Skip to content

Latest commit

 

History

History
113 lines (86 loc) · 8.42 KB

regressions-msmarco-v2-passage-augmented-d2q-t5.md

File metadata and controls

113 lines (86 loc) · 8.42 KB

Anserini Regressions: MS MARCO (V2) Passage Ranking

Models: BM25 with doc2query-T5 expansions on augmented passages

This page describes regression experiments for passage ranking on the augmented version of the MS MARCO (V2) passage corpus using the dev queries, which is integrated into Anserini's regression testing framework. Here, we expand the augmented passage corpus with doc2query-T5.

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.

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 msmarco-v2-passage-augmented-d2q-t5

Indexing

Typical indexing command:

target/appassembler/bin/IndexCollection \
  -collection MsMarcoV2PassageCollection \
  -input /path/to/msmarco-v2-passage-augmented-d2q-t5 \
  -index indexes/lucene-index.msmarco-v2-passage-augmented-d2q-t5/ \
  -generator DefaultLuceneDocumentGenerator \
  -threads 18 -storePositions -storeDocvectors -storeRaw \
  >& logs/log.msmarco-v2-passage-augmented-d2q-t5 &

The directory /path/to/msmarco-v2-passage-augmented-d2q-t5/ should be a directory containing the compressed jsonl files that comprise the corpus.

For additional details, see explanation of common indexing options.

Retrieval

Topics and qrels are stored in src/main/resources/topics-and-qrels/. These regression experiments use the dev queries and the dev2 queries.

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

target/appassembler/bin/SearchCollection \
  -index indexes/lucene-index.msmarco-v2-passage-augmented-d2q-t5/ \
  -topics src/main/resources/topics-and-qrels/topics.msmarco-v2-passage.dev.txt \
  -topicreader TsvInt \
  -output runs/run.msmarco-v2-passage-augmented-d2q-t5.bm25-default.topics.msmarco-v2-passage.dev.txt \
  -bm25 &
target/appassembler/bin/SearchCollection \
  -index indexes/lucene-index.msmarco-v2-passage-augmented-d2q-t5/ \
  -topics src/main/resources/topics-and-qrels/topics.msmarco-v2-passage.dev2.txt \
  -topicreader TsvInt \
  -output runs/run.msmarco-v2-passage-augmented-d2q-t5.bm25-default.topics.msmarco-v2-passage.dev2.txt \
  -bm25 &

target/appassembler/bin/SearchCollection \
  -index indexes/lucene-index.msmarco-v2-passage-augmented-d2q-t5/ \
  -topics src/main/resources/topics-and-qrels/topics.msmarco-v2-passage.dev.txt \
  -topicreader TsvInt \
  -output runs/run.msmarco-v2-passage-augmented-d2q-t5.bm25-default+rm3.topics.msmarco-v2-passage.dev.txt \
  -bm25 -rm3 &
target/appassembler/bin/SearchCollection \
  -index indexes/lucene-index.msmarco-v2-passage-augmented-d2q-t5/ \
  -topics src/main/resources/topics-and-qrels/topics.msmarco-v2-passage.dev2.txt \
  -topicreader TsvInt \
  -output runs/run.msmarco-v2-passage-augmented-d2q-t5.bm25-default+rm3.topics.msmarco-v2-passage.dev2.txt \
  -bm25 -rm3 &

Evaluation can be performed using trec_eval:

