--- title: "REG — paper by Wang et al. (2026) — summary and ideas" description: "Retrieval-augmented empathetic chatbots usually pick example dialogues by semantic closeness. 2 references." authors: ["Xu Wang", "Bo Wang", "Yang Xiang", "Yihong Tang", "Dongming Zhao", "Yuzifei", "Yuexian Hou"] year: 2026 venue: "Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)" source: "https://aclanthology.org/2026.acl-long.860/" doi: "10.18653/v1/2026.acl-long.860" licence: "CC BY 4.0 (ACL Anthology)" licence_url: "https://creativecommons.org/licenses/by/4.0/" summary_licence: "Elamur Cosmic License" canonical: "https://research.elamur.ai/free/en/papers/wang-2026-reg-retrieval-emotion-similarity-guiding/" --- [Elamur Research](https://md.research.elamur.ai/free/en/) · [this page in HTML](https://research.elamur.ai/free/en/papers/wang-2026-reg-retrieval-emotion-similarity-guiding/) [home](https://md.research.elamur.ai/free/en/) / [papers](https://md.research.elamur.ai/free/en/papers/) / [2026](https://md.research.elamur.ai/free/en/papers/2026/page/41/) / REG: Retrieval via Emotion Similarity for Guiding Empathetic Dialogue Generation # REG: Retrieval via Emotion Similarity for Guiding Empathetic Dialogue Generation **Authors:** [Xu Wang](https://md.research.elamur.ai/free/en/authors/xu-wang/), [Bo Wang](https://md.research.elamur.ai/free/en/authors/bo-wang/), [Yang Xiang](https://md.research.elamur.ai/free/en/authors/yang-xiang/), Yihong Tang, Dongming Zhao, Yuzifei and Yuexian Hou · **Year:** 2026 · **Venue:** [Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)](https://md.research.elamur.ai/free/en/venues/acl/) ([ACL 2026](https://md.research.elamur.ai/free/en/venues/acl/2026/)) · **Publisher:** [Association for Computational Linguistics](https://md.research.elamur.ai/free/en/publishers/association-for-computational-linguistics/) · **Licence:** [CC BY 4.0 (ACL Anthology)](https://creativecommons.org/licenses/by/4.0/) **Links:** [Publisher's page](https://aclanthology.org/2026.acl-long.860/) · [PDF](https://aclanthology.org/2026.acl-long.860.pdf) · [DOI 10.18653/v1/2026.acl-long.860](https://doi.org/10.18653/v1/2026.acl-long.860) · [OpenAlex](https://openalex.org/W7166880465) **Citations:** 0 citations (OpenAlex, 2026-10-14) ## Summary Retrieval-augmented empathetic chatbots usually pick example dialogues by semantic closeness. They can therefore pull in exemplars whose emotional tone contradicts the user's situation, such as a gloomy rain story for a cheerful one, which pushes the generator toward unsuitable replies. REG makes retrieval emotion-aware. Each context–response pair is tagged with four Emotion Attributes, binned into Low/Mid/High levels: - Sentiment, scored with VADER - Emotional Presence, Interpretation and Exploration, each predicted by a separate classifier Training has two stages. First, an initial generator G1 learns to respond conditioned on the context plus these attributes. Second, a retriever is trained by distillation from G1, with two additions: - **Token-level Retrieval** uses token-level scores of G1's outputs (e.g., BLEU) as reward signals, which reduces noise from the coarse labels. - **Retrieval Candidate Augmentation** injects randomized candidates, which increases exemplar diversity and helps avoid local optima. The retriever and a stronger generator G2 are then trained jointly. On EmpatheticDialogues, REG retrieves emotionally aligned exemplars much more often than earlier retrievers. Human judges also prefer its responses over strong baselines, including a fine-tuned Llama-3.1. This shows that emotional matching, not only topical matching, matters for empathetic generation. ## Key points - The four Emotion Attributes are Sentiment, Emotional Presence, Interpretation and Exploration. They are drawn from prior empathy research (Sharma et al.; Watson et al.) and fed to the generator as auxiliary embeddings. - The retriever is optimized by minimizing the KL divergence between its retrieval-score distribution and a distribution derived from G1's token-level metrics. - Exemplar alignment was checked on 3,000 samples. REG retrieved aligned exemplars in 59.5% of cases, compared with 38.0% for LEMPEx, 36.6% for a fixed retriever and 8.6% for plain RAG. - In human pairwise evaluation, REG won on Empathy against Llama-3.1-8B (SFT) by 58.0% vs 32.5%. Against CASE it won 56.3% vs 31.0%. Annotator agreement (κ) was about 0.47–0.57. - Baselines included Transformer, KEMP, CASE, EmpCRL, PECER, LEMPEx, Exemplary-empathy and Llama-3.1 (SFT). All experiments used the EmpatheticDialogues dataset. *Our summary of the paper, in our own words — not its text. The work belongs to its authors; read it at the links above.