Research map / Label-efficient and robust learning / Recognition and long-tail
Asymmetrical Hierarchical Networks with Attentive Interactions for Interpretable Review-Based Recommendation
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| Use asymmetric hierarchical attentive modules for user and item reviews | AHN · 5-core datasets (10 domains) | MSE | Helped |
| Use asymmetric hierarchical attentive modules for user and item reviews | AHN · 10-core datasets (9 domains) | MSE: up to 11.6% (DeepCoNN), 7.0% (D-ATT), 13.8% (MPCN), and 8.4% (HUITA) | Helped |
| Add learned attention to item sentence and review aggregation rather than average-pooling (ablation) | AHN · Videos_Games, Digital_Music, Automotive, Office_Products | MSE: 1.1286 -> 1.1138; 0.8205 -> 0.8172; 0.7506 -> 0.7314; 0.6951 -> 0.6825 | Helped |
| Use asymmetric user co-attention rather than item-style attention (ablation) | AHN · Videos_Games, Digital_Music, Automotive, Office_Products | MSE: 1.1604 -> 1.1138; 0.8246 -> 0.8172; 0.7467 -> 0.7314; 0.6941 -> 0.6825 | Helped |
| Weight affinity entries by item sentence and review attention (ablation) | AHN · Videos_Games, Digital_Music, Automotive, Office_Products | MSE: 1.1363 -> 1.1138; 0.8229 -> 0.8172; 0.7348 -> 0.7314; 0.6936 -> 0.6825 | Helped |
| Use a factorization machine prediction layer instead of a dot product (ablation) | AHN · Videos_Games, Digital_Music, Automotive, Office_Products | MSE: 1.1267 -> 1.1138; 0.8341 -> 0.8172; 0.7723 -> 0.7314; 0.7078 -> 0.6825 | Helped |
| Add gating mechanisms to attention (ablation) | AHN · Videos_Games, Digital_Music, Automotive, Office_Products | MSE: 1.1220 -> 1.1138; 0.8188 -> 0.8172; 0.7385 -> 0.7314; 0.6883 -> 0.6825 | Helped |
| Use hard review selections with Gumbel-Softmax pointers | MPCN · 10-core datasets | MSE | Hurt |
Closest papers by meaning
- Review-Enhanced Hierarchical Contrastive Learning for Recommendation (AAAI 2024 · similarity 0.86)
- ScoreNet: Consistency-driven Framework with Multi-side Information Fusion for Session-based Recommendation (AAAI 2025 · similarity 0.86)
- ContextGNN: Beyond Two-Tower Recommendation Systems (ICLR 2025 · similarity 0.85)
- Semantic Enhanced Heterogeneous Hypergraph Network for Collaborative Filtering (AAAI 2025 · similarity 0.85)
- Coherency Improved Explainable Recommendation via Large Language Model (AAAI 2025 · similarity 0.85)
- SHARP-Distill: A 68× Faster Recommender System with Hypergraph Neural Networks and Language Models (ICML 2025 · similarity 0.84)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.