Pith. sign in

REVIEW 4 major objections 5 minor 30 references

Improving Vietnamese Legal Document Retrieval using Synthetic Data

T0 review · 4 major / 5 minor · reviewed 2026-08-12 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read 500k synthetic queries lift Vietnamese legal retrieval to 74.61 MRR

desk verdict Useful new Vietnamese legal retrieval resources, but the headline gain from synthetic data is not isolated from in-domain supervision and a shared source with the test set. read the letter →

arxiv 2412.00657 v1 pith:7L2G7LTG submitted 2024-12-01 cs.IR cs.AI

Son Pham Tien , Hieu Nguyen Doan , An Nguyen Dai , Sang Dinh Viet This is my paper · ORCID
classification cs.IRcs.AI
keywords InformationRetrievalVietnameselegaltextsyntheticquerygenerationLLM-generateddatadenseColBERTQuery-as-Contextpre-trainingcontrastivefine-tuning
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

This paper argues that a large set of synthetic Vietnamese legal queries, generated by Llama 3 70B from real legal passages and filtered for quality, can stand in for the human-annotated training data that is missing for legal retrieval. The authors show that pre-training a PhoBERT (Vietnamese BERT) based retriever on these queries with a Query-as-Context masked auto-encoder objective, then fine-tuning with contrastive loss and hard negatives, produces large gains for both single-vector bi-encoders and multi-vector ColBERT models. On the new TVPL benchmark and on Legal Zalo 21, the best ColBERT model reaches MRR@10 (how high the correct passage ranks in the top ten results) of 74.61 and 84.08, outperforming the listed dense baselines. If correct, this gives a cheap, scalable path to building legal search and question-answering systems in Vietnamese and other low-resource legal domains.

What carries the argument

The load-bearing mechanism is an aspect-guided synthetic query generator coupled with a Query-as-Context pre-training objective. The generator prompts Llama 3 to identify 1-5 distinct aspects covered by a legal passage and write one question per aspect; this step alone raises the BGE-M3 passage hit rate at k=10 from 8.26% to 82.06% compared with direct prompting. Pre-training applies the CoT-MAE loss, in which an encoder reconstructs the masked passage while a decoder reconstructs the generated query from its unmasked tokens conditioned on the passage, so the encoder learns to align query and passage representations before any supervised fine-tuning. Fine-tuning then uses the InfoNCE contrastive loss with BGE-M3-mined hard negatives and in-batch negatives, and the final deployed model applies ColBERT's 2-bit residual compression.

What would settle it

Compute exact and near-duplicate overlap between the 143,261 passages used to generate synthetic queries and the 224,006 TVPL corpus passages, and measure retrieval-based similarity between the 507,152 synthetic queries and the 10,000 TVPL test queries. If even a modest fraction of test passages or test queries are covered by the synthetic generation passages, the reported TVPL gains are explained by leakage rather than by synthetic data improving generalization.

Watch

Extended reading notes

Core claim

The central discovery is that synthetic queries are not just a data-augmentation trick; they carry the training signal that would otherwise require tens of thousands of human-annotated legal QA pairs. Starting from 143,261 legal passages, the authors prompt Llama 3 to first list 1-5 aspects of a passage and then write one question for each aspect, producing 507,152 queries after BGE-M3-based filtering and removal of queries that explicitly reference the source passage. These pairs are used twice: to pre-train the encoder under the CoT-MAE Query-as-Context objective, and again as fine-tuning data alongside translated MS-MARCO and SQuAD 2.0 plus existing legal benchmarks. The result is consistent across settings: Query-as-Context pre-training improves over fine-tuning alone, ColBERT outperforms the bi-encoder, and the best ColBERT reaches MRR@10 of 74.61 on TVPL and 84.08 on Legal Zalo 21, ahead of the reported baselines. On the out-of-domain Zalo QA 19 benchmark the same model reaches MRR@10 of 72.38, close to much larger multilingual retrievers.

