---
title: "Hypothetical Document Embeddings (HyDE) (ID)"
source: "https://systems-analysis.info/int/Hypothetical_Document_Embeddings_(HyDE)_(ID)"
wiki: "systems-analysis.info/int"
article: "Hypothetical_Document_Embeddings_(HyDE)_(ID)"
language: "id"
categories:
  - "Category:Indonesian"
  - "Category:Large language models"
  - "Category:Pages with math errors"
  - "Category:Pages with math render errors"
  - "Category:Prompt engineering"
revision_id: 3115
wiki_created_at: 2026-09-06T23:15:55Z
wiki_modified_at: 2026-09-06T23:15:55Z
downloaded_at: 2026-09-07T22:55:15Z
---

# Hypothetical Document Embeddings (HyDE) (ID)

**Hypothetical Document Expansion (HyDE)** — metode peningkatan vector retrieval dan retrieval‑augmented generation (RAG), di mana large language model (LLM) berdasarkan kueri awal menghasilkan «dokumen hipotetis»; kemudian teks tersebut divektorisasi oleh encoder, dan pencarian dilakukan di antara dokumen nyata berdasarkan kedekatan dengan vektor yang diperoleh. Pendekatan ini memungkinkan penggunaan «pola relevansi» yang dikodekan oleh LLM, dan «membumikannya» pada korpus menggunakan dense embedding<sup>[\[1\]](https://systems-analysis.info/int/Hypothetical_Document_Embeddings_(HyDE)_(ID)#cite_note-1)</sup>.

## Definisi dan Intuisi

HyDE mengdekomposisi tugas pencarian menjadi dua tahap:

\(1\) LLM membuat «contoh jawaban yang relevan» (*hypothetical document*) terhadap kueri, sehingga memodelkan ciri-ciri relevansi;

\(2\) encoder kontrastif (mis., Contriever) mengubah teks tersebut menjadi vektor, yang digunakan untuk mengambil dokumen nyata dari indeks. Teks yang dihasilkan mungkin mengandung kesalahan faktual, namun yang penting adalah pola tematik dan terminologis yang ditangkap oleh encoder<sup>[\[2\]](https://systems-analysis.info/int/Hypothetical_Document_Embeddings_(HyDE)_(ID)#cite_note-2)</sup>.

## Sejarah dan Sumber

Gagasan memperluas pencarian dengan teks sintetis berakar pada karya-karya tentang perluasan kueri dan pseudo-relevance feedback (PRF): algoritma Rocchio dan language model relevansi<sup>[\[3\]](https://systems-analysis.info/int/Hypothetical_Document_Embeddings_(HyDE)_(ID)#cite_note-3)[\[4\]](https://systems-analysis.info/int/Hypothetical_Document_Embeddings_(HyDE)_(ID)#cite_note-4)</sup>. Untuk dense retrieval digunakan encoder yang dilatih secara kontrastif (Contriever)<sup>[\[5\]](https://systems-analysis.info/int/Hypothetical_Document_Embeddings_(HyDE)_(ID)#cite_note-5)</sup> dan Dense Passage Retrieval (DPR)<sup>[\[6\]](https://systems-analysis.info/int/Hypothetical_Document_Embeddings_(HyDE)_(ID)#cite_note-6)</sup>. Benchmark BEIR menstandarkan evaluasi zero‑shot<sup>[\[7\]](https://systems-analysis.info/int/Hypothetical_Document_Embeddings_(HyDE)_(ID)#cite_note-7)</sup>. Dalam konteks inilah HyDE diusulkan sebagai cara untuk «membawa» pengetahuan tentang relevansi ke dalam mode zero-shot melalui LLM tanpa fine-tuning encoder<sup>[\[8\]](https://systems-analysis.info/int/Hypothetical_Document_Embeddings_(HyDE)_(ID)#cite_note-8)</sup>.

## Metode dan Formalisasi

Misalkan korpus dokumen $\mathcal{D} = \{ d_{1},\ldots,d_{N}\}$, dan encoder teks $E:\text{text} \rightarrow {\mathbb{R}}^{n}$ mendefinisikan representasi vektor dokumen $\mathbf{v}_{d} = E(d)$. Untuk mengukur kedekatan digunakan cosine similarity atau dot product; catatan penting: \*\*dot product sama dengan cosine similarity hanya jika kedua vektor memiliki L2‑norm satuan\*\* ($\|\mathbf{u}\| = \|\mathbf{v}\| = 1$)<sup>[\[9\]](https://systems-analysis.info/int/Hypothetical_Document_Embeddings_(HyDE)_(ID)#cite_note-9)</sup>.

