Skip to content

Commit

Permalink
Update regressions for DeepImpact and uniCOIL (#1657)
Browse files Browse the repository at this point in the history
  • Loading branch information
lintool authored Oct 17, 2021
1 parent c67dc7f commit cf5c4fb
Show file tree
Hide file tree
Showing 12 changed files with 268 additions and 37 deletions.
31 changes: 16 additions & 15 deletions README.md
Original file line number Diff line number Diff line change
Expand Up @@ -50,25 +50,26 @@ For the most part, these runs are based on [_default_ parameter settings](https:
+ Regressions for [Wt10g](docs/regressions-wt10g.md), [Gov2](docs/regressions-gov2.md)
+ Regressions for [ClueWeb09 (Category B)](docs/regressions-cw09b.md), [ClueWeb12-B13](docs/regressions-cw12b13.md), [ClueWeb12](docs/regressions-cw12.md)
+ Regressions for [Tweets2011 (MB11 & MB12)](docs/regressions-mb11.md), [Tweets2013 (MB13 & MB14)](docs/regressions-mb13.md)
+ Regressions for Complex Answer Retrieval (CAR17): [[v1.5](docs/regressions-car17v1.5.md)] [[v2.0](docs/regressions-car17v2.0.md)] [[v2.0 with doc2query](docs/regressions-car17v2.0-doc2query.md)]
+ Regressions for MS MARCO Passage Ranking: [[base](docs/regressions-msmarco-passage.md)] [[doc2query](docs/regressions-msmarco-passage-doc2query.md)] [[docTTTTTquery](docs/regressions-msmarco-passage-docTTTTTquery.md)]
+ Regressions for MS MARCO Passage Ranking: [[DeepImpact](docs/regressions-msmarco-passage-deepimpact.md)] [[uniCOIL](docs/regressions-msmarco-passage-unicoil.md)]
+ Regressions for MS MARCO Document Ranking, Per Doc: [[base](docs/regressions-msmarco-doc.md)] [[docTTTTTquery](docs/regressions-msmarco-doc-docTTTTTquery-per-doc.md)]
+ Regressions for MS MARCO Document Ranking, Per Passage: [[base](docs/regressions-msmarco-doc-per-passage.md)] [[docTTTTTquery](docs/regressions-msmarco-doc-docTTTTTquery-per-passage.md)]
+ Regressions for the TREC 2019 Deep Learning Track (Passage): [[base](docs/regressions-dl19-passage.md)] [[docTTTTTquery](docs/regressions-dl19-passage-docTTTTTquery.md)]
+ Regressions for the TREC 2019 Deep Learning Track (Document), Per Doc: [[base](docs/regressions-dl19-doc.md)] [[docTTTTTquery](docs/regressions-dl19-doc-docTTTTTquery-per-doc.md)]
+ Regressions for the TREC 2019 Deep Learning Track (Document), Per Passage: [[base](docs/regressions-dl19-doc-per-passage.md)] [[docTTTTTquery](docs/regressions-dl19-doc-docTTTTTquery-per-passage.md)]
+ Regressions for the TREC 2020 Deep Learning Track (Passage): [[base](docs/regressions-dl20-passage.md)] [[docTTTTTquery](docs/regressions-dl20-passage-docTTTTTquery.md)]
+ Regressions for the TREC 2020 Deep Learning Track (Document), Per Doc: [[base](docs/regressions-dl20-doc.md)] [[docTTTTTquery](docs/regressions-dl20-doc-docTTTTTquery-per-doc.md)]
+ Regressions for the TREC 2020 Deep Learning Track (Document), Per Passage: [[base](docs/regressions-dl20-doc-per-passage.md)] [[docTTTTTquery](docs/regressions-dl20-doc-docTTTTTquery-per-passage.md)]
+ Regressions for MS MARCO (V2) Passage Ranking: [[base](docs/regressions-msmarco-v2-passage.md)] [[base on augmented corpus](docs/regressions-msmarco-v2-passage-augmented.md)]
+ Regressions for MS MARCO (V2) Document Ranking: [[base](docs/regressions-msmarco-v2-doc.md)] [[base on segmented corpus](docs/regressions-msmarco-v2-doc-segmented.md)]
+ Regressions for the TREC News Track (Background Linking Task): [[2018](docs/regressions-backgroundlinking18.md)] [[2019](docs/regressions-backgroundlinking19.md)] [[2020](docs/regressions-backgroundlinking20.md)]