Load-bearing premise

The result hinges on the TVPL evaluation being genuinely unseen, because the synthetic training queries and the TVPL test queries come from the same legal website and the paper never measures passage overlap or query similarity between them.

Editorial extensions

If this is right

  • The 507,152 synthetic query-passage pairs are released under CC BY 4.0, giving other Vietnamese NLP teams a ready-made retrieval training set.
  • Query-as-Context pre-training yields measurable gains over fine-tuning alone on in-domain benchmarks, suggesting the pre-training step will transfer to other low-resource legal languages.
  • Because 1-bit compressed ColBERT (647 MB) beats the 672 MB bi-encoder on MRR@10 and MAP@10 on TVPL, practitioners can get better accuracy without paying a storage penalty.
  • The improved out-of-domain Zalo QA 19 scores suggest the recipe transfers beyond legal text, despite being trained solely on legal data.
  • The aspect-guided prompting recipe and the BGE-M3 filtering step together form a reusable pipeline for generating synthetic retrieval data for domains where annotated data is scarce.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • If the TVPL evaluation passages overlap the passages used for synthetic query generation, the in-domain gains would be partly a training/evaluation leakage artifact; the paper does not report any overlap check, so an external held-out legal corpus would be the clean test.
  • The aspect-first prompting insight (generate aspects, then a question per aspect) likely generalizes to other specialized domains such as medical or financial regulation, and could be tested by measuring hit-rate improvements on those corpora.
  • The top-40 BGE-M3 filter may bias the synthetic set toward queries that dense retrieval can already answer, potentially under-representing hard, long-tail legal questions; an unfiltered sample comparison would reveal how much this matters.
  • A pure-synthetic training run (without the translated MS-MARCO and SQuAD data) would isolate how much of the gain comes from synthetic legal data versus multilingual data; the paper's ablations currently bundle both.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

4 major / 5 minor

Summary. The paper proposes a pipeline for improving Vietnamese legal passage retrieval: (1) collect legal documents from thuvienphapluat.vn and split them into 143,261 passages; (2) use Llama 3 70B with an aspect-guided prompt to generate over 620,000 synthetic queries, filtered by BGE-M3 top-40 retrieval and by excluding queries with deictic references, yielding 507,152 query-passage pairs; (3) pre-train PhoBERT-base-v2 with the Query-as-Context CoT-MAE objective; (4) fine-tune bi-encoder and ColBERT models with contrastive InfoNCE loss and BGE-M3-mined hard negatives on a mixture of MS-MARCO, SQuAD 2.0, Legal Zalo 21 (80%), the newly introduced TVPL dataset, and the synthetic queries. Evaluation on TVPL, Legal Zalo 21 (20%), and Vietnamese Wiki QA from Zalo AI Challenge 2019 shows that the proposed models outperform the listed sparse and dense baselines, with the best CoT-MAE ColBERT achieving MRR@10 of 74.61 on TVPL and 84.08 on Legal Zalo 21.

Significance. If the reported gains are attributable to the synthetic-data pipeline rather than to confounds, the work is a useful contribution to low-resource and Vietnamese legal retrieval: it provides a large synthetic query dataset, a new benchmark (TVPL), and a reproducible recipe combining LLM-generated queries with CoT-MAE pre-training and hard-negative contrastive fine-tuning. The authors also release the synthetic data and TVPL queries on Hugging Face, which is a concrete asset for the community. However, the central claim of 'strong improvement' depends on TVPL being a genuinely held-out benchmark and on a controlled comparison that isolates the synthetic-data component; the current manuscript does not provide either, so the significance is conditional on the results of the requested analyses.