HyDE mendefinisikan ulang representasi kueri melalui «dokumen hipotetis» yang dihasilkan oleh LLM. Secara formal:

$$
\begin{matrix}
 & \text{(1) Генерация гипотетического текста:} & & {\overset{\sim}{d}\; = \; G\!\left( q;\,{inst} \right),} \\
 & \text{(2) Эмбеддинг гипотетического текста:} & & {\mathbf{v}_{h}\; = \; E(\overset{\sim}{d}),} \\
 & \text{(3) Поиск ближайших соседей:} & & {\mathcal{R}_{k}(q)\; = \;{TopK}_{\, d \in \mathcal{D}}\; S\!\left( \mathbf{v}_{h},\mathbf{v}_{d} \right),}
\end{matrix}
$$

di mana $G$ — LLM dengan instruksi $inst$ (misalnya: «Tulis sebuah paragraf yang menjawab pertanyaan …»), $S$ — ukuran kemiripan (cosine atau IP dengan normalisasi), dan $\mathcal{R}_{k}(q)$ — himpunan dari $k$ dokumen dengan kemiripan maksimum<sup>[\[10\]](https://systems-analysis.info/int/Hypothetical_Document_Embeddings_(HyDE)_(ID)#cite_note-10)[\[11\]](https://systems-analysis.info/int/Hypothetical_Document_Embeddings_(HyDE)_(ID)#cite_note-11)</sup>.

Dalam praktik rekayasa, sering kali dihasilkan \*\*beberapa\*\* teks hipotetis dan representasinya diagregasi, yang meningkatkan stabilitas:

$$
{\overset{\sim}{d}}^{(j)} = G\!\left( q;\,{inst},\xi_{j} \right),\quad\mathbf{v}_{h}\; = \;\frac{1}{m}\sum\limits_{j = 1}^{m}E\!\left( {\overset{\sim}{d}}^{(j)} \right),
$$

di mana $\xi_{j}$ — parameter decoding stokastik (mis., temperature/top‑p). Ensembling semacam ini meningkatkan Recall dengan pertumbuhan latency yang moderat<sup>[\[12\]](https://systems-analysis.info/int/Hypothetical_Document_Embeddings_(HyDE)_(ID)#cite_note-12)</sup>.

### Pipeline Dasar HyDE

    # 1) prompt(query) -> hypothetical_doc
    # 2) embed(hypothetical_doc) -> v_h
    # 3) retrieve(index, v_h, k) -> candidates
    # 4) (optional) rerank(query, candidates) -> topN
    # 5) (для RAG) stuff / map-reduce / refine на topN

### Hubungan dengan Metode Lain (QE, doc2query, PRF)

- **QE (perluasan kueri)** menambahkan term pada kueri; HyDE sebaliknya menghasilkan seluruh «kuasi-dokumen», yang lebih sesuai dengan dense encoder<sup>[\[13\]](https://systems-analysis.info/int/Hypothetical_Document_Embeddings_(HyDE)_(ID)#cite_note-13)</sup>.
- **doc2query / docTTTTTquery** memperluas **dokumen** dengan kueri sintetis sebelum pengindeksan<sup>[\[14\]](https://systems-analysis.info/int/Hypothetical_Document_Embeddings_(HyDE)_(ID)#cite_note-14)[\[15\]](https://systems-analysis.info/int/Hypothetical_Document_Embeddings_(HyDE)_(ID)#cite_note-15)</sup>; HyDE memperluas **kueri** secara langsung (on the fly), tanpa memerlukan re-indexing.
- **PRF** (Rocchio, Relevance LM) memperbarui vektor kueri berdasarkan hasil teratas; HyDE mengekstrak «pola relevansi» langsung dari LLM, lalu «membumikannya» melalui retrieval dari korpus<sup>[\[16\]](https://systems-analysis.info/int/Hypothetical_Document_Embeddings_(HyDE)_(ID)#cite_note-16)</sup>.

## Integrasi dalam RAG dan Reranking

Dalam RAG, HyDE digunakan sebagai tahap pertama retrieval: dokumen hipotetis → embedding → k kandidat. Selanjutnya digunakan reranking: cross-encoder kelas BERT<sup>[\[17\]](https://systems-analysis.info/int/Hypothetical_Document_Embeddings_(HyDE)_(ID)#cite_note-17)</sup> atau late interaction ColBERT<sup>[\[18\]](https://systems-analysis.info/int/Hypothetical_Document_Embeddings_(HyDE)_(ID)#cite_note-18)</sup>. Untuk penggabungan daftar (mis., hybrid BM25+vector), umumnya digunakan RRF (*reciprocal rank fusion*): $\operatorname{RRF}(d) = \sum\limits_{r \in \mathcal{R}}\frac{1}{k + \operatorname{rank}_{r}(d)},\qquad k \approx 60.$ Metode RRF secara konsisten meningkatkan kualitas keseluruhan peringkat gabungan<sup>[\[19\]](https://systems-analysis.info/int/Hypothetical_Document_Embeddings_(HyDE)_(ID)#cite_note-19)</sup>.

