# Self-Supervised Contrastive BERT Fine-tuning for Fusion-based Reviewed-Item Retrieval

arXiv: https://arxiv.org/abs/2308.00762 · HTML: https://mlsanity.com/paper/2308.00762/

## Key facts

- Cited by 6 papers in the map (data through 14 Oct 2026).
- Builds on 14 of its 31 references that are in the map.
- First posted on arXiv 1 Aug 2023.

## Builds on (most cited first, 14 of 14)

- [BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding](https://mlsanity.com/paper/1810.04805/) (2018)
- [Adam: A Method for Stochastic Optimization](https://mlsanity.com/paper/1412.6980/) (2014)
- [Dense Passage Retrieval for Open-Domain Question Answering](https://mlsanity.com/paper/2004.04906/) (2020)
- [MS MARCO: A Human Generated MAchine Reading COmprehension Dataset](https://mlsanity.com/paper/1611.09268/) (2016)
- [Approximate Nearest Neighbor Negative Contrastive Learning for Dense Text Retrieval](https://mlsanity.com/paper/2007.00808/) (2020)
- [Passage Re-ranking with BERT](https://mlsanity.com/paper/1901.04085/) (2019)
- [Pretrained Transformers for Text Ranking: BERT and Beyond](https://mlsanity.com/paper/2010.06467/) (2020)
- [Efficient Natural Language Response Suggestion for Smart Reply](https://mlsanity.com/paper/1705.00652/) (2017)
- [Unsupervised Corpus Aware Language Model Pre-training for Dense Passage Retrieval](https://mlsanity.com/paper/2108.05540/) (2021)
- [Condenser: a Pre-training Architecture for Dense Retrieval](https://mlsanity.com/paper/2104.08253/) (2021)
- [Learning Dense Representations for Entity Retrieval](https://mlsanity.com/paper/1909.10506/) (2019)
- [On Interpretation and Measurement of Soft Attributes for Recommendation](https://mlsanity.com/paper/2105.09179/) (2021)
- [Design Patterns for Fusion-Based Object Retrieval](https://mlsanity.com/paper/1708.08715/) (2017)
- ["It doesn't look good for a date": Transforming Critiques into Preferences for Conversational Recommendation Systems](https://mlsanity.com/paper/2109.07576/) (2021)

## Built on by (most cited first, 6 of 6)

- [A Review of Modern Recommender Systems Using Generative Models (Gen-RecSys)](https://mlsanity.com/paper/2404.00579/) (2024)
- [Retrieval-Augmented Conversational Recommendation with Prompt-based Semi-Structured Natural Language State Tracking](https://mlsanity.com/paper/2406.00033/) (2024)
- [Elaborative Subtopic Query Reformulation for Broad and Indirect Queries in Travel Destination Recommendation](https://mlsanity.com/paper/2410.01598/) (2024)
- [QBR: A Question-Bank-Based Approach to Fine-Grained Legal Knowledge Retrieval for the General Public](https://mlsanity.com/paper/2505.04883/) (2025)
- [Refining Joint Text and Source Code Embeddings for Retrieval Task with Parameter-Efficient Fine-Tuning](https://mlsanity.com/paper/2405.04126/) (2024)
- [Large Language Model Driven Recommendation](https://mlsanity.com/paper/2408.10946/) (2024)

## Abstract (arXiv)

As natural language interfaces enable users to express increasingly complex natural language queries, there is a parallel explosion of user review content that can allow users to better find items such as restaurants, books, or movies that match these expressive queries. While Neural Information Retrieval (IR) methods have provided state-of-the-art results for matching queries to documents, they have not been extended to the task of Reviewed-Item Retrieval (RIR), where query-review scores must be aggregated (or fused) into item-level scores for ranking. In the absence of labeled RIR datasets, we extend Neural IR methodology to RIR by leveraging self-supervised methods for contrastive learning of BERT embeddings for both queries and reviews. Specifically, contrastive learning requires a choice of positive and negative samples, where the unique two-level structure of our item-review data combined with meta-data affords us a rich structure for the selection of these samples. For contrastive learning in a Late Fusion scenario, we investigate the use of positive review samples from the same item and/or with the same rating, selection of hard positive samples by choosing the least similar reviews from the same anchor item, and selection of hard negative samples by choosing the most similar reviews from different items. We also explore anchor sub-sampling and augmenting with meta-data. For a more end-to-end Early Fusion approach, we introduce contrastive item embedding learning to fuse reviews into single item embeddings. Experimental results show that Late Fusion contrastive learning for Neural RIR outperforms all other contrastive IR configurations, Neural IR, and sparse retrieval baselines, thus demonstrating the power of exploiting the two-level structure in Neural RIR approaches as well as the importance of preserving the nuance of individual review content via Late Fusion methods.