major comments (4)
  1. [§3.2, §4.1, Table 3] The TVPL benchmark may not be held out from the synthetic-data generation pipeline. Section 3.2 states that synthetic queries are generated from 143,261 passages scraped from thuvienphapluat.vn, and Section 4.1 states that the TVPL corpus (224,006 passages) and its 10,000 test queries are also sourced from that site's legal QA articles. The paper does not report any passage-level overlap analysis between the synthetic-generation set and the TVPL corpus/test set, nor any query-similarity analysis between synthetic queries and TVPL test queries. If a substantial fraction of TVPL test passages or near-duplicates were among the passages fed to Llama 3, the headline gains in Table 3 (e.g., CoT-MAE ColBERT MRR@10 74.61 on TVPL) could reflect memorization of passage content rather than a generalizable synthetic-data benefit. Please add a quantitative overlap analysis, ideally with exact-match and near-duplicate (e.g., n-gram or embedding-similarity) thresholds, and report the overlap rate for both the TVPL corpus and the TVPL test set.
  2. [§4.3, Table 3, §3.5] The baseline comparisons in Table 3 are not controlled for fine-tuning data and protocol. The dense baselines (vietnamese-sbert, vietnamese-bi-encoder, mE5base, BGE-M3) are used as pre-trained off-the-shelf retrievers, whereas the proposed models are fine-tuned on a large in-domain mixture that includes TVPL training queries and 80% of Legal Zalo 21. Part of the gap is therefore expected from in-domain supervised fine-tuning, not from synthetic data. Moreover, the paper does not ablate the synthetic component from the fine-tuning mixture: no model is trained on the same mixture with the 507,152 synthetic pairs removed, so the contribution of synthetic data to fine-tuning is not isolated. The cleanest within-paper comparison (with vs. without CoT-MAE pre-training, both fine-tuned identically) shows small gains (Table 3: MRR@10 +0.32 for bi-encoder, +0.71 for ColBERT on TVPL), which do not by themselves establish the abstract's 'strong improvement.' Please add an ablation that trains the same bi-encoder and ColBERT models on exactly the same fine-tuning mixture without the synthetic queries, and report the delta.
  3. [§4.5, Table 4] The out-of-domain results do not support the claim of generalizable improvement. In Table 4, CoT-MAE Bi-encoder (68.22 MRR@10) is numerically worse than Bi-encoder without pre-training (69.57), and the CoT-MAE ColBERT gain over ColBERT is 2.04 MRR@10 but still below BGE-M3 (72.38 vs. 76.69). The paper's statement that 'our models demonstrated improved performance on the out-of-domain dataset' is only partially true and is not quantified against the non-pre-trained models. Please either temper the claim or provide a statistical significance test and a direct discussion of the mixed out-of-domain results; as written, the out-of-domain evaluation undermines rather than supports the central attribution.
  4. [§3.3, §3.5, Table 3] The same retriever, BGE-M3, is used in two roles that shape the training distribution and is also a baseline in Table 3: it filters synthetic queries (Section 3.3) and mines hard negatives for fine-tuning (Section 3.5). While this is not circular in the strict sense—final evaluation uses separate test queries—it means the training distribution is biased toward passages that BGE-M3 can retrieve, and the baseline comparison against BGE-M3 in Table 3 is not apples-to-apples with respect to negative mining. Please discuss this potential bias explicitly and, if feasible, report an ablation that uses hard negatives mined with the trained model itself or with BM25 to check sensitivity.
minor comments (5)
  1. [Table 4] The mE5base row reports '72,76' with a comma as the decimal separator, while all other numbers in the table use periods; please standardize.
  2. [Reference [10]] The reference is titled 'Vienamese sbert' and the Hugging Face URL is informal; please correct the typo ('Vietnamese') and, if possible, include a version or accession date for the model.
  3. [§4.2] The description of fine-tuning says the bi-encoder was trained for 170,000 steps (5 epochs) and ColBERT for 290,000 steps (9 epochs), but the total number of training queries in the mixture is not reported; please state the per-epoch query count or the number of updates per dataset so a reader can assess whether the longer ColBERT schedule is a confound.
  4. [§3.3] The sentence 'Applying this method, we generated over 620,000 legal queries from 140,292 passages' uses a different passage count from the 143,261 quoted in §3.2; please reconcile the two numbers.
  5. [§5.1, Table 5] The passage hit rates in Table 5 (8.26% for basic prompting at k=10 vs. 82.06% for aspect-guided prompting) are striking; please clarify whether the basic-prompt queries were also filtered with BGE-M3, because the filtering step in §3.3 could differentially remove low-quality queries from the two prompt variants.