## Evaluasi pada Benchmark (BEIR dan lainnya)

Karya asli mengevaluasi HyDE dalam mode zero-shot pada TREC DL'19/20 (pencarian web) dan pada subset koleksi BEIR (Scifact, ArguAna, TREC‑COVID, FiQA, DBPedia, TREC‑NEWS, Climate‑FEVER). Fragmen hasil — *per 2023‑07*:

| Metode                    | DL19                   | DL20                   | Sumber                                                                                                           |
|---------------------------|------------------------|------------------------|------------------------------------------------------------------------------------------------------------------|
| BM25                      | 30.1 / 50.6 / 75.0     | 28.6 / 48.0 / 78.6     | <sup>[\[20\]](https://systems-analysis.info/int/Hypothetical_Document_Embeddings_(HyDE)_(ID)#cite_note-20)</sup> |
| Contriever (unsup.)       | 24.0 / 44.5 / 74.6     | 24.0 / 42.1 / 75.4     | <sup>[\[21\]](https://systems-analysis.info/int/Hypothetical_Document_Embeddings_(HyDE)_(ID)#cite_note-21)</sup> |
| **HyDE** (Contriever+LLM) | **41.8 / 61.3 / 88.0** | **38.2 / 57.9 / 84.4** | <sup>[\[22\]](https://systems-analysis.info/int/Hypothetical_Document_Embeddings_(HyDE)_(ID)#cite_note-22)</sup> |
| DPR (ft)                  | 36.5 / 62.2 / 76.9     | 41.8 / 65.3 / 81.4     | <sup>[\[23\]](https://systems-analysis.info/int/Hypothetical_Document_Embeddings_(HyDE)_(ID)#cite_note-23)</sup> |
| ANCE (ft)                 | 37.1 / 64.5 / 75.5     | 40.8 / 64.6 / 77.6     | <sup>[\[24\]](https://systems-analysis.info/int/Hypothetical_Document_Embeddings_(HyDE)_(ID)#cite_note-24)</sup> |

*TREC DL19/20 (pencarian web)* — mAP / nDCG@10 / Recall@1k

| Metode     | Scifact         | ArguAna         | TREC‑COVID      | FiQA        | DBPedia     | TREC‑NEWS       | Climate‑FEVER   | Sumber                                                                                                           |
|------------|-----------------|-----------------|-----------------|-------------|-------------|-----------------|-----------------|------------------------------------------------------------------------------------------------------------------|
| BM25       | 67.9 / 92.5     | 39.7 / 93.2     | **59.5 / 49.8** | 23.6 / 54.0 | 31.8 / 46.8 | 39.5 / 44.7     | 16.5 / 42.5     | <sup>[\[25\]](https://systems-analysis.info/int/Hypothetical_Document_Embeddings_(HyDE)_(ID)#cite_note-25)</sup> |
| Contriever | 64.9 / 92.6     | 37.9 / 90.1     | 27.3 / 17.2     | 24.5 / 56.2 | 29.2 / 45.3 | 34.8 / 42.3     | 15.5 / 44.1     | <sup>[\[26\]](https://systems-analysis.info/int/Hypothetical_Document_Embeddings_(HyDE)_(ID)#cite_note-26)</sup> |
| **HyDE**   | **69.1 / 96.4** | **46.6 / 97.9** | 59.3 / 41.4     | 27.3 / 62.1 | 36.8 / 47.2 | **44.0 / 50.9** | **22.3 / 53.0** | <sup>[\[27\]](https://systems-analysis.info/int/Hypothetical_Document_Embeddings_(HyDE)_(ID)#cite_note-27)</sup> |

*BEIR (pilihan dataset)* — nDCG@10 / Recall@100

HyDE juga meningkatkan MRR@100 pada dataset multibahasa Mr.TyDi (sw/ko/ja/bn) dibandingkan mContriever<sup>[\[28\]](https://systems-analysis.info/int/Hypothetical_Document_Embeddings_(HyDE)_(ID)#cite_note-28)</sup>.

## Rekomendasi Praktis

Kapan menggunakan HyDE

- Mode zero-shot/transfer (tidak ada label relevansi; «ketidakmiripan» domain dengan korpus pelatihan)<sup>[\[29\]](https://systems-analysis.info/int/Hypothetical_Document_Embeddings_(HyDE)_(ID)#cite_note-29)</sup>.
- Diperlukan peningkatan Recall@k dengan presisi yang dapat diterima — HyDE sering «membuka» area relevan dalam ruang vektor<sup>[\[30\]](https://systems-analysis.info/int/Hypothetical_Document_Embeddings_(HyDE)_(ID)#cite_note-30)</sup>.

