Workshop Program

The workshop on PDAV will be a half day event on Monday, October 14 in the morning, co-located with IEEE VIS 2024 in Tampa, Florida, USA.

Program

8:30 - 9:45 AM: Session 1

9:45 - 10:15 AM: Break

10:15 - 11:30 AM: Session 2

  • 10:15: Intoducing the Book on PDAV
    Jean-Daniel Fekete, Inria & Université Paris-Saclay, FR [Slides]
  • 10:35: Towards a Progressive Open Source Framework for SciVis and InfoVis [Paper] [Poster]
    Charles Gueunet, François Mazen
  • 10:50: Progressive Glimmer: Expanding Dimensionality in Multidimensional Scaling [Paper] [Poster]
    Marina Evers, David Hägele, Sören Döring, Daniel Weiskopf
  • 11:05:: Practical Challenges of Progressive Data Science in Healthcare [Paper] [Poster]
    Faisal Zaki Roshan, Abhishek Ahuja, Fateme Rajabiyazdi
  • 11:20: Q&A and Closing

Registration

The regular registration to the IEEE VIS 2024 Conference allows attending the workshop for free. Virtual participation costs $20 with the Virtual Participant Registration.

Keynote Presenters

Danyel Fisher

The Time for Progressive Visualization is Increasingly Now

We've been trying to make Progressive Analytics work for about 25 years, dating back at least from Hellerstein's CONTROL project in 1997. The computing world has changed since then: large-scale web services, mobile computing, cloud computing and the AI revolution have all moved the tradeoffs between computing, network transfer, and storage. This talk will argue that current trends and technologies make for a great fit for progressive analytics.

Bio

Danyel Fisher is a visualization researcher and independant consultant. He received his PhD from the University of California, Irvine. He carried out research in user experiences around data visualization at Microsoft Research; and built data analytics tools for DevOps at Honeycomb.io. He is passionate about understanding how technology can support sophisticated user needs.

Julien Tierny

A Progressive Perspective on Topological Data Analysis

Topological Data Analysis (TDA) is a recent area of computer science that focuses on discovering intrinsic structures hidden in data. Although most TDA algorithms have practicable complexities, their execution can still require a significant time for real-life datasets. Then, these algorithms can become a time bottleneck when they are integrated in large interactive systems. In this talk, I will present an adaptation, to the progressive setting, of two classical TDA algorithms, namely persistence diagram computation and barycenter evaluation. Extensive experiments on real-life datasets illustrate that this progressive strategy, in addition to the continuous visual feedback it provides, can even improve the overall computation time in certain cases. I will illustrate the utility of this approach in batch mode and interactive setups, where it respectively enables (1) the control of the execution time as well as (2) interactively updated previews of the topological features found in a dataset.

Bio

Julien Tierny received the Ph.D. degree in Computer Science from the University of Lille in 2008. From 2008 to 2010, he was a Fulbright researcher at the SCI Institute at the University of Utah. He is currently a CNRS research director, affiliated with Sorbonne University. His research expertise lies in topological methods for data analysis and visualization. In 2019, he was awarded a consolidator grant by the ERC. In 2023, he was awarded the "IMT Young Researcher Award" by the IMT and the French Academy of Science. He is also the founder and lead maintainer of the Topology ToolKit (TTK), an open source library for topological data analysis.

Barbara Hammer

Explaining incremental models

Incremental learning incorporates machine learning algorithms, which are capable of learning from continuous data streams using limited resources; hence it provides efficient tools for anytime model inference within progressive data science. Yet as many popular models are black box models in this context, it is an open challenge how to efficiently enhance such techniques with components, which enable a direct inspection by humans, i.e., explanations of incremental learners. Within the talk, I will present two novel technologies which transfer popular feature attribution methods to the incremental domain: incremental permutation feature importance (iPFI) as very efficient global black-box judgement of feature relevance at any point in time; incremental Shapley additive global importance (iSAGE) provides an alternative measure grounded in a formal game-theoretic approach. Both methods offer efficient possibilities to monitor variable importances and its changes in incremental of progressive tasks.

Bio

Barbara Hammer chairs the Machine Learning research group at the Research Institute for Cognitive Interaction Technology (CITEC) at Bielefeld University. After completing her doctorate at the University of Osnabrück in 1999, she was Professor of Theoretical Computer Science at Clausthal University of Technology and a visiting researcher in Bangalore, Paris, Padua and Pisa. Her areas of specialisation include trustworthy AI, lifelong machine learning, and the combination of symbolic and sub-symbolic representations. She is PI in the ERC Synergy Grant WaterFutures and in the DFG Transregio Contructing Explainability. Barbara Hammer has been active at IEEE CIS as member of chair of the Data Mining Technical committee and the Neural Networks Technical Committee. She has been elected as a review board member for Machine Learning of the German Research Foundation in 2024 and she represents computer science as a member of the selection committee for fellowships of the Alexander von Humboldt Foundation. She is member of the Scientific Directorate Schloss Dagstuhl. Further, she has been selected as member of Academia Europaea.

Supporters

Thanks to the transregional Collaborative Research Center "Quantitative Methods for Visual Computing" (SFB-TRR 161) for its support.

SBF