search

DFKI-NLP / InterroLang

InterroLang: Exploring NLP Models and Datasets through Dialogue-based Explanations [EMNLP 2023 Findings]
https://arxiv.org/abs/2310.05592
5 stars 1 forks source link

[Operations] New operation requests from User Study #115

Open qiaw99 opened 1 year ago

qiaw99 commented 1 year ago
nfelnlp commented 1 year ago

For "include + similar", we have to use the temp_dataset here: https://github.com/nfelnlp/InterroLang/blob/00aa371f15e57c82d25b215acf82acfd5420d324/actions/nlu/similarity.py#L90C50-L90C50

qiaw99 commented 1 year ago

@nfelnlp For now, I cann't really imagine a solution that would be able to implement include + similar. The reason is as follows:

  1. include relies on temp_dataset.
  2. When you call similar operation, e.g. show me the most similar one compared to instance 15. The corresponding parsed text is: filter 15 and similar 1. After doing this, the temp_dataset is already changed, containing only one instance which is id 15.
  3. As a result, the previous filtered temp_dataset by include operation will be "removed".
qiaw99 commented 1 year ago

In other words, in order to run any include related operations, user couldn't call any operations related to id (filter id and xxx)