tools/eval/trec_eval.9.0.4/trec_eval -c -m recall.100 src/main/resources/topics-and-qrels/qrels.msmarco-v2-passage.dev.txt runs/run.msmarco-v2-passage-augmented-d2q-t5.bm25-default.topics.msmarco-v2-passage.dev.txt
tools/eval/trec_eval.9.0.4/trec_eval -c -m recall.1000 src/main/resources/topics-and-qrels/qrels.msmarco-v2-passage.dev.txt runs/run.msmarco-v2-passage-augmented-d2q-t5.bm25-default.topics.msmarco-v2-passage.dev.txt
tools/eval/trec_eval.9.0.4/trec_eval -c -M 100 -m map -c -M 100 -m recip_rank src/main/resources/topics-and-qrels/qrels.msmarco-v2-passage.dev.txt runs/run.msmarco-v2-passage-augmented-d2q-t5.bm25-default.topics.msmarco-v2-passage.dev.txt
tools/eval/trec_eval.9.0.4/trec_eval -c -m recall.100 src/main/resources/topics-and-qrels/qrels.msmarco-v2-passage.dev2.txt runs/run.msmarco-v2-passage-augmented-d2q-t5.bm25-default.topics.msmarco-v2-passage.dev2.txt
tools/eval/trec_eval.9.0.4/trec_eval -c -m recall.1000 src/main/resources/topics-and-qrels/qrels.msmarco-v2-passage.dev2.txt runs/run.msmarco-v2-passage-augmented-d2q-t5.bm25-default.topics.msmarco-v2-passage.dev2.txt
tools/eval/trec_eval.9.0.4/trec_eval -c -M 100 -m map -c -M 100 -m recip_rank src/main/resources/topics-and-qrels/qrels.msmarco-v2-passage.dev2.txt runs/run.msmarco-v2-passage-augmented-d2q-t5.bm25-default.topics.msmarco-v2-passage.dev2.txt

tools/eval/trec_eval.9.0.4/trec_eval -c -m recall.100 src/main/resources/topics-and-qrels/qrels.msmarco-v2-passage.dev.txt runs/run.msmarco-v2-passage-augmented-d2q-t5.bm25-default+rm3.topics.msmarco-v2-passage.dev.txt
tools/eval/trec_eval.9.0.4/trec_eval -c -m recall.1000 src/main/resources/topics-and-qrels/qrels.msmarco-v2-passage.dev.txt runs/run.msmarco-v2-passage-augmented-d2q-t5.bm25-default+rm3.topics.msmarco-v2-passage.dev.txt
tools/eval/trec_eval.9.0.4/trec_eval -c -M 100 -m map -c -M 100 -m recip_rank src/main/resources/topics-and-qrels/qrels.msmarco-v2-passage.dev.txt runs/run.msmarco-v2-passage-augmented-d2q-t5.bm25-default+rm3.topics.msmarco-v2-passage.dev.txt
tools/eval/trec_eval.9.0.4/trec_eval -c -m recall.100 src/main/resources/topics-and-qrels/qrels.msmarco-v2-passage.dev2.txt runs/run.msmarco-v2-passage-augmented-d2q-t5.bm25-default+rm3.topics.msmarco-v2-passage.dev2.txt
tools/eval/trec_eval.9.0.4/trec_eval -c -m recall.1000 src/main/resources/topics-and-qrels/qrels.msmarco-v2-passage.dev2.txt runs/run.msmarco-v2-passage-augmented-d2q-t5.bm25-default+rm3.topics.msmarco-v2-passage.dev2.txt
tools/eval/trec_eval.9.0.4/trec_eval -c -M 100 -m map -c -M 100 -m recip_rank src/main/resources/topics-and-qrels/qrels.msmarco-v2-passage.dev2.txt runs/run.msmarco-v2-passage-augmented-d2q-t5.bm25-default+rm3.topics.msmarco-v2-passage.dev2.txt

Effectiveness

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

MAP@100 BM25 (default) +RM3
MS MARCO V2 Passage: Dev 0.1160 0.0870
MS MARCO V2 Passage: Dev2 0.1158 0.0879
MRR@100 BM25 (default) +RM3
MS MARCO V2 Passage: Dev 0.1172 0.0880
MS MARCO V2 Passage: Dev2 0.1170 0.0887
R@100 BM25 (default) +RM3
MS MARCO V2 Passage: Dev 0.5039 0.4742
MS MARCO V2 Passage: Dev2 0.5158 0.4905
R@1000 BM25 (default) +RM3
MS MARCO V2 Passage: Dev 0.7647 0.7625
MS MARCO V2 Passage: Dev2 0.7659 0.7649