Information Retrieval based on INLs
- Targeted audience
- Angewandte Informatik Bachelor
- Angewandte Informatik Master
- Komedia Bachelor
- Komedia Master
- Advanced programming skills (desired) Machine Learning (desired) Knowledge about IR; NLP, deep learning (desired)
Nutritions labels (INLs) are everywhere, in food packages, water, softdrink, milk bottles, medicines packs, etc. Over some time we learned to associate some nutritions with healty and some with unhealty life. Based on these associations we are able to judge whether we want to consume or reject what we see. We argue that this can also be the case in terms of reading consuption. In a traditional IR (Information Retrieval) documents are returned as a result of a search request. We know only that the returned documents are relevant to the user query. Other charecteristics are not provided. Once we start spending some time on the results and going through them we realize that there are only few documents that are really useful for our purposes. Note that the final relevant document selection is user dependent and might vary between users. We argue that documents retrieved by a search engine have also INLs and knowing them in advance whould help users to increase their search success hits and satisfaction.
In this project the aim is to develop an IR system that, unlike traditional IR systems, retrieves documents by also considering INLs. The project aims to have up to 15 students with strong programming (e.g. Python) and machine learning backgrounds (e.g. IM exam). It is also important that the student is familiar with the concept of IR and search engines. Knowledge about Natural Language Processing (NLP) and Text Processing are also useful skills the student should bring.
In our praxis project we already implemented some INLs. They are all packaged as a browser plugin. For every document users looks at the plugin computes a set of INLs and displays them. We aim to extend this browser plugin with the search facilities.