|
Size: 586
Comment:
|
← Revision 5 as of 2022-06-17 15:32:49 ⇥
Size: 617
Comment:
|
| Deletions are marked like this. | Additions are marked like this. |
| Line 5: | Line 5: |
| Use the existing [[https://github.com/ad-freiburg/qlever|QLever]] [[https://github.com/ad-freiburg/qlever/blob/master/docs/sparql_plus_text.md|text search system]] of our chair as basis for a search system for medical reports. The final product should be a search system that, given a query, yields a corresponding image retrieved from a matching medical report, as well as the report itself. Additional features of interest are synonym support, handling negations, ... |
Our SPARQL engine [[https://github.com/ad-freiburg/qlever|QLever]] supports search in structured data, but also in text. The text search part is still underdeveloped, both regarding what is possible and the efficiency. It would be great for this project, as well as for other projects, to develop it further. Beyond that, there are also many other interesting and non-trivial aspects that can be considered, such as entity recognition and linking, synonym support, proper handling of negations, etc. |
Type: Master project
Your interests / skills: C++, Natural Language Processing, Search Systems
Our SPARQL engine QLever supports search in structured data, but also in text. The text search part is still underdeveloped, both regarding what is possible and the efficiency. It would be great for this project, as well as for other projects, to develop it further.
Beyond that, there are also many other interesting and non-trivial aspects that can be considered, such as entity recognition and linking, synonym support, proper handling of negations, etc.