Visual analytics for machine learning

PN 6-4 (II)

Project description

This project addresses the fundamental research problem of visualization for machine learning (Vis4ML). Our goal is to open the black box of machine learning (ML) to human users, making ML models more transparent, controllable, and modifiable through visual analytics. In particular, we will consider the following research directions to address this objective. First, we plan to employ surrogates as vehicles to reduce the complexity of the full ML models and make meaningful information accessible to humans. Second, we will develop constraint visualizations to integrate physical constraints in Vis4ML. Third, we will investigate language-based and guided user interaction to ease the exploration of ML models. Finally, one research direction is to harness the power of ML to further improve Vis4ML. In a collaborative effort with other SimTech members, we will explore several application scenarios for visual analytics for ML.

Project information

Project title Visual analytics for machine learning
Project leaders Daniel Weiskopf (Ngoc Thang Vu)
Project staff Sebastian Künzel, doctoral researcher
Project duration February 2023 - December 2025
Project number PN 6-4 (II)

Publications PN 6-4 and PN 6-4 (II)

  1. 2024

    1. T. Munz-Körner and D. Weiskopf, “Exploring visual quality of multidimensional time series projections,” Visual Informatics, 2024, doi: https://doi.org/10.1016/j.visinf.2024.04.004.

  2. 2023

    1. N. Schäfer et al., “Visual Analysis of Scene-Graph-Based Visual Question Answering,” in Proceedings of the 16th International Symposium on Visual Information Communication and Interaction, in VINCI ’23. Guangzhou, China: Association for Computing Machinery, Oct. 2023, pp. 1–8. doi: 10.1145/3615522.3615547.

  3. 2022

    1. T. Munz, D. Väth, P. Kuznecov, N. T. Vu, and D. Weiskopf, “Visualization-based improvement of neural machine translation,” Computers & Graphics, vol. 103, pp. 45–60, 2022, doi: https://doi.org/10.1016/j.cag.2021.12.003.

  4. 2021

    1. T. Munz, D. Väth, P. Kuznecov, T. Vu, and D. Weiskopf, “Visual-Interactive Neural Machine Translation,” in Graphics Interface 2021, 2021. [Online]. Available: https://openreview.net/forum?id=DQHaCvN9xd
    2. R. Garcia, T. Munz, and D. Weiskopf, “Visual analytics tool for the interpretation of hidden states in recurrent neural networks,” Visual Computing for Industry, Biomedicine, and Art, vol. 4, Art. no. 24, Sep. 2021, doi: 10.1186/s42492-021-00090-0.

  5. 2020

    1. T. Munz, N. Schaefer, T. Blascheck, K. Kurzhals, E. Zhang, and D. Weiskopf, “Demo of a Visual Gaze Analysis System for Virtual Board Games,” in ACM Symposium on Eye Tracking Research and Applications, in ETRA ’20 Adjunct. Stuttgart, Germany: Association for Computing Machinery, 2020. doi: 10.1145/3379157.3391985.
    2. F. Heyen et al., “ClaVis: An Interactive Visual Comparison System for Classifiers,” in Proceedings of the International Conference on Advanced Visual Interfaces, in AVI ’20. Salerno, Italy: Association for Computing Machinery, 2020. doi: 10.1145/3399715.3399814.
    3. T. Munz, N. Schäfer, T. Blascheck, K. Kurzhals, E. Zhang, and D. Weiskopf, “Comparative Visual Gaze Analysis for Virtual Board Games,” The 13th International Symposium on Visual Information Communication and Interaction (VINCI 2020), 2020, doi: 10.1145/3430036.3430038.

  6. 2019

    1. T. Munz, M. Burch, T. van Benthem, Y. Poels, F. Beck, and D. Weiskopf, “Overlap-Free Drawing of Generalized Pythagoras Trees for Hierarchy Visualization,” in 2019 IEEE Visualization Conference (VIS), Oct. 2019, pp. 251–255. doi: 10.1109/VISUAL.2019.8933606.
    2. T. Munz, L. L. Chuang, S. Pannasch, and D. Weiskopf, “VisME: Visual microsaccades explorer,” Journal of Eye Movement Research, vol. 12, Art. no. 6, Dec. 2019, doi: 10.16910/jemr.12.6.5.

Data and software publications PN 6-4 and PN 6-4 (II)

  1. T. Munz-Körner and D. Weiskopf, “Visual Analysis System to Explore the Visual Quality of Multidimensional Time Series Projections.” DaRUS, 2024. doi: 10.18419/DARUS-3553.
  2. T. Munz-Körner and D. Weiskopf, “Supplemental Material for “Exploring Visual Quality of Multidimensional Time Series Projections.”” DaRUS, 2024. doi: 10.18419/DARUS-3965.
  3. T. Munz-Körner, S. Künzel, and D. Weiskopf, “Supplemental Material for “Visual-Explainable AI: The Use Case of Language Models.”” 2023. doi: 10.18419/darus-3456.
  4. N. Schäfer et al., “Model Parameters and Evaluation Data for our Visual Analysis System for Scene-Graph-Based Visual Question Answering.” 2023. doi: 10.18419/darus-3597.
  5. N. Schäfer et al., “Visual Analysis System for Scene-Graph-Based Visual Question Answering.” 2023. doi: 10.18419/darus-3589.
  6. T. Munz, D. Väth, P. Kuznecov, N. T. Vu, and D. Weiskopf, “NMTVis - Extended Neural Machine Translation Visualization System.” 2022. doi: 10.18419/darus-2124.
  7. T. Munz, D. Väth, P. Kuznecov, N. T. Vu, and D. Weiskopf, “NMTVis - Neural Machine Translation Visualization System.” 2021. doi: 10.18419/darus-1849.
  8. T. Munz, R. Garcia, and D. Weiskopf, “Visual Analytics System for Hidden States in Recurrent Neural Networks.” DaRUS, 2021. doi: 10.18419/DARUS-2052.
  9. T. Munz, D. Väth, P. Kuznecov, N. T. Vu, and D. Weiskopf, “NMTVis - Trained Models for our Visual Analytics System.” DaRUS, 2021. doi: 10.18419/DARUS-1850.
  10. T. Munz, N. Schäfer, T. Blascheck, K. Kurzhals, E. Zhang, and D. Weiskopf, “Supplemental Material for Comparative Visual Gaze Analysis for Virtual Board Games.” DaRUS, 2020. doi: 10.18419/DARUS-1130.
  11. T. Munz, “VisME software v1.2.” Zenodo, 2019. doi: 10.5281/ZENODO.3352236.
To the top of the page