Pengaturan tipikal

- **LLM dan prompt**: instruksi «Tulis sebuah paragraf yang menjawab pertanyaan …»; stokastisitas moderat (mis., *temperature*≈0.7)<sup>[\[31\]](https://systems-analysis.info/int/Hypothetical_Document_Embeddings_(HyDE)_(ID)#cite_note-31)</sup>.
- **Jumlah teks hipotetis**: 1–5; rata-rata embedding meningkatkan stabilitas<sup>[\[32\]](https://systems-analysis.info/int/Hypothetical_Document_Embeddings_(HyDE)_(ID)#cite_note-32)</sup>.
- **Embedder**: (m)Contriever tanpa fine-tuning; encoder yang sudah di-fine-tune juga dapat digunakan (efek HyDE tetap terjaga)<sup>[\[33\]](https://systems-analysis.info/int/Hypothetical_Document_Embeddings_(HyDE)_(ID)#cite_note-33)</sup>.
- **Normalisasi embedding**: L2‑norm; dot product ekuivalen dengan cosine<sup>[\[34\]](https://systems-analysis.info/int/Hypothetical_Document_Embeddings_(HyDE)_(ID)#cite_note-34)</sup>.
- **Hybrid retrieval**: BM25+vector dengan reranking berikutnya<sup>[\[35\]](https://systems-analysis.info/int/Hypothetical_Document_Embeddings_(HyDE)_(ID)#cite_note-35)</sup>.
- **Reranker**: Cross-Encoder (BERT re‑ranker)<sup>[\[36\]](https://systems-analysis.info/int/Hypothetical_Document_Embeddings_(HyDE)_(ID)#cite_note-36)</sup> atau ColBERT<sup>[\[37\]](https://systems-analysis.info/int/Hypothetical_Document_Embeddings_(HyDE)_(ID)#cite_note-37)</sup>.
- **Penggabungan** hasil dari berbagai strategi: RRF (*k*≈60)<sup>[\[38\]](https://systems-analysis.info/int/Hypothetical_Document_Embeddings_(HyDE)_(ID)#cite_note-38)</sup>.

Pemantauan kualitas/biaya

- Retrieval: nDCG@k, Recall@k, MRR; end‑to‑end RAG: EM/F1 atau metrik *groundedness* (RAGAS/TruLens)<sup>[\[39\]](https://systems-analysis.info/int/Hypothetical_Document_Embeddings_(HyDE)_(ID)#cite_note-39)[\[40\]](https://systems-analysis.info/int/Hypothetical_Document_Embeddings_(HyDE)_(ID)#cite_note-40)</sup>.
- Biaya/latency: didominasi oleh pembuatan LLM dan (jika ada) reranking; dioptimalkan dengan jumlah «teks hipotetis» dan panjang jawaban<sup>[\[41\]](https://systems-analysis.info/int/Hypothetical_Document_Embeddings_(HyDE)_(ID)#cite_note-41)</sup>.

## Keterbatasan dan Pertanyaan Terbuka

- **Halusinasi** teks hipotetis: LLM dapat memasukkan kesalahan faktual; «pembumian» melalui encoder dan korpus mengurangi risiko, namun tidak sepenuhnya menghilangkannya<sup>[\[42\]](https://systems-analysis.info/int/Hypothetical_Document_Embeddings_(HyDE)_(ID)#cite_note-42)</sup>.
- **Keterbatasan domain/bahasa**: keuntungan HyDE berkurang pada domain yang sangat terspesialisasi dan pada bahasa dengan sumber daya rendah<sup>[\[43\]](https://systems-analysis.info/int/Hypothetical_Document_Embeddings_(HyDE)_(ID)#cite_note-43)</sup>.
- **Latency dan biaya**: pembuatan LLM menambah keterlambatan dan biaya token; kritis untuk skenario online dan «teks hipotetis» yang panjang<sup>[\[44\]](https://systems-analysis.info/int/Hypothetical_Document_Embeddings_(HyDE)_(ID)#cite_note-44)</sup>.
- **Etika dan bias**: sebaiknya menggunakan LLM yang aman dan filter konten<sup>[\[45\]](https://systems-analysis.info/int/Hypothetical_Document_Embeddings_(HyDE)_(ID)#cite_note-45)</sup>.

