Skip to content

Latest commit

 

History

History
185 lines (155 loc) · 14.7 KB

regressions-neuclir22-ru-dt-splade.md

File metadata and controls

185 lines (155 loc) · 14.7 KB
Raw

Anserini Regressions: NeuCLIR22 — Russian (Document Translation)

This page presents document translation regression experiments for the TREC 2022 NeuCLIR Track, Russian, with the following configuration:

  • Queries: English
  • Documents: Machine-translated documents from Russian into English (corpus provided by the organizers)
  • Model: SPLADE CoCondenser SelfDistil

The exact configurations for these regressions are stored in this YAML file. Note that this page is automatically generated from this template as part of Anserini's regression pipeline, so do not modify this page directly; modify the template instead.

We make available a version of the corpus that has already been encoded with SPLADE CoCondenser SelfDistil, i.e., we performed model inference on every document and stored the output sparse vectors. Thus, no neural inference is required to reproduce these experiments; see instructions below.

From one of our Waterloo servers (e.g., orca), the following command will perform the complete regression, end to end:

python src/main/python/run_regression.py --index --verify --search --regression neuclir22-ru-dt-splade

Corpus Download

Download the corpus and unpack into collections/:

wget https://rgw.cs.uwaterloo.ca/pyserini/data/neuclir22-ru-en-splade.tar -P collections/
tar xvf collections/neuclir22-ru-en-splade.tar -C collections/

To confirm, neuclir22-ru-en-splade.tar is 5.5 GB and has MD5 checksum 828e44e1ab8af19e5e1224539abe347c. With the corpus downloaded, the following command will perform the remaining steps below:

python src/main/python/run_regression.py --index --verify --search --regression neuclir22-ru-dt-splade \
  --corpus-path collections/neuclir22-ru-en-splade

Indexing

Typical indexing command:

bin/run.sh io.anserini.index.IndexCollection \
  -collection JsonVectorCollection \
  -input /path/to/neuclir22-ru-en-splade \
  -generator DefaultLuceneDocumentGenerator \
  -index indexes/lucene-index.neuclir22-ru-en-splade \
  -threads 8 -impact -pretokenized -storeRaw \
  >& logs/log.neuclir22-ru-en-splade &

For additional details, see explanation of common indexing options.

Retrieval

After indexing has completed, you should be able to perform retrieval as follows:

bin/run.sh io.anserini.search.SearchCollection \
  -index indexes/lucene-index.neuclir22-ru-en-splade \
  -topics tools/topics-and-qrels/topics.neuclir22-en.splade.original-title.txt.gz \
  -topicReader TsvInt \
  -output runs/run.neuclir22-ru-en-splade.splade.topics.neuclir22-en.splade.original-title.txt \
  -impact -pretokenized &
bin/run.sh io.anserini.search.SearchCollection \
  -index indexes/lucene-index.neuclir22-ru-en-splade \
  -topics tools/topics-and-qrels/topics.neuclir22-en.splade.original-desc.txt.gz \
  -topicReader TsvInt \
  -output runs/run.neuclir22-ru-en-splade.splade.topics.neuclir22-en.splade.original-desc.txt \
  -impact -pretokenized &
bin/run.sh io.anserini.search.SearchCollection \
  -index indexes/lucene-index.neuclir22-ru-en-splade \
  -topics tools/topics-and-qrels/topics.neuclir22-en.splade.original-desc_title.txt.gz \
  -topicReader TsvInt \
  -output runs/run.neuclir22-ru-en-splade.splade.topics.neuclir22-en.splade.original-desc_title.txt \
  -impact -pretokenized &

