Abstract:

Data Visualization is effectively used to explore and explain complex data. In the talk, I will present three visual tools for different data domains that can help to formulate hypotheses in early stages of data discovery:

1) Data often comes with categorical information that assigns each item to one or multiple sets. We developed UpSet to explore the set intersections. Interactively, a user can filter and recombine intersections and calculate statistical measures on them. I will introduce the visual mapping and in a short demo, I will demonstrate the core features of our system.

2) Acquisition of biological data has become fast and cost-efficient. As a result, a domain scientist spends a large share of time in analyzing the experimental data. For example, RNAseq allows to sequence RNA which can then be used to investigate alternative splicing events within different conditions (different patient groups, different human tissues). ‘Vials’ provides a visual analysis tool to investigate this complex data. I will introduce the biological data, explain it’s mapping to visual variables, and show a short demo of the interactive system.

3) The increased interest in neural networks and their common use imposes the question about the ‘How ?’ - How do they work? What do they capture?. I will present a visual tool, LSTMVis, that allows investigation of state changes on trained models (LSTMs) and can be one puzzle piece in the quest of white-boxing neuronal networks.

All three projects resulted in open source tools which are used by domain experts
to help them to generate hypothesis as a first step towards insight.


Speaker:

Hendrik Strobelt is a Research Scientist at the Visual AI Lab with IBM Research Cambridge, MA. Hendrik was postdoctoral researcher at Harvard SEAS and NYU School of Engineering. He received his PhD in Computer Science from University of Konstanz, Germany and his MSc (Diplom) from TU Dresden, Germany.