.…”
Section: Discussion and Future Workmentioning
confidence: 99%
“…Training Data Some approaches use feedback on explanations for revising and improving the corpus of available training data in order to inform the learning system. This is usually achieved with a human-in-the-loop, who inspects visual explanations for data instances, revises them if necessary, and thus adds more training data [6]. In terms of IML, this can be understood as obtaining prior information from an expert user (or from world knowledge) at the final hypothesis which is represented as human interaction.…”
Section: Deriving Knowledge From Explanationsmentioning
confidence: 99%