Circularity Check

1 steps flagged · score 3.0 of 10

One supporting analysis is self-referential (BGE-M3 both filters and scores the synthetic queries), but the main retrieval results are not circular by construction.

  1. self definitional [Section 3.3 (Stage 3 filtering) and Section 5.1 (Table 5)]
    "For queries that are only shallowly relevant, we use the BGE-M3 dense retriever to filter out synthetic queries that cannot recover their input passage within the top 40 retrieved results. ... Performance is evaluated using passage hit rate (the percentage of queries retrieving their corresponding passage) and document hit rate ... We use the BGE-M3 dense retriever [4] to rank the top-k relevant passages for each query."

    The aspect-guided synthetic queries were retained only if BGE-M3 could recover their source passage within the top 40 retrieved results. Section 5.1 then evaluates the same BGE-M3 passage hit rate at k=40 (91.90%) as evidence that the aspect-guided prompt improves generation quality over the basic prompt. For queries drawn from the filtered set, the k=40 passage hit rate is the same criterion used to define the set, so the high score is at least partly guaranteed by the filter rather than independently measured. The comparison is also confounded because the basic-prompt row was not necessarily subjected to the same BGE-M3 filtering step. Thus the quality analysis in Table 5 reduces to the filtering input and cannot by itself validate the prompt design.

full rationale

The paper's central claim—that Llama-3-generated, BGE-M3-filtered synthetic queries, combined with Query-as-Context CoT-MAE pre-training and contrastive fine-tuning, improve Vietnamese legal retrieval—is an empirical result evaluated on held-out TVPL test queries and Legal Zalo 21 test split. No equation or construction makes these test predictions equal to the training inputs; supervised fine-tuning on a separate train/test split is not circular. The potential overlap between the thuvienphapluat.vn passages used for synthetic query generation and the TVPL benchmark corpus is a data-contamination risk, but the paper never states that TVPL test passages were included in the generation corpus, so it cannot be counted as demonstrated circularity under the requirement to exhibit a specific reduction. The only concrete self-referential step is the Section 5.1 quality analysis: BGE-M3 was used both to filter synthetic queries by top-40 recoverability and to score passage hit rate at k=40, making that supporting metric partially defined by its own filter. The sole self-citation of note (reference [20], a prior paper with overlapping authorship) appears as a baseline and translation approach, not as a load-bearing uniqueness claim or derivation. Overall, the main experimental contribution remains independent; the circularity is limited to a supporting analysis.

Assumptions & free parameters 5 free parameters · 6 assumptions · 0 invented entities

The central claim relies on assumptions about LLM-generated query quality, BGE-M3 as a reliable relevance filter, TVPL label correctness, translation fidelity of MS-MARCO and SQuAD, chunking adequacy, and absence of leakage between the synthetic training source and TVPL evaluation. No human evaluation or overlap analysis is provided for these assumptions. The listed free hyperparameters are design choices that affect dataset composition and the training objective; none are swept, so their individual contribution to the reported results is unknown.

