This repository accompanies the ACM TOIS paper "What can I cook with these ingredients?" - Understanding cooking-related information needs in conversational search

In this repository you find data that has been gathered when conducting in-situ experiments in a conversational cooking setting. These data include transcripts,annotations and code to analyse these data.

When you are here while reading our paper, you might have following questions:

Where can I find the transcripts?

The raw transcripts are located in the transcripts folder. All experiments were conducted in Germany. Therefore, we provide the original german version as well as an english version transcribed with DeepL. Since we put much more effort in the transcription and annotation process compared to our workshop paper (=pilot study), we also provide the annotated data from that paper (see folder transcripts/workshop_paper).

Where is the information need taxonomy?

A visual representation can be found in the annotation folder. InformationNeedTaxonomy.svg is the file you are looking for.

Where is the codebook?

A description of all information needs as well as examples for them are located in the annotation folder. annotation_schema_cookversational_search_german.xlsx is the german version, annotation_schema_cookversational_search.xlsx is the english one. The codebook from our pilot study is located in this folder, too.

Where is the code for the machine learning experiments?

You find code for both the baseline models, the BERT based models and the results in the experiments folder.

Is there a notebook that contains the statistical analyses?

Yes! Statistical Analysis.ipynb in the experimentsfolder is the relevant file.

Where is the CookversationalSearch dataset?

The fully annotated dataset with all information needs, german and english turns (translated with DeepL) is here: annotation/corpus/cookversational_search_dataset