## Tabel Perbandingan Metode

| Metode                    | Kelas                | Tempat teks dihasilkan                 | Encoder/indeks       | Reranker (tahap ke-2)        | Metrik tipikal (contoh)                            | Biaya/latency                            | Sumber                                                                                                                                                                                                                |
|---------------------------|----------------------|----------------------------------------|----------------------|------------------------------|----------------------------------------------------|------------------------------------------|-----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|
| **HyDE**                  | Query→*hypo‑doc*     | Di sisi kueri (LLM → paragraf)         | (m)Contriever; ANN   | BERT re‑rank / ColBERT / RRF | DL19 nDCG@10≈61.3; DL20≈57.9; ArguAna nDCG@10≈46.6 | \+ pembuatan LLM; + reranking (opsional) | <sup>[\[46\]](https://systems-analysis.info/int/Hypothetical_Document_Embeddings_(HyDE)_(ID)#cite_note-46)</sup>                                                                                                      |
| BM25                      | Leksikal             | —                                      | Indeks terbalik      | Opsional                     | lihat tabel (di atas)                              | Rendah (lexical)                         | <sup>[\[47\]](https://systems-analysis.info/int/Hypothetical_Document_Embeddings_(HyDE)_(ID)#cite_note-47)</sup>                                                                                                      |
| DPR / ANCE                | Dense (ft)           | —                                      | Bi‑encoder; ANN      | Opsional                     | DL19 nDCG@10≈62–65                                 | Sedang (tanpa LLM)                       | <sup>[\[48\]](https://systems-analysis.info/int/Hypothetical_Document_Embeddings_(HyDE)_(ID)#cite_note-48)[\[49\]](https://systems-analysis.info/int/Hypothetical_Document_Embeddings_(HyDE)_(ID)#cite_note-49)</sup> |
| doc2query / docTTTTTquery | Ekspansi dokumen     | Di sisi koleksi (sebelum pengindeksan) | BM25/sparse+expanded | Opsional                     | Peningkatan BM25 pada MS MARCO                     | Pembuatan offline tinggi; online cepat   | <sup>[\[50\]](https://systems-analysis.info/int/Hypothetical_Document_Embeddings_(HyDE)_(ID)#cite_note-50)[\[51\]](https://systems-analysis.info/int/Hypothetical_Document_Embeddings_(HyDE)_(ID)#cite_note-51)</sup> |
| PRF (Rocchio, RLM)        | QE berbasis feedback | Kueri (berdasarkan hasil teratas)      | Apa saja             | Opsional                     | Peningkatan Recall/risiko drift                    | \+ satu putaran retrieval tambahan       | <sup>[\[52\]](https://systems-analysis.info/int/Hypothetical_Document_Embeddings_(HyDE)_(ID)#cite_note-52)</sup>                                                                                                      |

Perbandingan HyDE dan pendekatan terkait

## Lihat Juga

- BM25
- Pencarian berbasis representasi vektor,
- RAG
- Pseudo-relevance feedback
- BEIR

## Daftar Pustaka

- Manning, C. D.; Raghavan, P.; Schütze, H. (2008). *Introduction to Information Retrieval*. Cambridge University Press. ISBN 978‑0521865715.
- Robertson, S.; Zaragoza, H. (2009). *The Probabilistic Relevance Framework: BM25 and Beyond*. Foundations and Trends in IR, 3(4), 333–389. DOI:10.1561/1500000019.

## Tautan

- Repositori HyDE: github.com/texttron/hyde.
- Dokumentasi: Haystack — HyDE: docs.haystack.deepset.ai.
- Dokumentasi: LangChain — HyDE Retriever: docs.langchain.com.