+ Regressions for Complex Answer Retrieval (CAR17): [v1.5](docs/regressions-car17v1.5.md), [v2.0](docs/regressions-car17v2.0.md), [v2.0 with doc2query](docs/regressions-car17v2.0-doc2query.md)
+ Regressions for MS MARCO Passage Ranking: [baselines](docs/regressions-msmarco-passage.md), [doc2query](docs/regressions-msmarco-passage-doc2query.md), [doc2query-T5](docs/regressions-msmarco-passage-docTTTTTquery.md)
+ Regressions for MS MARCO Passage Ranking: [DeepImpact](docs/regressions-msmarco-passage-deepimpact.md)
+ Regressions for MS MARCO Passage Ranking: [uniCOIL with doc2query-T5](docs/regressions-msmarco-passage-unicoil.md), [uniCOIL with TILDE](docs/regressions-msmarco-passage-unicoil-tilde-expansion.md)
+ Regressions for MS MARCO Document Ranking, Per Doc: [baselines](docs/regressions-msmarco-doc.md), [doc2query-T5](docs/regressions-msmarco-doc-docTTTTTquery-per-doc.md)
+ Regressions for MS MARCO Document Ranking, Per Passage: [baselines](docs/regressions-msmarco-doc-per-passage.md), [doc2query-T5](docs/regressions-msmarco-doc-docTTTTTquery-per-passage.md)
+ Regressions for TREC 2019 Deep Learning (Passage): [baselines](docs/regressions-dl19-passage.md), [doc2query-T5](docs/regressions-dl19-passage-docTTTTTquery.md)
+ Regressions for TREC 2019 Deep Learning (Document), Per Doc: [baselines](docs/regressions-dl19-doc.md), [doc2query-T5](docs/regressions-dl19-doc-docTTTTTquery-per-doc.md)
+ Regressions for TREC 2019 Deep Learning (Document), Per Passage: [baselines](docs/regressions-dl19-doc-per-passage.md), [doc2query-T5](docs/regressions-dl19-doc-docTTTTTquery-per-passage.md)
+ Regressions for TREC 2020 Deep Learning (Passage): [baselines](docs/regressions-dl20-passage.md), [doc2query-T5](docs/regressions-dl20-passage-docTTTTTquery.md)
+ Regressions for TREC 2020 Deep Learning (Document), Per Doc: [baselines](docs/regressions-dl20-doc.md), [doc2query-T5](docs/regressions-dl20-doc-docTTTTTquery-per-doc.md)
+ Regressions for TREC 2020 Deep Learning (Document), Per Passage: [baselines](docs/regressions-dl20-doc-per-passage.md), [doc2query-T5](docs/regressions-dl20-doc-docTTTTTquery-per-passage.md)
+ Regressions for MS MARCO (V2) Passage Ranking: [baselines](docs/regressions-msmarco-v2-passage.md), [baselines on augmented corpus](docs/regressions-msmarco-v2-passage-augmented.md)
+ Regressions for MS MARCO (V2) Document Ranking: [baselines](docs/regressions-msmarco-v2-doc.md), [baselines on segmented corpus](docs/regressions-msmarco-v2-doc-segmented.md)
+ Regressions for TREC News Tracks (Background Linking Task): [2018](docs/regressions-backgroundlinking18.md), [2019](docs/regressions-backgroundlinking19.md), [2020](docs/regressions-backgroundlinking20.md)
+ Regressions for [FEVER Fact Verification](docs/regressions-fever.md)
+ Regressions for [NTCIR-8 ACLIA (IR4QA subtask, Monolingual Chinese)](docs/regressions-ntcir8-zh.md)
+ Regressions for [CLEF 2006 Monolingual French](docs/regressions-clef06-fr.md)
+ Regressions for [TREC 2002 Monolingual Arabic](docs/regressions-trec02-ar.md)
+ Regressions for FIRE 2012: [[Monolingual Bengali](docs/regressions-fire12-bn.md)] [[Monolingual Hindi](docs/regressions-fire12-hi.md)] [[Monolingual English](docs/regressions-fire12-en.md)]
+ Regressions for FIRE 2012: [Monolingual Bengali](docs/regressions-fire12-bn.md), [Monolingual Hindi](docs/regressions-fire12-hi.md), [Monolingual English](docs/regressions-fire12-en.md)

