Pith. sign in

REVIEW 3 major objections 6 minor 40 references

ECLIPSE: Contrastive Dimension Importance Estimation with Pseudo-Irrelevance Feedback for Dense Retrieval

T0 review · 3 major / 6 minor · reviewed 2026-08-11 · deepseek-v4-flash

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

Pith's one-line read A training-free step that subtracts an averaged embedding of low-ranked 'irrelevant' documents from the dimension-importance signal improves dense retrieval across four TREC benchmarks.

desk verdict ECLIPSE is a plausible incremental extension of DIME that is let down by a tuning-on-test protocol; the method deserves a serious look, but the reported gains are not yet established. read the letter →

arxiv 2412.14967 v1 pith:DXZZ4ABJ submitted 2024-12-19 cs.IR

Giulio D'Erasmo , Giovanni Trappolini , Nicola Tonellotto , Fabrizio Silvestri This is my paper · ORCID
classification cs.IR
keywords denseretrievaldimensionimportanceestimationpseudo-relevancefeedbackpseudo-irrelevancedimensionalityreductioncontrastivelearningembeddingdimensionsTRECbenchmarks
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

Dimension IMportance Estimators (DIMEs) improve dense retrieval by scoring embedding dimensions according to how well they align a query with a pseudo-relevant document, but they never look at what makes a document irrelevant. ECLIPSE claims that subtracting the query's alignment with a 'moon' vector—the centroid of the lowest-ranked documents from the initial retrieval list—from the relevant alignment exposes the noisy dimensions and sharpens the dimension filter. The paper reports that this contrastive step gives statistically significant gains over PRF-based and LLM-based DIMEs and over the full-dimensional baseline on three in-domain and one out-of-domain benchmark, with average-precision improvements up to 19.50% over DIME baselines and 22.35% over using all dimensions. It further claims that what makes the negative documents useful is their low relevance score rather than their semantic content, since randomly sampling from the bottom of the list does not hurt. If this holds, ECLIPSE is a training-free plug-in improvement for existing DIME-based retrieval pipelines.

What carries the argument

The load-bearing object is the moon vector $m = \frac{1}{k^{-}}\sum_{i=0}^{k^{-}-1} d_{k-i}$, the average embedding of the bottom $k^{-}$ documents of the initial retrieval list, used in the Eclipse importance function $u_q(i) = \alpha(q_i\cdot s_i) - \beta(q_i\cdot m_i)$ alongside the relevant representative $s$. Rewriting the score as $q \odot (\alpha s - \beta m)$ shows that the moon vector enters through a residual vector: it is subtracted from the relevant signal before the query reweights each dimension. The mechanism does the work of suppressing dimensions that align with low-ranked documents while preserving those shared with the query and the relevant representative.

What would settle it

Run ECLIPSE beside its DIME baseline on a query set where the bottom of the initial ranking has been deliberately filled with relevant or topically related documents, and check whether ECLIPSE's AP and nDCG@10 fall at or below the baseline; a consistent drop would show that the method depends on the tail being truly irrelevant, not merely low-scoring.

Watch

Extended reading notes

Core claim

ECLIPSE's central claim is that query-dependent dimension importance should be estimated contrastively: for each dimension $i$ the score is $u_q(i) = \alpha(q_i\cdot s_i) - \beta(q_i\cdot m_i)$, where $s$ is the relevant representative embedding (the pseudo-relevant centroid for the PRF variant, or the LLM-generated document for the LLM variant) and $m$ is the moon vector, the centroid of the bottom $k^{-}$ documents from an initial ranked list of 1,000 retrieved documents. The term $q_i\cdot m_i$ estimates how strongly that dimension carries irrelevant content, and its subtraction suppresses such dimensions. Rewriting the score as $q \odot (\alpha s - \beta m)$ makes the mechanism explicit: the residual 'Eclipse vector' $\alpha s - \beta m$ is what remains of the relevant signal after the irrelevant centroid has been removed, and the elementwise product with $q$ then highlights the dimensions the query cares about. On DL19, DL20, DL-HARD, and Robust04, with ANCE, Contriever, and TAS-B, both variants beat their DIME counterparts and the full-dimensional baseline, with LLM-Eclipse the strongest; the gains persist at half the original dimensionality. The paper's RQ3 experiment, where randomly sampling negative documents from the lower end of the list performs like the exact bottom-$k^{-}$, supports the paper's stated conclusion that effectiveness comes from the documents' low relevance scores rather than their semantic content.

