An Updated Duet Model for Passage Re-ranking

By Bhaskar Mitra, Nick Craswell. 2019

54 citations recorded by the source corpus.

We propose several small modifications to Duet---a deep neural ranking model---and evaluate the updated model on the MS MARCO passage ranking task. We report significant improvements from the proposed changes based on an ablation study.

Subfields: Artificial Intelligence, Information Retrieval & Search

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In Information Retrieval & Search

Citations in the corpus

ColBERT: Efficient and Effective Passage Search via Contextualized Late Interaction over BERT (2020, background): “In this work, we compare against KNRM [5, 36], Duet [21, 23], Con-vKNRM [5], and fastText+ConvKNRM [11].”
Context-Aware Sentence/Passage Term Importance Estimation For First Stage Retrieval (2019, neutral): “Duet V2 [20] 0.243 +0% 0.245 +2% Conv-KNRM [9] 0.247 +2% 0.247 +3% FastText+Conv-KNRM [14] 0.277 +14% 0.290 +21% BERT Re-Ranker [22] 0.365 +50% 0.359 +50% depths, meaning a ranking from DeepCT-Index provided more relevant passages to a re-ranker.”
Document Expansion by Query Prediction (2019, background): “A new generation of neural ranking models offer solutions to the vocabulary mismatch problem based on continuous word representations and the ability to learn highly non-linear models of relevance; see recent overviews by Onal et al. (2018) and Mitra and Craswell (2019a).”
TREC Deep Learning Track: Reusable Test Collections in the Large Data Regime (2021, methodology): “, Duet [19, 20] or using word embeddings [14]—but does not fall into the “nnlm” category (3) trad: if the run exclusively uses traditional IR methods like BM25 [25] and RM3 [1].”
Conformer-Kernel with Query Term Independence for Document Retrieval (2020, methodology): “We adopt the Duet [Nanni et al., 2017, Mitra and Craswell, 2019b,a] framework wherein the term-document score is a linear combination of outputs from a latent and and an explicit matching models. st,d = w1 · BN(s(latent)t,d ) + w2 · BN(s (explicit) t,d ) + b (7) Where, {w1, w2, b} are learnable…”
Overview of the TREC 2020 Deep Learning Track (2021, methodology): “…[Devlin et al., 2018] or XLNet [Yang et al., 2019b] • nn: if the run employs some form of neural network based approach— e.g. , Duet [Mitra et al., 2017, Mitra and Craswell, 2019] or using word embeddings [Joulin et al., 2016]—but does not fall into the “nnlm” category • trad: if the run…”
Document Re-Ranking Model for Machine-Reading and Comprehension (2020, background): “218 Duet v2 [10] (official baseline) 0.”
Overview of the TREC 2019 deep learning track (2020, methodology): “…[Devlin et al., 2018] or XLNet [Yang et al., 2019b] • nn: if the run employs some form of neural network based approach— e.g. , Duet [Mitra et al., 2017, Mitra and Craswell, 2019] or using word embeddings [Joulin et al., 2016]—but does not fall into the “nnlm” category • trad: if the run…”
Expansion via Prediction of Importance with Contextualization (2020, methodology): “representations at index time to reduce the query-time cost. Interaction-based models combine signals from the query and the document at query time to compute the relevance score [13]. The Duet model [12] aims to achieve low query-time latency by combining signals from both a representation-based an”
Interpretable & Time-Budget-Constrained Contextualization for Re-Ranking (2020, methodology): “In the MatchPyramid model, we set the number of CNN layers to 5, each with kernel size 3 × 3 and 16 convolution channels.”
Incorporating Query Term Independence Assumption for Efficient Retrieval and Ranking using Deep Neural Networks (2019, methodology): “The particular implementation of Duet listed on the leaderboard includes modifications(3) to the original model [Mitra and Craswell, 2019].”
Let's measure run time! Extending the IR replicability infrastructure to include performance aspects (2019, methodology): “In contrast to MatchPyramid, the CNNs are only single layered, focusing on different n-gram sizes and single word-to-word interactions are modeled without a CNN. DUET [12] is a hybrid model applying CNNs to local interactions and single vector representation matching of the query and document.”