* 📚 Shelves (6): [🖥️ Computer Science](https://md.research.elamur.ai/free/en/shelves/computer-science/) · 16,667 works · [🤖 Artificial Intelligence](https://md.research.elamur.ai/free/en/shelves/computer-science/artificial-intelligence/) · 16,194 works · [🎭 Arts and Humanities](https://md.research.elamur.ai/free/en/shelves/arts-and-humanities/) · 16,039 works · [🗣️ Language and Linguistics](https://md.research.elamur.ai/free/en/shelves/arts-and-humanities/language-and-linguistics/) · 15,891 works · [💭 Psychology](https://md.research.elamur.ai/free/en/shelves/psychology/) · 1,552 works · [💭 Experimental and Cognitive Psychology](https://md.research.elamur.ai/free/en/shelves/psychology/experimental-and-cognitive-psychology/) · 870 works 🧭 Topics (3): [😊 Sentiment & emotion](https://md.research.elamur.ai/free/en/topics/sentiment-emotion/) · 972 works · [🗨️ Dialogue systems](https://md.research.elamur.ai/free/en/topics/dialogue-systems/) · 624 works · [🔍 Information retrieval & search](https://md.research.elamur.ai/free/en/topics/information-retrieval/) · 473 works ## Cite Cite as: Xu Wang, Bo Wang, Yang Xiang, Yihong Tang, Dongming Zhao, Yuzifei and Yuexian Hou. 2026. REG: Retrieval via Emotion Similarity for Guiding Empathetic Dialogue Generation. Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). The paper itself, not this page. BibTeX from the library's record of the work. ```bibtex @inproceedings{wang2026reg, title = {{REG: Retrieval via Emotion Similarity for Guiding Empathetic Dialogue Generation}}, author = {Xu Wang, Bo Wang, Yang Xiang, Yihong Tang, Dongming Zhao, Yuzifei and Yuexian Hou}, booktitle = {Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)}, year = {2026}, doi = {10.18653/v1/2026.acl-long.860}, url = {https://doi.org/10.18653/v1/2026.acl-long.860} } ``` - [← previous ReFreeKV: Towards Threshold-Free KV Cache Compression](https://md.research.elamur.ai/free/en/papers/ni-2026-refreekv-threshold-free-kv-cache/) - [next → Region-Grounded Report Generation for 3D Medical Imaging: A Fine-Grained Dataset and Graph-Enhanced Framework](https://md.research.elamur.ai/free/en/papers/nguyen-2026-region-grounded-report-generation-3d/) - [↑ back to papers of 2026](https://md.research.elamur.ai/free/en/papers/2026/page/41/) ## References 2 references, 0 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet. 1. Mingxiu Cai, Daling Wang, Shi Feng, and Yifei Zhang. 2024. Pecer: Empathetic response generation via dynamic personality extraction and contextual emotional reasoning. In ICASSP 2024-2024 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), pages 10631–10635. IEEE. [find](https://scholar.google.com/scholar?q=Mingxiu+Cai%2C+Daling+Wang%2C+Shi+Feng%2C+and+Yifei+Zhang%2E+2024%2E+Pecer%3A+Empathetic+response+generation+via+dynamic+personality+extraction+and+contextual+emotional+reasoning%2E+In+ICASSP+2024%2D2024+IEEE+Interna) 2. Changyu Chen, Yanran Li, Chen Wei, Jianwei Cui, Bin Wang, and Rui Yan. 2024. Empathetic response generation with relation-aware commonsense knowledge. In WSDM, pages 87–95. *Table 6: Human evaluation results (Win/Lose %). Ties are omitted for brevity. Kappa (κ) scores indicate moderate inter-annotat… [find](https://scholar.google.com/scholar?q=Changyu+Chen%2C+Yanran+Li%2C+Chen+Wei%2C+Jianwei+Cui%2C+Bin+Wang%2C+and+Rui+Yan%2E+2024%2E+Empathetic+response+generation+with+relation%2Daware+commonsense+knowledge%2E+In+WSDM%2C+pages+87%E2%80%9395%2E+%2ATable+6%3A+Human+evaluation+) *Updated 1 November 2026*