Load-bearing premise

The method's load-bearing premise is that documents at the bottom of the initial ranked list are genuinely irrelevant to the query, so their average embedding reliably points at noisy dimensions; if the tail contains relevant or topically related documents, the subtraction can suppress useful dimensions and hurt retrieval.

Editorial extensions

If this is right

  • Any DIME-based dense retrieval pipeline can adopt ECLIPSE without retraining the encoder, because the moon vector is built from the same initial ranked list the system already produces.
  • Because the gains persist when only half the embedding dimensions are kept, systems can retain fewer dimensions and still beat the full-dimensional baseline, cutting storage and score-computation cost.
  • LLM-Eclipse is the strongest variant in the paper's experiments, so pairing an LLM-generated relevant document with an irrelevant-document centroid is the variant that benefits most from the proposed contrastive subtraction.
  • The bottom tail of the ranked list functions as a reusable pseudo-negative signal even when its exact composition changes, since random sampling from the tail does not hurt.
  • The method transfers out of domain, with statistically significant gains on Robust04, so the contrastive subtraction is not tied to the MS MARCO training distribution.

Reading between the lines

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

  • A natural extension would treat the residual vector $\alpha s - \beta m$ as a modified query embedding, making the contrastive subtraction available to any downstream scorer rather than only to dimension filtering.
  • If only the low rank of the negative documents matters, then even documents from unrelated domains, or random noise vectors scaled to the embedding distribution, might serve as the moon centroid; this is a testable prediction the paper does not run.
  • For queries whose bottom-ranked documents are actually relevant, or where the top-$k$ list is dominated by one subtopic, the subtraction could suppress useful dimensions, so ambiguous and multi-intent queries are the natural boundary case to test.
  • The mechanism is conceptually close to contrastive losses in representation learning, which suggests the same pseudo-irrelevance subtraction could be tested in late-interaction models, where the per-token signals play the role of dimensions.
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

3 major / 6 minor

Summary. The paper proposes ECLIPSE, a training-free extension of DIME-based dimension importance estimation for dense retrieval. For a query q, it computes a "moon" centroid m of the bottom-ranked documents in an initial retrieval list and defines the importance of dimension i as u_q(i) = α(q_i·s_i) − β(q_i·m_i), where s is the pseudo-relevant ("sun") centroid or an LLM-generated document embedding (Eq. 4). The method is evaluated on TREC DL'19, DL'20, DL-HD, and RB'04 with ANCE, Contriever, and TAS-B, reporting average improvements over DIME baselines and the full-dimensional baseline (up to 19.50% and 22.35% in AP). The paper also studies the effect of the retrieved-list size k and of random sampling of the negative documents.

Significance. If the central empirical claim holds, ECLIPSE is a simple, generic, and potentially valuable enhancement to DIME-style dimension filtering: it adds one contrastive term to an existing importance function without retraining the encoder. The formulation in Eq. (4) is clear and the idea of using low-ranked documents as pseudo-irrelevant feedback is a natural, non-circular extension of pseudo-relevance feedback. The experimental coverage is broad (three dense retrievers, four benchmarks, PRF and LLM variants), and the ablations in RQ2/RQ3 provide useful information about the role of the negative set. However, the reported evaluation has a load-bearing methodological gap: the hyperparameters and the percentage of retained dimensions appear to be selected on the test queries, and the significance claims in the text are not fully supported by the paper's own table markers. These issues must be addressed before the main claim can be accepted.