## Catatan

1.  <span id="cite_note-1">[↑](https://systems-analysis.info/int/Hypothetical_Document_Embeddings_(HyDE)_(ID)#cite_ref-1) Gao, L.; Ma, X.; Lin, J.; Callan, J. (2023). ‘‘Precise Zero‑Shot Dense Retrieval without Relevance Labels (HyDE)’’. ACL 2023. pp. 1762–1777. DOI:10.18653/v1/2023.acl-long.99. <a href="https://arxiv.org/abs/2212.10496" class="external text" rel="nofollow">arXiv:2212.10496</a></span>
2.  <span id="cite_note-2">[↑](https://systems-analysis.info/int/Hypothetical_Document_Embeddings_(HyDE)_(ID)#cite_ref-2) Gao, L. et al. (2023). ACL 2023, §3.2. DOI:10.18653/v1/2023.acl-long.99.</span>
3.  <span id="cite_note-3">[↑](https://systems-analysis.info/int/Hypothetical_Document_Embeddings_(HyDE)_(ID)#cite_ref-3) Rocchio, J. (1971). ‘‘Relevance Feedback in Information Retrieval’’. In: Salton, G. (ed.) *The SMART Retrieval System*. Prentice‑Hall, pp. 313–323. ISBN 978‑0138145255.</span>
4.  <span id="cite_note-4">[↑](https://systems-analysis.info/int/Hypothetical_Document_Embeddings_(HyDE)_(ID)#cite_ref-4) Lavrenko, V.; Croft, W. B. (2001). ‘‘Relevance‑Based Language Models’’. SIGIR. DOI:10.1145/383952.383972.</span>
5.  <span id="cite_note-5">[↑](https://systems-analysis.info/int/Hypothetical_Document_Embeddings_(HyDE)_(ID)#cite_ref-5) Izacard, G. et al. (2021/2022). ‘‘Unsupervised Dense Information Retrieval with Contrastive Learning’’. <a href="https://arxiv.org/abs/2112.09118" class="external text" rel="nofollow">arXiv:2112.09118</a>.</span>
6.  <span id="cite_note-6">[↑](https://systems-analysis.info/int/Hypothetical_Document_Embeddings_(HyDE)_(ID)#cite_ref-6) Karpukhin, V. et al. (2020). ‘‘Dense Passage Retrieval for Open‑Domain QA’’. EMNLP. DOI:10.18653/v1/2020.emnlp-main.550.</span>
7.  <span id="cite_note-7">[↑](https://systems-analysis.info/int/Hypothetical_Document_Embeddings_(HyDE)_(ID)#cite_ref-7) Thakur, N. et al. (2021). ‘‘BEIR: A Heterogeneous Benchmark for Zero‑shot Evaluation of Information Retrieval Models’’. NeurIPS Datasets Track. <a href="https://arxiv.org/abs/2104.08663" class="external text" rel="nofollow">arXiv:2104.08663</a>.</span>
8.  <span id="cite_note-8">[↑](https://systems-analysis.info/int/Hypothetical_Document_Embeddings_(HyDE)_(ID)#cite_ref-8) Gao, L. et al. (2023). DOI:10.18653/v1/2023.acl-long.99.</span>
9.  <span id="cite_note-9">[↑](https://systems-analysis.info/int/Hypothetical_Document_Embeddings_(HyDE)_(ID)#cite_ref-9) Milvus Docs. ‘‘Similarity Metrics’’ — при L2‑нормализации векторов внутр. произведение эквивалентно косинусу. URL: <a href="https://milvus.io/docs/v2.2.x/metric.md" class="external free" rel="nofollow">https://milvus.io/docs/v2.2.x/metric.md</a></span>
10. <span id="cite_note-10">[↑](https://systems-analysis.info/int/Hypothetical_Document_Embeddings_(HyDE)_(ID)#cite_ref-10) Gao, L.; Ma, X.; Lin, J.; Callan, J. (2023). ‘‘Precise Zero‑Shot Dense Retrieval without Relevance Labels (HyDE)’’. ACL 2023, §3–4. arXiv:2212.10496. DOI:10.18653/v1/2023.acl-long.99.</span>
11. <span id="cite_note-11">[↑](https://systems-analysis.info/int/Hypothetical_Document_Embeddings_(HyDE)_(ID)#cite_ref-11) Izacard, G. et al. (2021/2022). ‘‘Unsupervised Dense Information Retrieval with Contrastive Learning (Contriever)’’. arXiv:2112.09118.</span>
12. <span id="cite_note-12">[↑](https://systems-analysis.info/int/Hypothetical_Document_Embeddings_(HyDE)_(ID)#cite_ref-12) Gao, L. et al. (2023). Прил. (ablation): влияние числа гипотетических текстов и параметров генерации. arXiv:2212.10496.</span>
13. <span id="cite_note-13">[↑](https://systems-analysis.info/int/Hypothetical_Document_Embeddings_(HyDE)_(ID)#cite_ref-13) Gao, L. et al. (2023). DOI:10.18653/v1/2023.acl-long.99.</span>
14. <span id="cite_note-14">[↑](https://systems-analysis.info/int/Hypothetical_Document_Embeddings_(HyDE)_(ID)#cite_ref-14) Nogueira, R. et al. (2019). ‘‘Document Expansion by Query Prediction’’ (doc2query). <a href="https://arxiv.org/abs/1904.08375" class="external text" rel="nofollow">arXiv:1904.08375</a>.</span>
