Recap of 26th International Conference on Discovery Science
Oct 20, 2023
We are excited to share that two of our PhD students, Pavlin Poličar and Jaka Kokošar, presented their papers describing our lab's ongoing research into biomarker discovery and temporal visualizations at the recent Discovery Science 2023 conference held in Porto, Portugal. Their research is greatly contributing to the science behind the tools in Orange, especially in the field of visualizations, bioinformatics and survival analysis. As a team, we want to make our work accessible to everyone, so we invite you to check our open-access papers down below.
1. Refining Temporal Visualizations Using the Directional Coherence Loss
We've developed a simple technique to augment dimensionality reduction approaches such as t-SNE and UMAP that generate embeddings more suited to temporal data. Using publicly accessible data on the COVID-19 pandemic in Slovenia, we show that our approach can reveal interesting temporal patterns in the Slovenian healthcare system that could not be observed in the standard visualizations. Check out the open-access paper.
2. Gene Interactions in Survival Data Analysis: A Data-Driven Approach Using Restricted Mean Survival Time and Literature Mining
The vast landscape of survival-related gene interactions is complex. Our approach, which combines biology-inspired algorithms with in-depth literature mining, offers a streamlined and efficient way to interpret this data. This research contributes to an evolving toolbox for survival analysis in Orange. Check out the open-access paper.
We are thrilled to be a part of the Discovery Science community, where outstanding research is brought to light.