major comments (3)
  1. [Section 5, Table 1; Section 4 (Hyperparameters)] No validation split is described anywhere in the paper. Section 5 states that Table 1 reports "the best result among varying the percentage of retained dimensions," and Section 4 defines grids for k+, k−, α, and β without specifying how these are chosen. If these choices are made on the same test queries (43, 54, and 50 queries for DL'19, DL'20, and DL-HD, respectively) and the best configuration is then used for paired significance tests, the p-values are invalid and the reported 19.50% and 22.35% improvements can be inflated by selection. Please add a held-out validation split for all hyperparameters and the dimension-budget selection, or report performance across the whole grid with appropriate multiple-testing correction, and re-run significance tests on the single pre-specified configuration.
  2. [Section 5, Table 1 (significance claims)] The text says that "both PRF Eclipse and LLM Eclipse show statistically significant improvement with respect to their DIME counterparts and Baseline," but the table's own superscripts contradict this for several cells. For example, ANCE DL'19 PRF Eclipse AP (0.406) and ANCE DL'20 PRF Eclipse AP (0.408) carry only superscript a (significant vs. Baseline, not vs. PRF DIME), and several other Eclipse cells lack superscript b. The paper should quantify how many of the 24 comparisons are significant and qualify the claim accordingly.
  3. [Section 3, Eq. (3); Section 5, RQ3 (Table 3)] The RQ3 experiment only shows that random sampling from the last 30, 100, or 150 documents is equivalent to using the exact bottom-k− documents; it does not test whether those documents are actually irrelevant to the query. Consequently, the Section 6 conclusion that effectiveness "stems primarily from their low relevance scores, rather than from any specific semantic properties" is not established by the reported experiments. Please add an explicit test with documents of known relevance (e.g., labeled non-relevant or topically unrelated documents) as the negative set, or at least discuss why the current experiment is sufficient to separate relevance from semantic content.
minor comments (6)
  1. [Section 5, Table 2] The claim that "Eclipse consistently outperforms the DIMEs baseline, even with reduced dimensionality" is contradicted by the negative improvements for TAS-B LLM Eclipse on RB'04 (-2.35% AP, -2.04% nDCG@10) and by several non-significant or negative cells for other models; please qualify the claim.
  2. [Section 4 (Hyperparameters)] The grid for α and β is described as "positive values increasing linearly from 0.1 up to 1" but the step size is not specified; please list the exact values used.
  3. [Section 3] The text refers to "Figure 3" when discussing the query "What is an active margin?", but the corresponding figure is labeled Fig. 1 in the manuscript; please fix the cross-references.
  4. [Abstract] The phrase "mAP(AP)" is redundant; please use one metric name consistently.
  5. [Section 5, Table 1 caption] The caption states that superscripts a and b indicate significant improvement over Baseline and standard DIMEs, respectively, but some reported best values carry no superscript (e.g., ANCE DL'20 LLM Eclipse nDCG@10 0.665); please clarify whether an unmarked best value is simply not statistically significant.
  6. [General] No code or data availability statement is included; adding one would improve reproducibility, especially for the LLM-generated pseudo-relevant documents, which are otherwise not exactly reproducible.

Circularity Check

1 steps flagged · score 4.0 of 10

Core dimension-importance formula is not circular, but the headline gains are partly self-confirming because all hyperparameters and the retained-dimension fraction are selected on the same tiny test sets used for the significance tests.

  1. fitted input called prediction [Section 4 (Hyperparameters) and Section 5 (Results for RQ1, Table 1)]
    "We define four primary hyperparameters that influence different aspects of the model’s decision-making process:k+, k−, α, and β. ... We report the best result among varying the percentage of retained dimensions. ... The bold represents the best result for each dataset and metric"

    The reported AP/nDCG improvements are the best values selected over the k+, k-, alpha, beta grid and over the retained-dimension percentage, evaluated on the same TREC DL'19/DL'20/DL-HD/RB'04 test queries used for the paired significance tests; no validation split or separate tuning protocol is described. Reporting 'the best result among varying the percentage of retained dimensions' makes the submitted numbers the maximum over that search dimension, and the p-values are computed only after this selection. The headline 19.50%/22.35% gain is therefore partly a resubstitution optimum on the test collections rather than an independent estimate, so the empirical claim is partially self-confirming. The formula Eq.

