Towards better analysis of machine learning models: A visual analytics perspective

By Shixia Liu, Xiting Wang, Mengchen Liu, Jun Zhu. Visual Informatics, 2017

396 citations recorded by the source corpus.

Abstract Interactive model analysis, the process of understanding, diagnosing, and refining a machine learning model with the help of interactive visualization, is very important for users to efficiently solve real-world artificial intelligence and data mining problems. Dramatic advances in big data analytics have led to a wide variety of interactive model analysis tasks. In this paper, we present a comprehensive analysis and interpretation of this rapidly developing area. Specifically, we classify the relevant work into three categories: understanding, diagnosis, and refinement. Each category is exemplified by recent influential work. Possible future research opportunities are also explored and discussed.

Subfields: Artificial Intelligence, Computer Vision and Pattern Recognition, Signal Processing

View Paper: https://www.semanticscholar.org/paper/9bd71d6b87d9e5b8394dc6bc8bc2fca7b7deb083

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In Computer Vision and Pattern Recognition

In Signal Processing

Citations in the corpus

Manifold: A Model-Agnostic Framework for Interpretation and Diagnosis of Machine Learning Models (2018, background): “As discussed earlier, many existing visual analytics solutions focus on interpreting the internal working mechanism of a specific model type [12, 21, 25, 32, 37, 41].”
Explaining anomalies detected by autoencoders using Shapley Additive Explanations (2021, background): “…eﬀectiveness of the explanation for them, such experts are not always available and do not always know speciﬁcally what they want to investigate (in terms of which instances may be anomalous and where to start their investigation of a potentially anomalous instance) before using a prototype [34].”
Do Convolutional Neural Networks Learn Class Hierarchy? (2017, neutral): “[40]: understand, diagnose, and improve.”
Visual Interaction with Deep Learning Models through Collaborative Semantic Inference (2019, methodology): “While previous work has categorized and described the vast design space for visual interpretation of machine learning models [24,33,60, 61], the problem of co-designing models, visualizations, and interactions remains challenging.”
RuleMatrix: Visualizing and Understanding Classifiers with Rules (2018, methodology): “Visualization has been increasingly used to support the understanding, diagnosis and refinement of machine learning models [1, 22].”
The importance of interpretability and visualization in machine learning for applications in medicine and health care (2019, methodology): “Interpretation would lead here to the design of better models, which would be the ﬁnal goal described in [27], namely reﬁnement .”
Visual Analytics in Deep Learning: An Interrogative Survey for the Next Frontiers (2018, neutral): “Regarding deep neural networks, related surveys include a recent book chapter that only discusses visualization of deep neural networks related to the field of computer vision [22], an unpublished essay that proposes a preliminary taxonomy for visualization techniques [23], and a paper that focuses ”
Explainable Artificial Intelligence: a Systematic Review (2020, methodology): “Refinement The capacity of a method to guide experts in improving the performance/robustness of a model [90]”
Interpreting Black-Box Models: A Review on Explainable Artificial Intelligence (2023, methodology): “Examples of lossless visualisation methods include dimensionality reduction techniques [135, 136] (such as PCA or t-SNE), scatter plots, and heat maps.”