ColBERT-PRF: Semantic Pseudo-Relevance Feedback for Dense Passage and Document Retrieval

Abstract

Pseudo-relevance feedback mechanisms, from Rocchio to the relevance models, have shown the usefulness of expanding and reweighting the users’ initial queries using information occurring in an initial set of retrieved documents, known as the pseudo-relevant set. Recently, dense retrieval – through the use of neural contextual language models such as BERT for analysing the documents’ and queries’ contents and computing their relevance scores – has shown a promising performance on several information retrieval tasks still relying on the traditional inverted index for identifying documents relevant to a query. Two different dense retrieval families have emerged: the use of single embedded representations for each passage and query, e.g., using BERT’s [CLS] token, or via multiple representations, e.g., using an embedding for each token of the query and document (exemplified by ColBERT). In this work, we conduct the first study into the potential for multiple representation dense retrieval to be enhanced using pseudo-relevance feedback and present our proposed approach ColBERT-PRF. In particular, based on the pseudo-relevant set of documents identified using a first-pass dense retrieval, ColBERT-PRF extracts the representative feedback embeddings from the document embeddings of the pseudo-relevant set. Among the representative feedback embeddings, the embeddings that most highly discriminate among documents are employed as the expansion embeddings, which are then added to the original query representation. We show that these additional expansion embeddings both enhance the effectiveness of a reranking of the initial query results as well as an additional dense retrieval operation. Indeed, experiments on the MSMARCO passage ranking dataset show that MAP can be improved by upto 26% on the TREC 2019 query set and 10% on the TREC 2020 query set by the application of our proposed ColBERT-PRF method on a ColBERT dense retrieval approach. We further validate the effectiveness of our proposed pseudo-relevance feedback technique for a dense retrieval model on MSMARCO document ranking and TREC Robust04 document ranking tasks. For instance, ColBERT-PRF exhibits upto 21% and 14% improvement in MAP over the ColBERT E2E model on the MSMARCO document ranking TREC 2019 and TREC 2020 query sets, respectively. Additionally, we study the effectiveness of variants of the ColBERT-PRF model with different weighting methods. Finally, we show that ColBERT-PRF can be made more efficient, attaining upto 4.54x speedup over the default ColBERT-PRF model, and with little impact on effectiveness, through the application of approximate scoring and different clustering methods.

Pemrosesan Query dan Pemeringkatan Hasil dalam Information Retrieval: Sebuah Kajian Literatur

“…f. Evaluasi Pada dasarnya, ada dua kriteria berbeda untuk menilai kualitas IRS. Menurut [19], yang pertama adalah Precission. Kemampuan IRS untuk mengembalikan dokumen (informasi) terkait, serta keakuratan dan ketepatan dari dokumen yang diambil, seperti yang ditunjukkan oleh persamaan (2).…”

“…Recall merupakan proporsi dokumen yang telah ditemukan dan terkait dengan kueri. membantu menghitung metrik IR lainnya yaitu F dengan persamaan berikut[19]…”

Learn to be Fair without Labels: A Distribution-based Learning Framework for Fair Ranking

“…It is worth exploring more complex retrieval models that bring different re-ranking candidates. However, given the GPU memory limitation, we could not replicate the retrieval model, ColBERT-PRF [28], used by one of the TREC participants [14] that produces the best TREC evaluation scores. We leave this part as one direction of our feature work.…”

A Purely Entity-Based Semantic Search Approach for Document Retrieval

“…Moreover, Gao and Callan (2022) [25] proposed the MORES+ model, which is a re-ranking method, and tested it on two classical IR collections: Robust04 and ClueWeb09. Wang et al (2023) [28] proposed a ranking method named ColBERT-PRF. They evaluated it on the MSMARCO and TREC Robust collections for document ranking tasks.…”