free parameters (5)
  • Aspect count range in query generation prompt = 1 to 5 aspects per passage
    Prompt design choice in Section 3.3, selected after comparing with basic prompting in Section 5.1; it directly shapes diversity and relevance of the 507k synthetic queries.
  • BGE-M3 top-40 filtering threshold = k = 40
    Section 3.3 discards synthetic queries whose passage is not in the top 40 BGE-M3 results; the threshold is hand-picked and controls which queries enter pre-training and fine-tuning.
  • Hard negatives per query = 7 (bi-encoder), 15 (ColBERT)
    Section 3.5 uses these counts in the InfoNCE loss; neither is swept, and both affect the learned margin between positive and negative passages.
  • Fine-tuning training steps = 170,000 steps (bi-encoder), 290,000 steps (ColBERT)
    Section 4.2 chooses 5 and 9 epochs with no early-stopping or step sweep; ColBERT was trained longer because it 'continued improving throughout'.
  • InfoNCE temperature tau = 1
    Section 3.5 fixes tau=1; not tuned, but it scales all similarity logits and is a free choice in the loss function.
assumptions (6)
  • domain assumption Llama 3 70B generates relevant and diverse Vietnamese legal questions from passages.
    Section 3.3 relies on LLM output quality; the paper checks passage hit rate with BGE-M3 but gives no human evaluation of query relevance or diversity.
  • domain assumption BGE-M3's ranking is a valid proxy for relevance when filtering synthetic queries and mining hard negatives.
    Sections 3.3 and 3.5 use BGE-M3 both to accept/reject synthetic queries and to select hard negatives; any systematic error in BGE-M3 propagates into the training data.
  • domain assumption TVPL test labels, queries from legal QA articles on thuvienphapluat.vn, are correct relevance judgments.
    Section 4.1 constructs TVPL from legal QA articles; no protocol, annotator agreement, or human verification is described.
  • domain assumption Google Translate preserves query-passage relevance when translating MS-MARCO and SQuAD 2.0 into Vietnamese.
    Section 4.1 uses translated English datasets; translation errors can create mislabeled positives and negatives, and no quality check is reported.
  • domain assumption Chunking legal documents at 256 tokens keeps passage-level retrieval labels meaningful.
    Section 4.2 chunks Legal Zalo 21 passages to PhoBERT's 256-token limit; relevant information may be split across chunks, changing the retrieval task.
  • domain assumption Overlap between the passages used for synthetic query generation and the TVPL evaluation corpus does not materially inflate in-domain scores.
    Sections 3.2 and 4.1 both use thuvienphapluat.vn; no overlap or leakage analysis is provided, so valid held-out evaluation is assumed.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Improving Vietnamese Legal Document Retrieval using Synthetic Data." pith.science (2026). https://pith.science/paper/7L2G7LTG

@misc{pith2026241200657,
  author       = {Pith},
  title        = {Pith review of: Improving Vietnamese Legal Document Retrieval using Synthetic Data},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/7L2G7LTG}},
  note         = {Machine review of arXiv:2412.00657}
}
read the original abstract

In the field of legal information retrieval, effective embedding-based models are essential for accurate question-answering systems. However, the scarcity of large annotated datasets poses a significant challenge, particularly for Vietnamese legal texts. To address this issue, we propose a novel approach that leverages large language models to generate high-quality, diverse synthetic queries for Vietnamese legal passages. This synthetic data is then used to pre-train retrieval models, specifically bi-encoder and ColBERT, which are further fine-tuned using contrastive loss with mined hard negatives. Our experiments demonstrate that these enhancements lead to strong improvement in retrieval accuracy, validating the effectiveness of synthetic data and pre-training techniques in overcoming the limitations posed by the lack of large labeled datasets in the Vietnamese legal domain.

Figures

Figures reproduced from arXiv: 2412.00657 by the authors.

Figure 1
Figure 1. Workflow for generating synthetic queries and fine-tuning retrieval models using Vietnamese legal texts. Our methodology is divided into several key stages: Stage 1: Collect legal text data. This initial stage involves the collec￾tion and preprocessing of legal documents. The documents are split into smaller passages suitable for further processing. Stage 2: Generate queries using Llama 3 prompt with legal pas￾sages… view at source ↗
Figure 2
Figure 2. The shortened prompt template we used to generate synthetic queries from legal text passages, with placeholders for input documents and few-shot examples omitted. performance in our testing, to filter out queries whose corresponding passages did not appear in the top 40 relevant results. Additionally, we excluded queries that directly referred to the passage using terms like "quy định này" or "thông tư này". This pr… view at source ↗
Figure 3
Figure 3. Top 20 domains by number of queries in synthetic query dataset. For evaluation, we used 20% of the Legal Zalo 21 dataset and 10,000 test queries from TVPL for in-domain testing. Additionally, we employed the Viet￾namese Wiki Question Answering dataset from the Zalo AI Challenge 2019 for out-of-domain evaluation. 4.2 Model Pre-training and Fine-tuning We used the synthetic dataset for pre-training, with text segmente… view at source ↗