bin/run.sh io.anserini.search.SearchCollection \
  -index indexes/lucene-index.neuclir22-ru-en-splade \
  -topics tools/topics-and-qrels/topics.neuclir22-en.splade.original-title.txt.gz \
  -topicReader TsvInt \
  -output runs/run.neuclir22-ru-en-splade.splade+rm3.topics.neuclir22-en.splade.original-title.txt \
  -impact -pretokenized -rm3 -collection JsonVectorCollection &
bin/run.sh io.anserini.search.SearchCollection \
  -index indexes/lucene-index.neuclir22-ru-en-splade \
  -topics tools/topics-and-qrels/topics.neuclir22-en.splade.original-desc.txt.gz \
  -topicReader TsvInt \
  -output runs/run.neuclir22-ru-en-splade.splade+rm3.topics.neuclir22-en.splade.original-desc.txt \
  -impact -pretokenized -rm3 -collection JsonVectorCollection &
bin/run.sh io.anserini.search.SearchCollection \
  -index indexes/lucene-index.neuclir22-ru-en-splade \
  -topics tools/topics-and-qrels/topics.neuclir22-en.splade.original-desc_title.txt.gz \
  -topicReader TsvInt \
  -output runs/run.neuclir22-ru-en-splade.splade+rm3.topics.neuclir22-en.splade.original-desc_title.txt \
  -impact -pretokenized -rm3 -collection JsonVectorCollection &

bin/run.sh io.anserini.search.SearchCollection \
  -index indexes/lucene-index.neuclir22-ru-en-splade \
  -topics tools/topics-and-qrels/topics.neuclir22-en.splade.original-title.txt.gz \
  -topicReader TsvInt \
  -output runs/run.neuclir22-ru-en-splade.splade+rocchio.topics.neuclir22-en.splade.original-title.txt \
  -impact -pretokenized -rocchio -collection JsonVectorCollection &
bin/run.sh io.anserini.search.SearchCollection \
  -index indexes/lucene-index.neuclir22-ru-en-splade \
  -topics tools/topics-and-qrels/topics.neuclir22-en.splade.original-desc.txt.gz \
  -topicReader TsvInt \
  -output runs/run.neuclir22-ru-en-splade.splade+rocchio.topics.neuclir22-en.splade.original-desc.txt \
  -impact -pretokenized -rocchio -collection JsonVectorCollection &
bin/run.sh io.anserini.search.SearchCollection \
  -index indexes/lucene-index.neuclir22-ru-en-splade \
  -topics tools/topics-and-qrels/topics.neuclir22-en.splade.original-desc_title.txt.gz \
  -topicReader TsvInt \
  -output runs/run.neuclir22-ru-en-splade.splade+rocchio.topics.neuclir22-en.splade.original-desc_title.txt \
  -impact -pretokenized -rocchio -collection JsonVectorCollection &

Evaluation can be performed using trec_eval:

bin/trec_eval -c -m ndcg_cut.20 tools/topics-and-qrels/qrels.neuclir22-ru.txt runs/run.neuclir22-ru-en-splade.splade.topics.neuclir22-en.splade.original-title.txt
python -m pyserini.eval.trec_eval -c -m judged.20 tools/topics-and-qrels/qrels.neuclir22-ru.txt runs/run.neuclir22-ru-en-splade.splade.topics.neuclir22-en.splade.original-title.txt
bin/trec_eval -c -m recall.1000 tools/topics-and-qrels/qrels.neuclir22-ru.txt runs/run.neuclir22-ru-en-splade.splade.topics.neuclir22-en.splade.original-title.txt
bin/trec_eval -c -m map tools/topics-and-qrels/qrels.neuclir22-ru.txt runs/run.neuclir22-ru-en-splade.splade.topics.neuclir22-en.splade.original-title.txt
bin/trec_eval -c -m ndcg_cut.20 tools/topics-and-qrels/qrels.neuclir22-ru.txt runs/run.neuclir22-ru-en-splade.splade.topics.neuclir22-en.splade.original-desc.txt
python -m pyserini.eval.trec_eval -c -m judged.20 tools/topics-and-qrels/qrels.neuclir22-ru.txt runs/run.neuclir22-ru-en-splade.splade.topics.neuclir22-en.splade.original-desc.txt
bin/trec_eval -c -m recall.1000 tools/topics-and-qrels/qrels.neuclir22-ru.txt runs/run.neuclir22-ru-en-splade.splade.topics.neuclir22-en.splade.original-desc.txt
bin/trec_eval -c -m map tools/topics-and-qrels/qrels.neuclir22-ru.txt runs/run.neuclir22-ru-en-splade.splade.topics.neuclir22-en.splade.original-desc.txt
bin/trec_eval -c -m ndcg_cut.20 tools/topics-and-qrels/qrels.neuclir22-ru.txt runs/run.neuclir22-ru-en-splade.splade.topics.neuclir22-en.splade.original-desc_title.txt
python -m pyserini.eval.trec_eval -c -m judged.20 tools/topics-and-qrels/qrels.neuclir22-ru.txt runs/run.neuclir22-ru-en-splade.splade.topics.neuclir22-en.splade.original-desc_title.txt
bin/trec_eval -c -m recall.1000 tools/topics-and-qrels/qrels.neuclir22-ru.txt runs/run.neuclir22-ru-en-splade.splade.topics.neuclir22-en.splade.original-desc_title.txt
bin/trec_eval -c -m map tools/topics-and-qrels/qrels.neuclir22-ru.txt runs/run.neuclir22-ru-en-splade.splade.topics.neuclir22-en.splade.original-desc_title.txt