full rationale

No definitional circularity is present: u_q(i)=alpha(q_i*s_i)-beta(q_i*m_i) is an original weighted contrastive combination whose inputs (top-document centroid, bottom-document centroid, query embedding) do not encode the retrieval metrics being reported. The cited DIME framework [11] is a published baseline shared with one of the present authors, but it is used as an external comparator and for a baseline hyperparameter choice (k+=1), not as an unverified uniqueness theorem, so it is not load-bearing in a circular sense. The main reduction is methodological: Section 4 specifies hyperparameter ranges without any validation split, and Section 5 reports the best result over the retained-dimension percentage on the same 43/54/50-query test sets, with significance tests applied after that selection. This makes the reported gains partly artifacts of test-set selection rather than out-of-sample predictions, while the core modeling idea remains independent. Score 4 reflects this partial empirical self-confirmation without alleging that Eq. (4) is defined in terms of its own output.

Assumptions & free parameters 5 free parameters · 4 assumptions · 1 invented entities

The method's performance rests on a small set of tuned weights (α, β, k+, k-, dimension-retention percentage) and on assumptions borrowed from the DIME literature, notably the Manifold Clustering Hypothesis and the reliability of bottom-ranked documents as irrelevant. The only new conceptual entity is the moon centroid.

free parameters (5)
  • alpha (α) = tuned over grid 0.1 to 1
    Weight for the relevant representative embedding in Eq. (4). Selected to maximize test-set performance.
  • beta (β) = tuned over grid 0.1 to 1
    Weight for the irrelevant representative embedding in Eq. (4). Selected to maximize test-set performance.
  • k+ (number of top documents for the sun vector) = tuned over {2, ..., 14}
    Number of pseudo-relevant documents averaged to build s in the PRF variant (Section 4).
  • k- (number of bottom documents for the moon vector) = tuned over {2, ..., 6}
    Number of pseudo-irrelevant documents averaged to build m (Section 4).
  • percentage of retained dimensions = selected as best per cell
    The paper reports 'the best result among varying the percentage of retained dimensions' (Section 5), effectively choosing the dimension budget on the test set.
assumptions (4)
  • domain assumption The Manifold Clustering Hypothesis holds: query-relevant documents lie in a low-dimensional query-dependent subspace, so eliminating dimensions is beneficial.
    Adopted from Faggioli et al. [11] to justify dimension pruning; if false, the whole DIME framework and ECLIPSE lose their basis.
  • ad hoc to paper Bottom-ranked documents in the initial retrieval list are reliable pseudo-irrelevant examples.
    The moon vector m is built from the bottom k- documents (Eq. 3). The method assumes these are irrelevant; RQ3 tests semantic dependence but not the correctness of the irrelevance label.
  • domain assumption The dot product between query and feedback embeddings measures per-dimension importance (u_q = q_i * s_i), and relevance decomposes multiplicatively across dimensions.
    This is the DIME scoring assumption inherited from [11], used in Eqs. (1)-(4).
  • ad hoc to paper The Hadamard product with the query and the linear difference of centroids preserves and amplifies relevant signals.
    Eq. (5) rewrites the method as q ⊙ (αs - βm); the claim that this 'amplifies contrast' is a design heuristic, not proven.
invented entities (1)
  • moon vector m (pseudo-irrelevant centroid)
    purpose: Aggregates the bottom-k- retrieved document embeddings to represent irrelevant content to be subtracted.
    A new construct introduced in Eq. (3); it is a mean of selected document vectors and has no external validation beyond the retrieval results reported in this paper.

how reviews work

0 comments
Cite this review

Pith. "Pith review of ECLIPSE: Contrastive Dimension Importance Estimation with Pseudo-Irrelevance Feedback for Dense Retrieval." pith.science (2026). https://pith.science/paper/DXZZ4ABJ