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

30 extracted references · 25 canonical work pages

  1. [1]

    arXiv preprint arXiv:2303.08774 (2023)

    Achiam, J., Adler, S., Agarwal, S., Ahmad, L., Akkaya, I., Aleman, F.L., Almeida, D., Altenschmidt, J., Altman, S., Anadkat, S., et al.: GPT-4 Technical Report. arXiv preprint arXiv:2303.08774 (2023)

    arXiv 2023

  2. [2]

    AI@Meta: Llama 3 Model Card (2024), https://github.com/meta-llama/ llama3/blob/main/MODEL_CARD.md

    work page 2024

  3. [3]

    In: Proceedings of the 45th International ACM SIGIR Conference on Research and Development in Information Retrieval

    Bonifacio, Luiz and Abonizio, Hugo and Fadaee, Marzieh and Nogueira, Rodrigo: Inpars: Unsupervised dataset generation for information retrieval. In: Proceedings of the 45th International ACM SIGIR Conference on Research and Development in Information Retrieval. p. 2387–2392. Association for Computing Machinery (2022)

    work page 2022

  4. [4]

    In: Ku, L.W., Martins, A., Srikumar, V

    Chen, J., Xiao, S., Zhang, P., Luo, K., Lian, D., Liu, Z.: M3-embedding: Multi- linguality, multi-functionality, multi-granularity text embeddings through self- knowledge distillation. In: Ku, L.W., Martins, A., Srikumar, V. (eds.) Findings of the Association for Computational Linguistics: ACL 2024. pp. 2318–2335. As- sociation for Computational Linguisti...

    work page 2024

  5. [5]

    Devlin, J., Chang, M., Lee, K., Toutanova, K.: BERT: Pre-training of Deep Bidi- rectional Transformers for Language Understanding. In: Proceedings of the 2019 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, NAACL-HLT 2019, Minneapolis, MN, USA, June 2-7, 2019, Volume 1 (Long and Short...

    work page 2019

  6. [6]

    In: Proceedings of the 2021 Conference on Empirical Methods in Natural Language Processing, EMNLP 2021

    Gao, L., Callan, J.: Condenser: a Pre-training Architecture for Dense Retrieval. In: Proceedings of the 2021 Conference on Empirical Methods in Natural Language Processing, EMNLP 2021. pp. 981–993. Association for Computational Linguistics (2021)

    work page 2021

  7. [7]

    In: Proceedings of the 60th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), ACL 2022

    Gao, L., Callan, J.: Unsupervised Corpus Aware Language Model Pre-training for Dense Passage Retrieval. In: Proceedings of the 60th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), ACL 2022. pp. 2843–2853. Association for Computational Linguistics (2022)

    work page 2022

  8. [8]

    In: Proceedings of the 2021 Conference on Empirical Methods in Natural Language Processing

    Gao, T., Yao, X., Chen, D.: SimCSE: Simple Contrastive Learning of Sentence Embeddings. In: Proceedings of the 2021 Conference on Empirical Methods in Natural Language Processing. pp. 6894–6910. Association for Computational Lin- guistics (Nov 2021)

    work page 2021

Show all 30 references
  1. [9]

    Gunasekar, S., Zhang, Y., Aneja, J., Mendes, C.C.T., Giorno, A.D., Gopi, S., Java- heripi, M., Kauffmann, P.C., de Rosa, G.H., Saarikivi, O., Salim, A., Shah, S., Behl, H., Wang, X., Bubeck, S., Eldan, R., Kalai, A.T., Lee, Y.T., Li, Y.: Textbooks are all you need (2024),https...