## Reproduction Guides

Expand Down
3 changes: 1 addition & 2 deletions docs/experiments-msmarco-passage-unicoil-tilde-expansion.md
Original file line number Diff line number Diff line change
Expand Up @@ -2,8 +2,7 @@

This page describes how to reproduce experiments using uniCOIL with TILDE document expansion on the MS MARCO passage corpus, as described in the following paper:

> Shengyao Zhuang and Guido Zuccon. [Fast Passage Re-ranking with Contextualized Exact Term
Matching and Efficient Passage Expansion.](https://arxiv.org/pdf/2108.08513) _arXiv:2108.08513_.
> Shengyao Zhuang and Guido Zuccon. [Fast Passage Re-ranking with Contextualized Exact Term Matching and Efficient Passage Expansion.](https://arxiv.org/pdf/2108.08513) _arXiv:2108.08513_.
The original uniCOIL model is described here:

Expand Down
14 changes: 7 additions & 7 deletions docs/regressions-msmarco-passage-deepimpact.md
Original file line number Diff line number Diff line change
Expand Up @@ -17,9 +17,9 @@ Typical indexing command:
```
nohup sh target/appassembler/bin/IndexCollection -collection JsonVectorCollection \
-input /path/to/msmarco-passage-deepimpact \
-index indexes/lucene-index.msmarco-passage-deepimpact.raw \
-index indexes/lucene-index.msmarco-passage-deepimpact \
-generator DefaultLuceneDocumentGenerator \
-threads 16 -impact -pretokenized -storeRaw \
-threads 16 -impact -pretokenized \
>& logs/log.msmarco-passage-deepimpact &
```

Expand All @@ -36,7 +36,7 @@ The regression experiments here evaluate on the 6980 dev set questions; see [thi
After indexing has completed, you should be able to perform retrieval as follows:

```
nohup target/appassembler/bin/SearchCollection -index indexes/lucene-index.msmarco-passage-deepimpact.raw \
nohup target/appassembler/bin/SearchCollection -index indexes/lucene-index.msmarco-passage-deepimpact \
-topicreader TsvInt -topics src/main/resources/topics-and-qrels/topics.msmarco-passage.dev-subset.deepimpact.tsv.gz \
-output runs/run.msmarco-passage-deepimpact.deepimpact.topics.msmarco-passage.dev-subset.deepimpact.tsv.gz \
-impact -pretokenized &
Expand Down Expand Up @@ -71,12 +71,12 @@ In order to reproduce results reported in the paper, we need to convert to MS MA

```bash
python tools/scripts/msmarco/convert_trec_to_msmarco_run.py \
--input runs/run.msmarco-passage-deepimpact.deepimpact.topics.msmarco-passage.dev-subset.deep-impact.tsv.gz \
--output runs/run.msmarco-passage-deepimpact.deepimpact.topics.msmarco-passage.dev-subset.deep-impact.tsv.gz.msmarco --quiet
--input runs/run.msmarco-passage-deepimpact.deepimpact.topics.msmarco-passage.dev-subset.deepimpact.tsv.gz \
--output runs/run.msmarco-passage-deepimpact.deepimpact.topics.msmarco-passage.dev-subset.deepimpact.tsv.gz.msmarco --quiet