15. <span id="cite_note-15">[↑](https://systems-analysis.info/int/Hypothetical_Document_Embeddings_(HyDE)_(ID)#cite_ref-15) Nogueira, R.; Lin, J. (2019). ‘‘From doc2query to docTTTTTquery’’ (tech report). <a href="https://cs.uwaterloo.ca/~jimmylin/publications/Nogueira_Lin_2019_docTTTTTquery-v2.pdf" class="external text" rel="nofollow">PDF</a></span>
16. <span id="cite_note-16">[↑](https://systems-analysis.info/int/Hypothetical_Document_Embeddings_(HyDE)_(ID)#cite_ref-16) Rocchio, J. (1971); Lavrenko & Croft (2001), см. выше.</span>
17. <span id="cite_note-17">[↑](https://systems-analysis.info/int/Hypothetical_Document_Embeddings_(HyDE)_(ID)#cite_ref-17) Nogueira, R.; Cho, K. (2019). ‘‘Passage Re‑ranking with BERT’’. <a href="https://arxiv.org/abs/1901.04085" class="external text" rel="nofollow">arXiv:1901.04085</a>.</span>
18. <span id="cite_note-18">[↑](https://systems-analysis.info/int/Hypothetical_Document_Embeddings_(HyDE)_(ID)#cite_ref-18) Khattab, O.; Zaharia, M. (2020). ‘‘ColBERT: Efficient and Effective Passage Search via Contextualized Late Interaction over BERT’’. SIGIR. DOI:10.1145/3397271.3401075; <a href="https://arxiv.org/abs/2004.12832" class="external text" rel="nofollow">arXiv:2004.12832</a>.</span>
19. <span id="cite_note-19">[↑](https://systems-analysis.info/int/Hypothetical_Document_Embeddings_(HyDE)_(ID)#cite_ref-19) Cormack, G. V.; Clarke, C. L. A.; Büttcher, S. (2009). ‘‘Reciprocal Rank Fusion Outperforms Condorcet and Nearly Optimally Combines Rankings’’. SIGIR. DOI:10.1145/1571941.1572114.</span>
20. <span id="cite_note-20">[↑](https://systems-analysis.info/int/Hypothetical_Document_Embeddings_(HyDE)_(ID)#cite_ref-20) Gao, L. et al. (2023). Табл. 1. DOI:10.18653/v1/2023.acl-long.99.</span>
21. <span id="cite_note-21">[↑](https://systems-analysis.info/int/Hypothetical_Document_Embeddings_(HyDE)_(ID)#cite_ref-21) Izacard, G. et al. (2022); сводные метрики — в Gao et al., 2023, табл. 1. arXiv:2112.09118.</span>
22. <span id="cite_note-22">[↑](https://systems-analysis.info/int/Hypothetical_Document_Embeddings_(HyDE)_(ID)#cite_ref-22) Gao, L. et al. (2023). Табл. 1.</span>
23. <span id="cite_note-23">[↑](https://systems-analysis.info/int/Hypothetical_Document_Embeddings_(HyDE)_(ID)#cite_ref-23) Karpukhin, V. et al. (2020); сводные — в Gao et al., 2023.</span>
24. <span id="cite_note-24">[↑](https://systems-analysis.info/int/Hypothetical_Document_Embeddings_(HyDE)_(ID)#cite_ref-24) Xiong, L. et al. (2021). ICLR. <a href="https://arxiv.org/abs/2007.00808" class="external text" rel="nofollow">arXiv:2007.00808</a>.</span>
25. <span id="cite_note-25">[↑](https://systems-analysis.info/int/Hypothetical_Document_Embeddings_(HyDE)_(ID)#cite_ref-25) Thakur, N. et al. (2021); сводные — в Gao et al., 2023, табл. 2. arXiv:2104.08663.</span>
26. <span id="cite_note-26">[↑](https://systems-analysis.info/int/Hypothetical_Document_Embeddings_(HyDE)_(ID)#cite_ref-26) Izacard, G. et al. (2022); сводные — в Gao et al., 2023, табл. 2.</span>
27. <span id="cite_note-27">[↑](https://systems-analysis.info/int/Hypothetical_Document_Embeddings_(HyDE)_(ID)#cite_ref-27) Gao, L. et al. (2023). Табл. 2.</span>
28. <span id="cite_note-28">[↑](https://systems-analysis.info/int/Hypothetical_Document_Embeddings_(HyDE)_(ID)#cite_ref-28) Gao, L. et al. (2023). Табл. 3. DOI:10.18653/v1/2023.acl-long.99.</span>
29. <span id="cite_note-29">[↑](https://systems-analysis.info/int/Hypothetical_Document_Embeddings_(HyDE)_(ID)#cite_ref-29) Gao, L. et al. (2023). §4–5.</span>
30. <span id="cite_note-30">[↑](https://systems-analysis.info/int/Hypothetical_Document_Embeddings_(HyDE)_(ID)#cite_ref-30) Gao, L. et al. (2023). §4.2–4.3.</span>
31. <span id="cite_note-31">[↑](https://systems-analysis.info/int/Hypothetical_Document_Embeddings_(HyDE)_(ID)#cite_ref-31) Gao, L. et al. (2023). §4.1.</span>
32. <span id="cite_note-32">[↑](https://systems-analysis.info/int/Hypothetical_Document_Embeddings_(HyDE)_(ID)#cite_ref-32) Haystack Docs. ‘‘Hypothetical Document Embeddings (HyDE)’’ (инженерная справка). <a href="https://docs.haystack.deepset.ai/docs/hypothetical-document-embeddings-hyde" class="external text" rel="nofollow">docs.haystack.deepset.ai</a></span>
33. <span id="cite_note-33">[↑](https://systems-analysis.info/int/Hypothetical_Document_Embeddings_(HyDE)_(ID)#cite_ref-33) Gao, L. et al. (2023). Табл. 6.</span>