@misc{pith2026241214967,
  author       = {Pith},
  title        = {Pith review of: ECLIPSE: Contrastive Dimension Importance Estimation with Pseudo-Irrelevance Feedback for Dense Retrieval},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/DXZZ4ABJ}},
  note         = {Machine review of arXiv:2412.14967}
}
read the original abstract

Recent advances in Information Retrieval have leveraged high-dimensional embedding spaces to improve the retrieval of relevant documents. Moreover, the Manifold Clustering Hypothesis suggests that despite these high-dimensional representations, documents relevant to a query reside on a lower-dimensional, query-dependent manifold. While this hypothesis has inspired new retrieval methods, existing approaches still face challenges in effectively separating non-relevant information from relevant signals. We propose a novel methodology that addresses these limitations by leveraging information from both relevant and non-relevant documents. Our method, ECLIPSE, computes a centroid based on irrelevant documents as a reference to estimate noisy dimensions present in relevant ones, enhancing retrieval performance. Extensive experiments on three in-domain and one out-of-domain benchmarks demonstrate an average improvement of up to 19.50% (resp. 22.35%) in mAP(AP) and 11.42% (resp. 13.10%) in nDCG@10 w.r.t. the DIME-based baseline (resp. the baseline using all dimensions). Our results pave the way for more robust, pseudo-irrelevance-based retrieval systems in future IR research.

Figures

Figures reproduced from arXiv: 2412.14967 by the authors.

Figure 1
Figure 1. Comparison of retrieval results (Top Relevant document and First FPs [PITH_FULL_IMAGE:figures/full_fig_p006_1.png] view at source ↗
Figure 2
Figure 2. Performance comparison of PRF-Eclipse on DL ’19 (a) and RB ’04 (b) collections, showing AP as the percentage of retained dimensions increases. Dif￾ferent k values represent the cardinality of retrieved document sets Dq. Smaller k (e.g. k=50) includes only highly relevant documents, while larger k (e.g. k = 50, 000) gradually incorporates less relevant documents, affecting retrieval performance. 0.2 0.4 0.6 0.8 1.0 P… view at source ↗
Figure 3
Figure 3. Performance comparison of LLM-Eclipse on DL ’19 (a) and RB ’04 (b) collections, showing AP as the percentage of retained dimensions increases. Dif￾ferent k values represent the cardinality of retrieved document sets Dq. Smaller k (e.g. k=50) includes only highly relevant documents, while larger k (e.g. k = 50, 000) gradually incorporates less relevant documents, affecting retrieval performance [PITH_FULL_IMAGE:figu… view at source ↗

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

40 extracted references · 14 canonical work pages

  1. [1]

    In: Mitkov, R., Angelova, G

    Al Sharou, K., Li, Z., Specia, L.: Towards a better understanding of noise in natural language processing. In: Mitkov, R., Angelova, G. (eds.) Proceedings of Eclipse: Contrastive DIME with Pseudo-Irrelevance Feedback 11 0.2 0.4 0.6 0.8 1.0 Percentage of retained dimensions 0.10 0.20 0.30 0.40AP ANCE 0.2 0.4 0.6 0.8 1.0 Percentage of retained dimensions 0....

    work page 1901

  2. [2]

    Ba, S..: Discovering topics with neural topic models built from plsa assumptions (2019), https://arxiv.org/abs/1911.10924

    work page Pith review arXiv 2019

  3. [3]

    org/abs/1611.09268

    Bajaj, P., Campos, D., Craswell, N., Deng, L., Gao, J., et al., X.L.: Ms marco: A human generated machine reading comprehension dataset (2018),https://arxiv. org/abs/1611.09268

    arXiv 2018

  4. [4]

    IEEE Transactions on Pattern Analysis and Machine Intelligence 35(8), 1798–1828 (2013).https://doi.org/10.1109/TPAMI.2013.50

    Bengio, Y., Courville, A., Vincent, P.: Representation learning: A review and new perspectives. IEEE Transactions on Pattern Analysis and Machine Intelligence 35(8), 1798–1828 (2013).https://doi.org/10.1109/TPAMI.2013.50

    doi:10.1109/tpami.2013.50 2013

  5. [5]

