# Towards Bridging Review Sparsity in Recommendation with Textual Edge Graph Representation

arXiv: https://arxiv.org/abs/2508.01128 · HTML: https://mlsanity.com/paper/2508.01128/

## Key facts

- Cited by 1 papers in the map (data through 14 Oct 2026).
- Builds on 26 of its 59 references that are in the map.
- First posted on arXiv 2 Aug 2025.

## Builds on (most cited first, 26 of 26)

- [Sentence-BERT: Sentence Embeddings using Siamese BERT-Networks](https://mlsanity.com/paper/1908.10084/) (2019)
- [BERTScore: Evaluating Text Generation with BERT](https://mlsanity.com/paper/1904.09675/) (2019)
- [Variational Graph Auto-Encoders](https://mlsanity.com/paper/1611.07308/) (2016)
- [LightGCN: Simplifying and Powering Graph Convolution Network for Recommendation](https://mlsanity.com/paper/2002.02126/) (2020)
- [Neural Graph Collaborative Filtering](https://mlsanity.com/paper/1905.08108/) (2019)
- [GAIN: Missing Data Imputation using Generative Adversarial Nets](https://mlsanity.com/paper/1806.02920/) (2018)
- [Text Embeddings Reveal (Almost) As Much As Text](https://mlsanity.com/paper/2310.06816/) (2023)
- [Handling Missing Data with Graph Representation Learning](https://mlsanity.com/paper/2010.16418/) (2020)
- [Matrix Completion and Low-Rank SVD via Fast Alternating Least Squares](https://mlsanity.com/paper/1410.2596/) (2014)
- [UltraGCN: Ultra Simplification of Graph Convolutional Networks for Recommendation](https://mlsanity.com/paper/2110.15114/) (2021)
- [Fine-Grained Spoiler Detection from Large-Scale Review Corpora](https://mlsanity.com/paper/1905.13416/) (2019)
- [A Survey on Data Augmentation in Large Model Era](https://mlsanity.com/paper/2401.15422/) (2024)
- [Data Augmentation using Large Language Models: Data Perspectives, Learning Paradigms and Challenges](https://mlsanity.com/paper/2403.02990/) (2024)
- [Edgeformers: Graph-Empowered Transformers for Representation Learning on Textual-Edge Networks](https://mlsanity.com/paper/2302.11050/) (2023)
- [Disentangling Semantics and Syntax in Sentence Embeddings with Pre-trained Language Models](https://mlsanity.com/paper/2104.05115/) (2021)
- [Tensor Factorization via Matrix Factorization](https://mlsanity.com/paper/1501.07320/) (2015)
- [Action Sequence Augmentation for Early Graph-based Anomaly Detection](https://mlsanity.com/paper/2010.10016/) (2020)
- [WGCN: Graph Convolutional Networks with Weighted Structural Features](https://mlsanity.com/paper/2104.14060/) (2021)
- [Large Language Model with Graph Convolution for Recommendation](https://mlsanity.com/paper/2402.08859/) (2024)
- [Position: Text Embeddings Should Capture Implicit Semantics, Not Just Surface Meaning](https://mlsanity.com/paper/2506.08354/) (2025)
- [Identifying Helpful Sentences in Product Reviews](https://mlsanity.com/paper/2104.09792/) (2021)
- [TEG-DB: A Comprehensive Dataset and Benchmark of Textual-Edge Graphs](https://mlsanity.com/paper/2406.10310/) (2024)
- [Shilling Black-box Review-based Recommender Systems through Fake Review Generation](https://mlsanity.com/paper/2306.16526/) (2023)
- [Establishing strong imputation performance of a denoising autoencoder in a wide range of missing data problems](https://mlsanity.com/paper/2004.02584/) (2020)
- [Bias-Variance Tradeoff of Graph Laplacian Regularizer](https://mlsanity.com/paper/1706.00544/) (2017)
- [Link Prediction on Textual Edge Graphs](https://mlsanity.com/paper/2405.16606/) (2024)

## Built on by (most cited first, 1 of 1)

- [Reasoning-Based Personalized Generation for Users with Sparse Data](https://mlsanity.com/paper/2602.21219/) (2026)

## Abstract (arXiv)

Textual reviews enrich recommender systems with fine-grained preference signals and enhanced explainability. However, in real-world scenarios, users rarely leave reviews, resulting in severe sparsity that undermines the effectiveness of existing models. A natural solution is to impute or generate missing reviews to enrich the data. However, conventional imputation techniques -- such as matrix completion and LLM-based augmentation -- either lose contextualized semantics by embedding texts into vectors, or overlook structural dependencies among user-item interactions. To address these shortcomings, we propose TWISTER (ToWards Imputation on Sparsity with Textual Edge Graph Representation), a unified framework that imputes missing reviews by jointly modeling semantic and structural signals. Specifically, we represent user-item interactions as a Textual-Edge Graph (TEG), treating reviews as edge attributes. To capture relational context, we construct line-graph views and employ a large language model as a graph-aware aggregator. For each interaction lacking a textual review, our model aggregates the neighborhood's natural-language representations to generate a coherent and personalized review. Experiments on the Amazon and Goodreads datasets show that TWISTER consistently outperforms traditional numeric, graph-based, and LLM baselines, delivering higher-quality imputed reviews and, more importantly, enhanced recommendation performance. In summary, TWISTER generates reviews that are more helpful, authentic, and specific, while smoothing structural signals for improved recommendations.