You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
@inproceedings{petri2013exploring,
title={Exploring the magic of WAND},
author={Petri, Matthias and Culpepper, J Shane and Moffat, Alistair},
booktitle={Proceedings of the 18th Australasian Document Computing Symposium},
pages={58--65},
year={2013}
}
I believe if you're using inverted index and token - docs list, using the WAND Top-K Retrieval Algorithm can speedup retrieval for small K in large documents. I'm not sure whether it's relevant to this project. I've once implemented this https://raw.githubusercontent.com/hockyy/ir-pa-2/main/bsbi.py
The text was updated successfully, but these errors were encountered:
I have an implementation of another optimization algorithm for top-k retrieval on my local branch. Unfortunately, it slows down the retrieval because (I suspect) it requires more instructions to be executed (even if they are applied to less data).
Current implementation heavily relies on vector computations, which are fairly optimized on modern CPUs.
I will let you know if WAND improves efficiency over the current implementation.
AmenRa
changed the title
[Question/Feature Request] Use WAND Top-K Retrieval
[Feature Request] Use WAND Top-K Retrieval
Dec 22, 2022
Hi, I have a working WAND implementation, but it is slower than brute force vector operations.
I am now considering more advanced WAND-based approaches. I hope to add one soon.
Unfortunately, I don't think this will happen anytime soon. The lexical retriever is already reasonably efficient, and there are other things I prefer to prioritize.
I believe if you're using inverted index and token - docs list, using the WAND Top-K Retrieval Algorithm can speedup retrieval for small K in large documents. I'm not sure whether it's relevant to this project. I've once implemented this https://raw.githubusercontent.com/hockyy/ir-pa-2/main/bsbi.py
The text was updated successfully, but these errors were encountered: