Back

Fossil-explorer.com: An efficient interactive approach to exploring fossil data

Wu, J.; Xu, H.-H.; Pan, Z.; Niu, Z.

2022-06-16 paleontology
10.1101/2022.06.12.495851 bioRxiv
Show abstract

Fossils today are increasingly being digitized and documented by multi-modal data obtained from visual data (i.e., photos and tomographic images), as well as text, locations, geological ages, and other chemical and physical measurements. Popular online websites such as PBDB and GBDB offer visual explorations of specimens localities, but they have limited multi-modal data visualization abilities and face challenges related to visual obscuration and insufficient interaction/exploration. Here, we present fossil-explorer.com, a continuously developing open-source online tool for assisting paleontologists with interactively exploring fossil collections. The tool is designed to address the issues of visual clutter, limited data types, and insufficient interactions. It is intuitive and endorsed by paleontologists. We have also quantitatively evaluated the tool by measuring the interaction scaling performance. The results show that it provides sublinear interaction performance and thus is able to deal efficiently with millions-level data. The current fossil-explorer.com demonstrates the Ordovician to Silurian graptolite fossil multimedia dataset, which is significant in global stratigraphy and shale gas exploration. The extended version also facilitates the use of Deepbone (http://deepbone.org), worlds most comprehensive database of vertebrate paleontology database. We developed the code for fossil-explorer.com to be open access and will continue to improve it.

Matching journals

The top 1 journal accounts for 50% of the predicted probability mass.

50% of probability mass above

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.