34. <span id="cite_note-34">[↑](https://systems-analysis.info/int/Hypothetical_Document_Embeddings_(HyDE)_(ID)#cite_ref-34) Milvus Docs. ‘‘Similarity Metrics’’.</span>
35. <span id="cite_note-35">[↑](https://systems-analysis.info/int/Hypothetical_Document_Embeddings_(HyDE)_(ID)#cite_ref-35) Haystack × Milvus Integration (официальная док.). <a href="https://haystack.deepset.ai/integrations/milvus-document-store" class="external text" rel="nofollow">haystack.deepset.ai</a></span>
36. <span id="cite_note-36">[↑](https://systems-analysis.info/int/Hypothetical_Document_Embeddings_(HyDE)_(ID)#cite_ref-36) Nogueira, R.; Cho, K. (2019). arXiv:1901.04085.</span>
37. <span id="cite_note-37">[↑](https://systems-analysis.info/int/Hypothetical_Document_Embeddings_(HyDE)_(ID)#cite_ref-37) Khattab, O.; Zaharia, M. (2020). DOI:10.1145/3397271.3401075.</span>
38. <span id="cite_note-38">[↑](https://systems-analysis.info/int/Hypothetical_Document_Embeddings_(HyDE)_(ID)#cite_ref-38) Cormack, G. V. et al. (2009). DOI:10.1145/1571941.1572114.</span>
39. <span id="cite_note-39">[↑](https://systems-analysis.info/int/Hypothetical_Document_Embeddings_(HyDE)_(ID)#cite_ref-39) Manning, C. D.; Raghavan, P.; Schütze, H. (2008). *Introduction to Information Retrieval*. Cambridge Univ. Press. ISBN 978‑0521865715.</span>
40. <span id="cite_note-40">[↑](https://systems-analysis.info/int/Hypothetical_Document_Embeddings_(HyDE)_(ID)#cite_ref-40) Es, S. et al. (2023). ‘‘RAGAS: Automated Evaluation of Retrieval‑Augmented Generation’’. <a href="https://arxiv.org/abs/2309.15217" class="external text" rel="nofollow">arXiv:2309.15217</a>.</span>
41. <span id="cite_note-41">[↑](https://systems-analysis.info/int/Hypothetical_Document_Embeddings_(HyDE)_(ID)#cite_ref-41) Gao, L. et al. (2023). §5.</span>
42. <span id="cite_note-42">[↑](https://systems-analysis.info/int/Hypothetical_Document_Embeddings_(HyDE)_(ID)#cite_ref-42) Gao, L. et al. (2023). §3.2; §4.1. DOI:10.18653/v1/2023.acl-long.99.</span>
43. <span id="cite_note-43">[↑](https://systems-analysis.info/int/Hypothetical_Document_Embeddings_(HyDE)_(ID)#cite_ref-43) Gao, L. et al. (2023). Табл. 3; §4.4.</span>
44. <span id="cite_note-44">[↑](https://systems-analysis.info/int/Hypothetical_Document_Embeddings_(HyDE)_(ID)#cite_ref-44) Gao, L. et al. (2023). §4–5.</span>
45. <span id="cite_note-45">[↑](https://systems-analysis.info/int/Hypothetical_Document_Embeddings_(HyDE)_(ID)#cite_ref-45) Ouyang, L. et al. (2022). ‘‘Training language models to follow instructions with human feedback (InstructGPT)’’. NeurIPS. <a href="https://arxiv.org/abs/2203.02155" class="external text" rel="nofollow">arXiv:2203.02155</a>.</span>
46. <span id="cite_note-46">[↑](https://systems-analysis.info/int/Hypothetical_Document_Embeddings_(HyDE)_(ID)#cite_ref-46) Gao, L. et al. (2023). Табл. 1–2.</span>
47. <span id="cite_note-47">[↑](https://systems-analysis.info/int/Hypothetical_Document_Embeddings_(HyDE)_(ID)#cite_ref-47) Robertson, S.; Zaragoza, H. (2009). ‘‘The Probabilistic Relevance Framework: BM25 and Beyond’’. Found. Trends IR. DOI:10.1561/1500000019.</span>
48. <span id="cite_note-48">[↑](https://systems-analysis.info/int/Hypothetical_Document_Embeddings_(HyDE)_(ID)#cite_ref-48) Karpukhin, V. et al. (2020). DOI:10.18653/v1/2020.emnlp-main.550.</span>
49. <span id="cite_note-49">[↑](https://systems-analysis.info/int/Hypothetical_Document_Embeddings_(HyDE)_(ID)#cite_ref-49) Xiong, L. et al. (2021). <a href="https://arxiv.org/abs/2007.00808" class="external text" rel="nofollow">arXiv:2007.00808</a>.</span>
50. <span id="cite_note-50">[↑](https://systems-analysis.info/int/Hypothetical_Document_Embeddings_(HyDE)_(ID)#cite_ref-50) Nogueira, R. et al. (2019). arXiv:1904.08375.</span>
51. <span id="cite_note-51">[↑](https://systems-analysis.info/int/Hypothetical_Document_Embeddings_(HyDE)_(ID)#cite_ref-51) Nogueira, R.; Lin, J. (2019). tech report.</span>
52. <span id="cite_note-52">[↑](https://systems-analysis.info/int/Hypothetical_Document_Embeddings_(HyDE)_(ID)#cite_ref-52) Rocchio, J. (1971). SMART; Lavrenko & Croft (2001) SIGIR.</span>