    2024

  2. [10]

    https://huggingface.co/keepitreal/ vietnamese-sbert, online; accessed 18 September 2024

    keepitreal: Vienamese sbert. https://huggingface.co/keepitreal/ vietnamese-sbert, online; accessed 18 September 2024

    2024

  3. [11]

    In: Proceedings of the 43rd Inter- national ACM SIGIR Conference on Research and Development in Information Retrieval

    Khattab, O., Zaharia, M.: ColBERT: Efficient and Effective Passage Search via Contextualized Late Interaction over BERT. In: Proceedings of the 43rd Inter- national ACM SIGIR Conference on Research and Development in Information Retrieval. p. 39–48. Association for Computing M...

    2020

  4. [12]

    In: Pro- ceedings of the 28th International Conference on Computational Linguistics

    Kien, P.M., Nguyen, H.T., Bach, N.X., Tran, V., Le Nguyen, M., Phuong, T.M.: Answering legal questions by learning neural attentive text representation. In: Pro- ceedings of the 28th International Conference on Computational Linguistics. pp. 988–998 (2020)

    2020

  5. [13]

    CoRR abs/2009.10270 (2020)

    Liang, D., Xu, P., Shakeri, S., dos Santos, C.N., Nallapati, R., Huang, Z., Xi- ang, B.: Embedding-based Zero-shot Retrieval through Query Generation. CoRR abs/2009.10270 (2020)

    2020 arXiv

  6. [14]

    How to Train Your Dragon: Diverse Augmentation Towards Generalizable Dense Retrieval

    Lin, S.C., Asai, A., Li, M., Oguz, B., Lin, J., Mehdad, Y., Yih, W.t., Chen, X.: "How to Train Your Dragon: Diverse Augmentation Towards Generalizable Dense Retrieval". In: Findings of the Association for Computational Linguistics: EMNLP

  7. [15]

    Zero-shot Neural Passage Retrieval via Domain-targeted Synthetic Question Generation

    Ma, J., Korotkov, I., Yang, Y., Hall, K., McDonald, R.: "Zero-shot Neural Passage Retrieval via Domain-targeted Synthetic Question Generation". In: Proceedings of the 16th Conference of the European Chapter of the Association for Compu- tational Linguistics: Main Volume. pp. 1...

    2021

  8. [16]

    CoRR abs/2306.02707 (2023)

    Mukherjee, S., Mitra, A., Jawahar, G., Agarwal, S., Palangi, H., Awadallah, A.: Orca: Progressive Learning from Complex Explanation Traces of GPT-4. CoRR abs/2306.02707 (2023)

    2023 arXiv

  9. [17]

    In: Besold, T.R., Bordes, A., d’Avila Garcez, A.S., Wayne, G

    Nguyen, T., Rosenberg, M., Song, X., Gao, J., Tiwary, S., Majumder, R., Deng, L.: MS MARCO: A human generated machine reading comprehension dataset. In: Besold, T.R., Bordes, A., d’Avila Garcez, A.S., Wayne, G. (eds.) Proceedings of the Improving Vietnamese Legal Document Retr...

    2016

  10. [18]

    CoRRabs/2209.14494 (2022)

    Pham, N., Nguyen, H., Do, T.: Multi-stage Information Retrieval for Vietnamese Legal Texts. CoRRabs/2209.14494 (2022)

    2022 arXiv

  11. [19]

    In: Proceedings of the 12th International Symposium on Information and Communication Technology

    PhamDuy,A.,LeThanh,H.:AQuestion-AnsweringSystemforVietnamesePublic Administrative Services. In: Proceedings of the 12th International Symposium on Information and Communication Technology. pp. 85–92 (2023)

    2023

  12. [20]

    arXiv e-prints (Mar 2024)