bin/trec_eval -c -m ndcg_cut.20 tools/topics-and-qrels/qrels.neuclir22-ru.txt runs/run.neuclir22-ru-en-splade.splade+rm3.topics.neuclir22-en.splade.original-title.txt
python -m pyserini.eval.trec_eval -c -m judged.20 tools/topics-and-qrels/qrels.neuclir22-ru.txt runs/run.neuclir22-ru-en-splade.splade+rm3.topics.neuclir22-en.splade.original-title.txt
bin/trec_eval -c -m recall.1000 tools/topics-and-qrels/qrels.neuclir22-ru.txt runs/run.neuclir22-ru-en-splade.splade+rm3.topics.neuclir22-en.splade.original-title.txt
bin/trec_eval -c -m map tools/topics-and-qrels/qrels.neuclir22-ru.txt runs/run.neuclir22-ru-en-splade.splade+rm3.topics.neuclir22-en.splade.original-title.txt
bin/trec_eval -c -m ndcg_cut.20 tools/topics-and-qrels/qrels.neuclir22-ru.txt runs/run.neuclir22-ru-en-splade.splade+rm3.topics.neuclir22-en.splade.original-desc.txt
python -m pyserini.eval.trec_eval -c -m judged.20 tools/topics-and-qrels/qrels.neuclir22-ru.txt runs/run.neuclir22-ru-en-splade.splade+rm3.topics.neuclir22-en.splade.original-desc.txt
bin/trec_eval -c -m recall.1000 tools/topics-and-qrels/qrels.neuclir22-ru.txt runs/run.neuclir22-ru-en-splade.splade+rm3.topics.neuclir22-en.splade.original-desc.txt
bin/trec_eval -c -m map tools/topics-and-qrels/qrels.neuclir22-ru.txt runs/run.neuclir22-ru-en-splade.splade+rm3.topics.neuclir22-en.splade.original-desc.txt
bin/trec_eval -c -m ndcg_cut.20 tools/topics-and-qrels/qrels.neuclir22-ru.txt runs/run.neuclir22-ru-en-splade.splade+rm3.topics.neuclir22-en.splade.original-desc_title.txt
python -m pyserini.eval.trec_eval -c -m judged.20 tools/topics-and-qrels/qrels.neuclir22-ru.txt runs/run.neuclir22-ru-en-splade.splade+rm3.topics.neuclir22-en.splade.original-desc_title.txt
bin/trec_eval -c -m recall.1000 tools/topics-and-qrels/qrels.neuclir22-ru.txt runs/run.neuclir22-ru-en-splade.splade+rm3.topics.neuclir22-en.splade.original-desc_title.txt
bin/trec_eval -c -m map tools/topics-and-qrels/qrels.neuclir22-ru.txt runs/run.neuclir22-ru-en-splade.splade+rm3.topics.neuclir22-en.splade.original-desc_title.txt