python tools/scripts/msmarco/msmarco_passage_eval.py \
collections/msmarco-passage/qrels.dev.small.tsv \
runs/run.msmarco-passage-deepimpact.deepimpact.topics.msmarco-passage.dev-subset.deep-impact.tsv.gz.msmarco
tools/topics-and-qrels/qrels.msmarco-passage.dev-subset.txt \
runs/run.msmarco-passage-deepimpact.deepimpact.topics.msmarco-passage.dev-subset.deepimpact.tsv.gz.msmarco
```

The results should be as follows:
Expand Down
91 changes: 91 additions & 0 deletions docs/regressions-msmarco-passage-unicoil-tilde-expansion.md
Original file line number Diff line number Diff line change
@@ -0,0 +1,91 @@
# Anserini: Regressions for uniCOIL w/ TILDE on [MS MARCO Passage](https://github.com/microsoft/MSMARCO-Passage-Ranking)

This page documents regression experiments for uniCOIL w/ TILDE document expansion on the MS MARCO Passage Ranking Task, which is integrated into Anserini's regression testing framework.
The model is described in the following paper:

> Shengyao Zhuang and Guido Zuccon. [Fast Passage Re-ranking with Contextualized Exact Term Matching and Efficient Passage Expansion.](https://arxiv.org/pdf/2108.08513) _arXiv:2108.08513_.
For more complete instructions on how to run end-to-end experiments, refer to [this page](experiments-msmarco-passage-unicoil-tilde-expansion.md).

The exact configurations for these regressions are stored in [this YAML file](../src/main/resources/regression/msmarco-passage-unicoil-tilde-expansion.yaml).
Note that this page is automatically generated from [this template](../src/main/resources/docgen/templates/msmarco-passage-unicoil-tilde-expansion.template) as part of Anserini's regression pipeline, so do not modify this page directly; modify the template instead.

## Indexing

Typical indexing command:

```
nohup sh target/appassembler/bin/IndexCollection -collection JsonVectorCollection \
-input /path/to/msmarco-passage-unicoil-tilde-expansion \
-index indexes/lucene-index.msmarco-passage-unicoil-tilde-expansion \
-generator DefaultLuceneDocumentGenerator \
-threads 16 -impact -pretokenized \
>& logs/log.msmarco-passage-unicoil-tilde-expansion &
```

The directory `/path/to/msmarco-passage-unicoil-tilde-expansion/` should be a directory containing the compressed `jsonl` files that comprise the corpus.
See [this page](experiments-msmarco-passage-unicoil-tilde-expansion.md) for additional details.

For additional details, see explanation of [common indexing options](common-indexing-options.md).

## Retrieval

Topics and qrels are stored in [`src/main/resources/topics-and-qrels/`](../src/main/resources/topics-and-qrels/).
The regression experiments here evaluate on the 6980 dev set questions; see [this page](experiments-msmarco-passage.md) for more details.

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

```
nohup target/appassembler/bin/SearchCollection -index indexes/lucene-index.msmarco-passage-unicoil-tilde-expansion \
-topicreader TsvInt -topics src/main/resources/topics-and-qrels/topics.msmarco-passage.dev-subset.unicoil-tilde-expansion.tsv.gz \
-output runs/run.msmarco-passage-unicoil-tilde-expansion.unicoil.topics.msmarco-passage.dev-subset.unicoil-tilde-expansion.tsv.gz \
-impact -pretokenized &
```

Evaluation can be performed using `trec_eval`:

```
tools/eval/trec_eval.9.0.4/trec_eval -m map -c -m recip_rank -c -m recall.1000 -c src/main/resources/topics-and-qrels/qrels.msmarco-passage.dev-subset.txt runs/run.msmarco-passage-unicoil-tilde-expansion.unicoil.topics.msmarco-passage.dev-subset.unicoil-tilde-expansion.tsv.gz
```

## Effectiveness

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

MAP | uniCOIL |
:---------------------------------------|-----------|
[MS MARCO Passage: Dev](https://github.com/microsoft/MSMARCO-Passage-Ranking)| 0.3560 |


MRR | uniCOIL |
:---------------------------------------|-----------|
[MS MARCO Passage: Dev](https://github.com/microsoft/MSMARCO-Passage-Ranking)| 0.3606 |


R@1000 | uniCOIL |
:---------------------------------------|-----------|
[MS MARCO Passage: Dev](https://github.com/microsoft/MSMARCO-Passage-Ranking)| 0.9646 |

The above runs are in TREC output format and evaluated with `trec_eval`.
In order to reproduce results reported in the paper, we need to convert to MS MARCO output format and then evaluate:

```bash
python tools/scripts/msmarco/convert_trec_to_msmarco_run.py \
--input runs/run.msmarco-passage-unicoil-tilde-expansion.unicoil.topics.msmarco-passage.dev-subset.unicoil-tilde-expansion.tsv.gz \
--output runs/run.msmarco-passage-unicoil-tilde-expansion.unicoil.topics.msmarco-passage.dev-subset.unicoil-tilde-expansion.tsv.gz.msmarco --quiet

python tools/scripts/msmarco/msmarco_passage_eval.py \
tools/topics-and-qrels/qrels.msmarco-passage.dev-subset.txt \
runs/run.msmarco-passage-unicoil-tilde-expansion.unicoil.topics.msmarco-passage.dev-subset.unicoil-tilde-expansion.tsv.gz.msmarco
```

The results should be as follows:

```
#####################
MRR @10: 0.34957184927457136
QueriesRanked: 6980
#####################
```

This corresponds to the effectiveness reported in the paper.
6 changes: 3 additions & 3 deletions docs/regressions-msmarco-passage-unicoil.md
Original file line number Diff line number Diff line change
Expand Up @@ -17,9 +17,9 @@ Typical indexing command:
```
nohup sh target/appassembler/bin/IndexCollection -collection JsonVectorCollection \
-input /path/to/msmarco-passage-unicoil \
-index indexes/lucene-index.msmarco-passage-unicoil.raw \
-index indexes/lucene-index.msmarco-passage-unicoil \
-generator DefaultLuceneDocumentGenerator \
-threads 16 -impact -pretokenized -storeRaw \
-threads 16 -impact -pretokenized \
>& logs/log.msmarco-passage-unicoil &
```

Expand All @@ -36,7 +36,7 @@ The regression experiments here evaluate on the 6980 dev set questions; see [thi
After indexing has completed, you should be able to perform retrieval as follows:

```
nohup target/appassembler/bin/SearchCollection -index indexes/lucene-index.msmarco-passage-unicoil.raw \
nohup target/appassembler/bin/SearchCollection -index indexes/lucene-index.msmarco-passage-unicoil \
-topicreader TsvInt -topics src/main/resources/topics-and-qrels/topics.msmarco-passage.dev-subset.unicoil.tsv.gz \
-output runs/run.msmarco-passage-unicoil.unicoil.topics.msmarco-passage.dev-subset.unicoil.tsv.gz \
-impact -pretokenized &
Expand Down
2 changes: 1 addition & 1 deletion docs/regressions-msmarco-v2-passage.md
Original file line number Diff line number Diff line change
@@ -1,4 +1,4 @@
# Anserini: Regressions for [MS MARCO (V2) passage Ranking](https://microsoft.github.io/msmarco/TREC-Deep-Learning.html)
# Anserini: Regressions for [MS MARCO (V2) Passage Ranking](https://microsoft.github.io/msmarco/TREC-Deep-Learning.html)

