Lmst

#TREC2022DL

If you are at #TREC2022, please join us for the Deep Learning Track session which just started! 🤗 #TREC2022DL #TRECDL #TREC2023DL

Quick follow-up: Here's our poster summarizing the TREC 2022 Deep Learning track. #TREC2022DL #TREC2022 #TRECDL

An academic poster titled "Overview of the TREC 2022 Deep Learning Track".

Was a pleasure to present the 4th Deep Learning Track overview at #TREC2022. We had healthy track participation and observed bigger gaps between trad. and NN based methods, and better distribution of system performances across queries. #TREC2022DL #TRECDL

Track homepage: https://microsoft.github.io/msmarco/TREC-Deep-Learning

If you are attending #TREC2022, we hope to see you at the Deep Learning Track session this Wednesday!

Slide titled "TREC Deep Learning: Ranking in the large-data regime" describing the goal and history of the track and the focus for the 2022 edition of the track.

Slide presenting the participant statistics for the track this year.

Plots comparing traditional IR based methods ("trad") and deep learning based methods ("nnlm" and "nn") for both the passage and document ranking tasks for TREC 2022 Deep Learning Track.

Couple of boxplots showing distribution of system performances for each of the test queries in the TREC 2022 Deep Learning Track.

Client Info

Server: https://mastodon.social

Version: 2025.07

Repository: https://github.com/cyevgeniy/lmst