    In: The 23rd International Conference on Information Integration and Web Intelligence

    Buzaaba, H., Amagasa, T.: A scheme for efficient question answering with low dimension reconstructed embeddings. In: The 23rd International Conference on Information Integration and Web Intelligence. p. 303–310. iiWAS2021, Association for Computing Machinery, New York, NY, USA (2022). https://doi.org/10. 1145/3487664.3487705, https://doi.org/10.1145/34876...

    work page arXiv 2022

  6. [6]

    Chen, T., Kornblith, S., Norouzi, M., Hinton, G.: A simple framework for con- trastive learning of visual representations (2020),https://arxiv.org/abs/2002. 05709

    work page 2020

  7. [7]

    Craswell, N., Mitra, B., Yilmaz, E., Campos, D.: Overview of the trec 2020 deep learning track (2021),https://arxiv.org/abs/2102.07662

    arXiv 2021

  8. [8]

    Craswell, N., Mitra, B., Yilmaz, E., Campos, D., Voorhees, E.M.: Overview of the trec 2019 deep learning track (2020),https://arxiv.org/abs/2003.07820

    arXiv 2020

Show all 40 references
  1. [9]

    In: European Conference on Information Retrieval

    Datta, S., Ganguly, D., MacAvaney, S., Greene, D.: A deep learning approach for selective relevance feedback. In: European Conference on Information Retrieval. pp. 189–204. Springer (2024)

    2024

  2. [10]

    In: Burstein, J., Doran, C., Solorio, T

    Devlin, J., Chang, M.W., Lee, K., Toutanova, K.: BERT: Pre-training of deep bidirectional transformers for language understanding. In: Burstein, J., Doran, C., Solorio, T. (eds.) Proceedings of the 2019 Conference of the North Amer- ican Chapter of the Association for Computat...

    2019 doi

  3. [11]

    In: Proceedings of the 47th International ACM SIGIR Conference on Research and Development in Information Retrieval

    Faggioli, G., Ferro, N., Perego, R., Tonellotto, N.: Dimension importance esti- mation for dense information retrieval. In: Proceedings of the 47th International ACM SIGIR Conference on Research and Development in Information Retrieval. p. 1318–1328. SIGIR ’24, Association for...

    2024

  4. [12]

    Goodfellow, I.J., Shlens, J., Szegedy, C.: Explaining and harnessing adversarial examples (2015), https://arxiv.org/abs/1412.6572

    2015 arXiv

  5. [13]

    Hofstätter, S., Lin, S.C., Yang, J.H., Lin, J., Hanbury, A.: Efficiently teaching an effective dense retriever with balanced topic aware sampling (2021),https: //arxiv.org/abs/2104.06967

    2021 arXiv

  6. [14]

    Scandinavian Journal of Statistics6(2), 65–70 (1979),http://www.jstor.org/stable/4615733

    Holm, S.: A simple sequentially rejective multiple test procedure. Scandinavian Journal of Statistics6(2), 65–70 (1979),http://www.jstor.org/stable/4615733

    1979

  7. [15]

    arXiv preprint arXiv:2112.09118 (2021) 14 D’Erasmo et al

    Izacard, G., Caron, M., Hosseini, L., Riedel, S., Bojanowski, P., Joulin, A., Grave, E.: Unsupervised dense information retrieval with contrastive learning. arXiv preprint arXiv:2112.09118 (2021) 14 D’Erasmo et al

    2021 arXiv

  8. [16]

    In: 2015 38th International Convention on Information and Commu- nication Technology, Electronics and Microelectronics (MIPRO)

    Jović, A., Brkić, K., Bogunović, N.: A review of feature selection methods with applications. In: 2015 38th International Convention on Information and Commu- nication Technology, Electronics and Microelectronics (MIPRO). pp. 1200–1205 (2015). https://doi.org/10.1109/MIPRO.201...