    Quang Duc, N., Hai Son, L., Nhan, N.D., Dich Nhat Minh, N., Thanh Huong, L., Viet Sang, D.: Towards Comprehensive Vietnamese Retrieval-Augmented Genera- tion and Large Language Models. arXiv e-prints (Mar 2024)

    2024

  13. [21]

    In: Proceedings of the 56th Annual Meeting of the As- sociation for Computational Linguistics (Volume 2: Short Papers)

    Rajpurkar, P., Jia, R., Liang, P.: Know What You Don’t Know: Unanswerable Questions for SQuAD. In: Proceedings of the 56th Annual Meeting of the As- sociation for Computational Linguistics (Volume 2: Short Papers). pp. 784–789. Association for Computational Linguistics (Jul 2018)

    2018

  14. [22]

    Reimers, N., Gurevych, I.: Sentence-BERT: Sentence Embeddings using Siamese BERT-Networks. In: Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing and the 9th International Joint Conference on Natural Language Processing, EMNLP-IJCNLP 2019. p...

    2019

  15. [23]

    In: Proceed- ings of the 2022 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, NAACL 2022

    Santhanam, K., Khattab, O., Saad-Falcon, J., Potts, C., Zaharia, M.: ColBERTv2: Effective and Efficient Retrieval via Lightweight Late Interaction. In: Proceed- ings of the 2022 Conference of the North American Chapter of the Association for Computational Linguistics: Human La...

    2022

  16. [24]

    In: Thirty-fifth Conference on Neural Information Processing Systems Datasets and Benchmarks Track (Round 2) (2021)

    Thakur, N., Reimers, N., Rücklé, A., Srivastava, A., Gurevych, I.: BEIR: A het- erogeneous benchmark for zero-shot evaluation of information retrieval models. In: Thirty-fifth Conference on Neural Information Processing Systems Datasets and Benchmarks Track (Round 2) (2021)

    2021

  17. [25]

    In: 2022 14th International Conference on Knowledge and Systems Engineering (KSE)

    Van, H.N., Nguyen, D., Nguyen, P.M., Le Nguyen, M.: Miko team: Deep learning approach for legal question answering in alqac 2022. In: 2022 14th International Conference on Knowledge and Systems Engineering (KSE). pp. 1–5. IEEE (2022)

    2022

  18. [26]

    In: Bouamor, H., Pino, J., Bali, K

    W, X., Ma, G., Qian, W., Lin, Z., Hu, S.: Query-as-context pre-training for dense passage retrieval. In: Bouamor, H., Pino, J., Bali, K. (eds.) Proceedings of the 2023 Conference on Empirical Methods in Natural Language Processing. pp. 1906–1916. Association for Computational ...

    2023

  19. [27]

    In: Ku, L.W., Martins, A., Srikumar, V

    Wang, L., Yang, N., Huang, X., Yang, L., Majumder, R., Wei, F.: Improving text embeddings with large language models. In: Ku, L.W., Martins, A., Srikumar, V. (eds.) Proceedings of the 62nd Annual Meeting of the Association for Compu- tational Linguistics (Volume 1: Long Papers...

    2024

  20. [28]

    CoRRabs/2402.05672 (2024)

    Wang, L., Yang, N., Huang, X., Yang, L., Majumder, R., Wei, F.: Multilingual E5 Text Embeddings: A Technical Report. CoRRabs/2402.05672 (2024)

    2024 arXiv

  21. [29]

    Wu, Xing and Ma, Guangyuan and Lin, Meng and Lin, Zijia and Wang, Zhongyuan and Hu, Songlin: Contextual masked auto-encoder for dense passage retrieval. In: Proceedings of the Thirty-Seventh AAAI Conference on Artificial Intelligence and Thirty-Fifth Conference on Innovative A...

    2023

  22. [2023]

    6385–6400

    pp. 6385–6400. Association for Computational Linguistics (Dec 2023)

    2023

Pith tools

Reviewed August 12, 2026 · model on record in the stance chip above.