# CALMRec: Causally Aligned Language Memory for Long-Horizon Recommendation

arXiv: https://arxiv.org/abs/2607.23647 · HTML: https://mlsanity.com/paper/2607.23647/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 13 of its 22 references that are in the map.
- First posted on arXiv 26 Jul 2026.

## Builds on (most cited first, 13 of 13)

- [Conservative Q-Learning for Offline Reinforcement Learning](https://mlsanity.com/paper/2006.04779/) (2020)
- [Self-Attentive Sequential Recommendation](https://mlsanity.com/paper/1808.09781/) (2018)
- [BERT4Rec: Sequential Recommendation with Bidirectional Encoder Representations from Transformer](https://mlsanity.com/paper/1904.06690/) (2019)
- [Session-based Recommendations with Recurrent Neural Networks](https://mlsanity.com/paper/1511.06939/) (2015)
- [A Survey on Large Language Models for Recommendation](https://mlsanity.com/paper/2305.19860/) (2023)
- [Recommendations as Treatments: Debiasing Learning and Evaluation](https://mlsanity.com/paper/1602.05352/) (2016)
- [KuaiRec: A Fully-observed Dataset and Insights for Evaluating Recommender Systems](https://mlsanity.com/paper/2202.10842/) (2022)
- [XRec: Large Language Models for Explainable Recommendation](https://mlsanity.com/paper/2406.02377/) (2024)
- [Multi-Task Fusion via Reinforcement Learning for Long-Term User Satisfaction in Recommender Systems](https://mlsanity.com/paper/2208.04560/) (2022)
- [LLM2Rec: Large Language Models Are Powerful Embedding Models for Sequential Recommendation](https://mlsanity.com/paper/2506.21579/) (2025)
- [User Retention-oriented Recommendation with Decision Transformer](https://mlsanity.com/paper/2303.06347/) (2023)
- [UniM: A Unified Any-to-Any Interleaved Multimodal Benchmark](https://mlsanity.com/paper/2603.05075/) (2026)
- [LLM-Enhanced Reinforcement Learning for Long-Term User Satisfaction in Interactive Recommendation](https://mlsanity.com/paper/2601.19585/) (2026)

## Abstract (arXiv)

Large language models (LLMs) can summarize heterogeneous user evidence in natural language, but current LLM recommenders often collapse enduring preferences, transient intent, and exposure-induced behavior into one profile. This makes recommendation vulnerable to feedback loops: repeated exposure is mistaken for preference, immediate clicks dominate delayed satisfaction, and fluent explanations need not reflect the ranking decision. We propose our method, a model-agnostic framework for long-horizon recommendation. Our method uses a frozen multimodal language model to convert item content and feedback into evidence-grounded semantic atoms, then maintains separate short-term, long-term, and exposure memories. Propensity-weighted updates reduce policy-induced exposure bias, while a conservative offline critic reranks candidates for delayed satisfaction under a behavior-support constraint. Explanations use only influential evidence atoms and are checked by counterfactual deletion. We provide an identification result and evaluate the framework in e-commerce-like, news-like, and short-video-like environments. Across ten seeds, our method improves discounted long-term value over the strongest alternative by 6.1%, 7.6%, and 6.7%, respectively. Twenty-seed paired ablations show significant value drops after removing propensity correction (0.739 +/- 0.191) or conservative support regularization (0.523 +/- 0.234). A frozen instruction language model also more than doubles semantic-atom NDCG over TF-IDF on a held-out paraphrase benchmark.