    2015

  9. [17]

    In: Webber, B., Cohn, T., He, Y., Liu, Y

    Karpukhin, V., Oguz, B., Min, S., Lewis, P., Wu, L., Edunov, S., Chen, D., Yih, W.t.: Dense passage retrieval for open-domain question answering. In: Webber, B., Cohn, T., He, Y., Liu, Y. (eds.) Proceedings of the 2020 Conference on Empirical Methods in Natural Language Proces...

    2020

  10. [18]

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

    Khattab, O., Zaharia, M.: Colbert: Efficient and effective passage search via con- textualized late interaction over bert. In: Proceedings of the 43rd International ACM SIGIR Conference on Research and Development in Information Retrieval. p. 39–48. SIGIR ’20, Association for ...

    2020

  11. [19]

    ACM Comput

    Li, J., Cheng, K., Wang, S., Morstatter, F., Trevino, R.P., Tang, J., Liu, H.: Feature selection: A data perspective. ACM Comput. Surv.50(6) (Dec 2017). https:// doi.org/10.1145/3136625, https://doi.org/10.1145/3136625

    2017 doi

  12. [20]

    IEEE Transactions on Knowledge and Data Engineering32(8), 1475–1488 (2020)

    Li, W., Zhang, Y., Sun, Y., Wang, W., Li, M., Zhang, W., Lin, X.: Approxi- mate nearest neighbor search on high dimensional data — experiments, analyses, and improvement. IEEE Transactions on Knowledge and Data Engineering32(8), 1475–1488 (2020). https://doi.org/10.1109/TKDE.2...

    2020

  13. [21]

    In: Goldberg, Y., Kozareva, Z., Zhang, Y

    Liu, Z., Zhang, H., Xiong, C., Liu, Z., Gu, Y., Li, X.: Dimension reduction for ef- ficient dense retrieval via conditional autoencoder. In: Goldberg, Y., Kozareva, Z., Zhang, Y. (eds.) Proceedings of the 2022 Conference on Empirical Methods in Nat- ural Language Processing. p...

    2022

  14. [22]

    Transactions of the Association for Computa- tional Linguistics9, 329–345 (04 2021).https://doi.org/10.1162/tacl_a_00369, https://doi.org/10.1162/tacl_a_00369

    Luan,Y.,Eisenstein,J.,Toutanova,K.,Collins, M.:Sparse,Dense,andAttentional Representations for Text Retrieval. Transactions of the Association for Computa- tional Linguistics9, 329–345 (04 2021).https://doi.org/10.1162/tacl_a_00369, https://doi.org/10.1162/tacl_a_00369

    2021 doi

  15. [23]

    In: Proceed- ings of the 18th ACM Conference on Information and Knowledge Management

    Lv, Y., Zhai, C.: Adaptive relevance feedback in information retrieval. In: Proceed- ings of the 18th ACM Conference on Information and Knowledge Management. p. 255–264. CIKM ’09, Association for Computing Machinery, New York, NY, USA(2009). https://doi.org/10.1145/1645953.164...

    2009

  16. [24]

    In: Moens, M.F., Huang, X., Specia, L., Yih, S.W.t

    Ma, X., Li, M., Sun, K., Xin, J., Lin, J.: Simple and effective unsupervised re- dundancy elimination to compress dense vectors for passage retrieval. In: Moens, M.F., Huang, X., Specia, L., Yih, S.W.t. (eds.) Proceedings of the 2021 Con- ference on Empirical Methods in Natura...

    2021 doi

  17. [25]

    In: Proceedings of the 44th International ACM SIGIR Conference on Research and Development in Information Retrieval

    Mackie, I., Dalton, J., Yates, A.: How deep is your learning: the dl-hard an- notated deep learning dataset. In: Proceedings of the 44th International ACM SIGIR Conference on Research and Development in Information Retrieval. p. 2335–2341. SIGIR ’21, Association for Computing ...