This page documents regression experiments for passage ranking on the MS MARCO (V2) passage corpus, which is integrated into Anserini's regression testing framework.
For more complete instructions on how to run end-to-end experiments, refer to [this page](experiments-msmarco-v2.md).
Expand Down
Original file line number Diff line number Diff line change
Expand Up @@ -51,12 +51,12 @@ In order to reproduce results reported in the paper, we need to convert to MS MA

```bash
python tools/scripts/msmarco/convert_trec_to_msmarco_run.py \
--input runs/run.msmarco-passage-deepimpact.deepimpact.topics.msmarco-passage.dev-subset.deep-impact.tsv.gz \
--output runs/run.msmarco-passage-deepimpact.deepimpact.topics.msmarco-passage.dev-subset.deep-impact.tsv.gz.msmarco --quiet
--input runs/run.msmarco-passage-deepimpact.deepimpact.topics.msmarco-passage.dev-subset.deepimpact.tsv.gz \
--output runs/run.msmarco-passage-deepimpact.deepimpact.topics.msmarco-passage.dev-subset.deepimpact.tsv.gz.msmarco --quiet

python tools/scripts/msmarco/msmarco_passage_eval.py \
collections/msmarco-passage/qrels.dev.small.tsv \
runs/run.msmarco-passage-deepimpact.deepimpact.topics.msmarco-passage.dev-subset.deep-impact.tsv.gz.msmarco
tools/topics-and-qrels/qrels.msmarco-passage.dev-subset.txt \
runs/run.msmarco-passage-deepimpact.deepimpact.topics.msmarco-passage.dev-subset.deepimpact.tsv.gz.msmarco
```

The results should be as follows:
Expand Down
Original file line number Diff line number Diff line change
@@ -0,0 +1,71 @@
# Anserini: Regressions for uniCOIL w/ TILDE on [MS MARCO Passage](https://github.com/microsoft/MSMARCO-Passage-Ranking)

This page documents regression experiments for uniCOIL w/ TILDE document expansion on the MS MARCO Passage Ranking Task, which is integrated into Anserini's regression testing framework.
The model is described in the following paper:

> Shengyao Zhuang and Guido Zuccon. [Fast Passage Re-ranking with Contextualized Exact Term Matching and Efficient Passage Expansion.](https://arxiv.org/pdf/2108.08513) _arXiv:2108.08513_.

For more complete instructions on how to run end-to-end experiments, refer to [this page](experiments-msmarco-passage-unicoil-tilde-expansion.md).

The exact configurations for these regressions are stored in [this YAML file](../src/main/resources/regression/msmarco-passage-unicoil-tilde-expansion.yaml).
Note that this page is automatically generated from [this template](../src/main/resources/docgen/templates/msmarco-passage-unicoil-tilde-expansion.template) as part of Anserini's regression pipeline, so do not modify this page directly; modify the template instead.

## Indexing

Typical indexing command:

```
${index_cmds}
```

The directory `/path/to/msmarco-passage-unicoil-tilde-expansion/` should be a directory containing the compressed `jsonl` files that comprise the corpus.
See [this page](experiments-msmarco-passage-unicoil-tilde-expansion.md) for additional details.

For additional details, see explanation of [common indexing options](common-indexing-options.md).

## Retrieval

Topics and qrels are stored in [`src/main/resources/topics-and-qrels/`](../src/main/resources/topics-and-qrels/).
The regression experiments here evaluate on the 6980 dev set questions; see [this page](experiments-msmarco-passage.md) for more details.

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

```
${ranking_cmds}
```

Evaluation can be performed using `trec_eval`:

```
${eval_cmds}
```

## Effectiveness

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

${effectiveness}

The above runs are in TREC output format and evaluated with `trec_eval`.
In order to reproduce results reported in the paper, we need to convert to MS MARCO output format and then evaluate:

```bash
python tools/scripts/msmarco/convert_trec_to_msmarco_run.py \
--input runs/run.msmarco-passage-unicoil-tilde-expansion.unicoil.topics.msmarco-passage.dev-subset.unicoil-tilde-expansion.tsv.gz \
--output runs/run.msmarco-passage-unicoil-tilde-expansion.unicoil.topics.msmarco-passage.dev-subset.unicoil-tilde-expansion.tsv.gz.msmarco --quiet

python tools/scripts/msmarco/msmarco_passage_eval.py \
tools/topics-and-qrels/qrels.msmarco-passage.dev-subset.txt \
runs/run.msmarco-passage-unicoil-tilde-expansion.unicoil.topics.msmarco-passage.dev-subset.unicoil-tilde-expansion.tsv.gz.msmarco
```

The results should be as follows:

```
#####################
MRR @10: 0.34957184927457136
QueriesRanked: 6980
#####################
```

This corresponds to the effectiveness reported in the paper.
Loading

0 comments on commit cf5c4fb

Please sign in to comment.