Skip to content

Latest commit

 

History

History
98 lines (71 loc) · 6.49 KB

regressions-dl20-doc-segmented.wp-ca.md

File metadata and controls

98 lines (71 loc) · 6.49 KB
Raw

Anserini Regressions: TREC 2020 Deep Learning Track (Document)

Models: various bag-of-words approaches on segmented documents using CompositeAnalyzer.

This page describes experiments, integrated into Anserini's regression testing framework, on the TREC 2020 Deep Learning Track document ranking task.

Note that the NIST relevance judgments provide far more relevant documents per topic, unlike the "sparse" judgments provided by Microsoft (these are sometimes called "dense" judgments to emphasize this contrast). For additional instructions on working with MS MARCO document collection, refer to this page.

  • Indexing Condition: each MS MARCO document is first segmented into passages, each passage is treated as a unit of indexing
  • Expansion Condition: none

In the passage (i.e., segment) indexing condition, we select the score of the highest-scoring passage from a document as the score for that document to produce a document ranking; this is known as the MaxP technique.

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.

Note that in November 2021 we discovered issues in our regression tests, documented here. As a result, we have had to rebuild all our regressions from the raw corpus. These new versions yield end-to-end scores that are slightly different, so if numbers reported in a paper do not exactly match the numbers here, this may be the reason.

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 dl20-doc-segmented.wp-ca

Indexing

Typical indexing command:

bin/run.sh io.anserini.index.IndexCollection \
  -threads 16 \
  -collection JsonCollection \
  -input /path/to/msmarco-doc-segmented \
  -generator DefaultLuceneDocumentGenerator \
  -index indexes/lucene-inverted.msmarco-v1-doc-segmented.wp-ca/ \
  -storePositions -storeDocvectors -storeRaw -analyzeWithHuggingFaceTokenizer bert-base-uncased -useCompositeAnalyzer \
  >& logs/log.msmarco-doc-segmented &

The directory /path/to/msmarco-doc-segmented/ should be a directory containing the segmented corpus in Anserini's jsonl format. See this page for how to prepare the corpus.

For additional details, see explanation of common indexing options.

Retrieval

Topics and qrels are stored here, which is linked to the Anserini repo as a submodule. The regression experiments here evaluate on the 45 topics for which NIST has provided judgments as part of the TREC 2020 Deep Learning Track. The original data can be found here.

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

bin/run.sh io.anserini.search.SearchCollection \
  -index indexes/lucene-inverted.msmarco-v1-doc-segmented.wp-ca/ \
  -topics tools/topics-and-qrels/topics.dl20.txt \
  -topicReader TsvInt \
  -output runs/run.msmarco-doc-segmented.bm25-default.topics.dl20.txt \
  -bm25 -hits 10000 -selectMaxPassage -selectMaxPassage.delimiter "#" -selectMaxPassage.hits 1000 -analyzeWithHuggingFaceTokenizer bert-base-uncased -useCompositeAnalyzer &

Evaluation can be performed using trec_eval:

bin/trec_eval -c -M 100 -m map tools/topics-and-qrels/qrels.dl20-doc.txt runs/run.msmarco-doc-segmented.bm25-default.topics.dl20.txt
bin/trec_eval -c -m ndcg_cut.10 tools/topics-and-qrels/qrels.dl20-doc.txt runs/run.msmarco-doc-segmented.bm25-default.topics.dl20.txt
bin/trec_eval -c -m recall.100 tools/topics-and-qrels/qrels.dl20-doc.txt runs/run.msmarco-doc-segmented.bm25-default.topics.dl20.txt
bin/trec_eval -c -m recall.1000 tools/topics-and-qrels/qrels.dl20-doc.txt runs/run.msmarco-doc-segmented.bm25-default.topics.dl20.txt

Effectiveness

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

AP@100 BM25 (default)
DL20 (Doc) 0.3568
nDCG@10 BM25 (default)
DL20 (Doc) 0.5196
R@100 BM25 (default)
DL20 (Doc) 0.5912
R@1000 BM25 (default)
DL20 (Doc) 0.7856

Explanation of settings:

  • The setting "default" refers the default BM25 settings of k1=0.9, b=0.4.

Note that in the official evaluation for document ranking, all runs were truncated to top-100 hits per query (whereas all top-1000 hits per query were retained for passage ranking). Thus, average precision is computed to depth 100 (i.e., AP@100); nDCG@10 remains unaffected. Remember that we keep qrels of all relevance grades, unlike the case for passage ranking, where relevance grade 1 needs to be discarded when computing certain metrics. Here, we retrieve 1000 hits per query, but measure AP at cutoff 100 (e.g., AP@100). Thus, the experimental results reported here are directly comparable to the results reported in the track overview paper.

Note that #1721 slightly change the results, since we corrected underlying issues with data preparation.