bin/trec_eval -c -m ndcg_cut.20 tools/topics-and-qrels/qrels.neuclir22-ru.txt runs/run.neuclir22-ru-en-splade.splade+rocchio.topics.neuclir22-en.splade.original-title.txt
python -m pyserini.eval.trec_eval -c -m judged.20 tools/topics-and-qrels/qrels.neuclir22-ru.txt runs/run.neuclir22-ru-en-splade.splade+rocchio.topics.neuclir22-en.splade.original-title.txt
bin/trec_eval -c -m recall.1000 tools/topics-and-qrels/qrels.neuclir22-ru.txt runs/run.neuclir22-ru-en-splade.splade+rocchio.topics.neuclir22-en.splade.original-title.txt
bin/trec_eval -c -m map tools/topics-and-qrels/qrels.neuclir22-ru.txt runs/run.neuclir22-ru-en-splade.splade+rocchio.topics.neuclir22-en.splade.original-title.txt
bin/trec_eval -c -m ndcg_cut.20 tools/topics-and-qrels/qrels.neuclir22-ru.txt runs/run.neuclir22-ru-en-splade.splade+rocchio.topics.neuclir22-en.splade.original-desc.txt
python -m pyserini.eval.trec_eval -c -m judged.20 tools/topics-and-qrels/qrels.neuclir22-ru.txt runs/run.neuclir22-ru-en-splade.splade+rocchio.topics.neuclir22-en.splade.original-desc.txt
bin/trec_eval -c -m recall.1000 tools/topics-and-qrels/qrels.neuclir22-ru.txt runs/run.neuclir22-ru-en-splade.splade+rocchio.topics.neuclir22-en.splade.original-desc.txt
bin/trec_eval -c -m map tools/topics-and-qrels/qrels.neuclir22-ru.txt runs/run.neuclir22-ru-en-splade.splade+rocchio.topics.neuclir22-en.splade.original-desc.txt
bin/trec_eval -c -m ndcg_cut.20 tools/topics-and-qrels/qrels.neuclir22-ru.txt runs/run.neuclir22-ru-en-splade.splade+rocchio.topics.neuclir22-en.splade.original-desc_title.txt
python -m pyserini.eval.trec_eval -c -m judged.20 tools/topics-and-qrels/qrels.neuclir22-ru.txt runs/run.neuclir22-ru-en-splade.splade+rocchio.topics.neuclir22-en.splade.original-desc_title.txt
bin/trec_eval -c -m recall.1000 tools/topics-and-qrels/qrels.neuclir22-ru.txt runs/run.neuclir22-ru-en-splade.splade+rocchio.topics.neuclir22-en.splade.original-desc_title.txt
bin/trec_eval -c -m map tools/topics-and-qrels/qrels.neuclir22-ru.txt runs/run.neuclir22-ru-en-splade.splade+rocchio.topics.neuclir22-en.splade.original-desc_title.txt

Effectiveness

With the above commands, you should be able to reproduce the following results:

MAP SPLADE +RM3 +Rocchio
NeuCLIR 2022 (Russian): title (original English queries) 0.3745 0.3151 0.3909
NeuCLIR 2022 (Russian): desc (original English queries) 0.3294 0.2717 0.3378
NeuCLIR 2022 (Russian): desc+title (original English queries) 0.3559 0.3026 0.3659
nDCG@20 SPLADE +RM3 +Rocchio
NeuCLIR 2022 (Russian): title (original English queries) 0.4865 0.4154 0.4836
NeuCLIR 2022 (Russian): desc (original English queries) 0.4193 0.3552 0.4243
NeuCLIR 2022 (Russian): desc+title (original English queries) 0.4573 0.4000 0.4604
J@20 SPLADE +RM3 +Rocchio
NeuCLIR 2022 (Russian): title (original English queries) 0.3741 0.3469 0.3816
NeuCLIR 2022 (Russian): desc (original English queries) 0.3702 0.3219 0.3680
NeuCLIR 2022 (Russian): desc+title (original English queries) 0.3838 0.3518 0.3846
Recall@1000 SPLADE +RM3 +Rocchio
NeuCLIR 2022 (Russian): title (original English queries) 0.8538 0.8169 0.8686
NeuCLIR 2022 (Russian): desc (original English queries) 0.8376 0.7529 0.8238
NeuCLIR 2022 (Russian): desc+title (original English queries) 0.8513 0.7704 0.8544

Reproduction Log*

To add to this reproduction log, modify this template and run bin/build.sh to rebuild the documentation.