    2021

  18. [26]

    Mikolov, T., Chen, K., Corrado, G., Dean, J.: Efficient estimation of word repre- sentations in vector space (2013),https://arxiv.org/abs/1301.3781

    2013 arXiv

  19. [27]

    Nguyen, K.A., im Walde, S.S., Vu, N.T.: Neural-based noise filtering from word embeddings (2016), https://arxiv.org/abs/1610.01874

    2016 arXiv

  20. [28]

    OpenAI, Achiam, J., Adler, S., Agarwal, S., Ahmad, L., Akkaya, I., et al.: Gpt-4 technical report (2024),https://arxiv.org/abs/2303.08774

    2024 arXiv

  21. [29]

    Radford, A., Narasimhan, K., Salimans, T., Sutskever, I.: Improving language un- derstanding by generative pre-training. Tech. rep., OpenAI (2018)

    2018

  22. [30]

    arXiv preprint arXiv:2309.05922 (2023)

    Rawte,V.,Sheth,A.,Das,A.:Asurveyofhallucinationinlargefoundationmodels. arXiv preprint arXiv:2309.05922 (2023)

    2023 arXiv

  23. [31]

    arXiv preprint arXiv:1908.10084 (2019)

    Reimers, N.: Sentence-bert: Sentence embeddings using siamese bert-networks. arXiv preprint arXiv:1908.10084 (2019)

    2019 arXiv

  24. [32]

    Robertson, S., Zaragoza, H.: The probabilistic relevance framework: Bm25 and beyond. Found. Trends Inf. Retr.3(4), 333–389 (Apr 2009). https://doi.org/ 10.1561/1500000019, https://doi.org/10.1561/1500000019

    2009 doi

  25. [33]

    Prentice Hall, Engle- wood Cliffs, New Jersey (1971)

    Rocchio, J.: Relevance Feedback in Information Retrieval. Prentice Hall, Engle- wood Cliffs, New Jersey (1971)

    1971

  26. [34]

    Biometrika 52(3/4), 591–611 (1965), http://www.jstor.org/stable/ 2333709

    Shapiro, S.S., Wilk, M.B.: An analysis of variance test for normality (complete samples). Biometrika 52(3/4), 591–611 (1965), http://www.jstor.org/stable/ 2333709

    1965

  27. [35]

    Biometrika6(1), 1–25 (1908),http://www

    Student: The probable error of a mean. Biometrika6(1), 1–25 (1908),http://www. jstor.org/stable/2331554

    1908

  28. [36]

    In: Proceedings of the Thirteenth Text REtrieval Conference (TREC 2004)

    Voorhees, E.M.: Overview of the trec 2004 robust track. In: Proceedings of the Thirteenth Text REtrieval Conference (TREC 2004). NIST Special Publication 500-261, National Institute of Standards and Technology (NIST), Gaithersburg, MD (2004)

    2004

  29. [37]

    Biometrics Bulletin 1(6), 80–83 (1945),http://www.jstor.org/stable/3001968

    Wilcoxon, F.: Individual comparisons by ranking methods. Biometrics Bulletin 1(6), 80–83 (1945),http://www.jstor.org/stable/3001968

    1945

  30. [38]

    In: International Conference on Learning Representations (2021), https://openreview.net/forum?id=zeFrfgyZln

    Xiong, L., Xiong, C., Li, Y., Tang, K.F., Liu, J., Bennett, P.N., Ahmed, J., Over- wijk, A.: Approximate nearest neighbor negative contrastive learning for dense text retrieval. In: International Conference on Learning Representations (2021), https://openreview.net/forum?id=zeFrfgyZln

    2021

  31. [39]

    ACM Trans

    Xu, J., Croft, W.B.: Improving the effectiveness of information retrieval with local context analysis. ACM Trans. Inf. Syst.18(1), 79–112 (Jan 2000).https://doi. org/10.1145/333135.333138, https://doi.org/10.1145/333135.333138

  32. [40]

    Yin, Z., Shen, Y.: On the dimensionality of word embedding (2018), https:// arxiv.org/abs/1812.04224

    2018 arXiv

Pith tools

Reviewed August 11, 2026 · model on record in the stance chip above.