The paper introduces Explanatory World Models (EWM), an attribution framework that leverages explanations from a frontier language model to identify earlier events that plausibly contributed to a delayed outcome. EWM redistributes reward recursively while preserving the undiscounted episode return, and is evaluated on long‑horizon RL benchmarks, a public sequential‑recommendation task, and an industry‑scale generative recommender, consistently outperforming baseline methods, especially when rewards are highly delayed. The study demonstrates that structured explanations can provide reusable credit signals for learning from delayed outcomes.
arXiv:2610.07349v1 Announce Type: cross Abstract: Group Relative Policy Optimization (GRPO) avoids a separate critic by estimating advantages from rollout groups. For multi-turn agents, however, traj...
arXiv:2607. 21090v1 Announce Type: cross Abstract: We propose a Reinforcement Learning (RL) method to directly optimize the faithfulness of self-explanations - the extent to which a model's generated reasoning accurately reflects its internal decision-making process.
The paper introduces MARI, a Memory-Augmented Recommendation system that uses a Decision Memory Bank to store Structured Decision Memories (SDMs) capturing users’ goals, constraints, and trade-offs. Unlike prior methods that flatten heterogeneous behaviors into token sequences, MARI retrieves relevant SDMs to enhance LLM reasoning, improving performance on next-item prediction and a new Difficult Choice Prediction task while keeping latency low. The approach also provides interpretable, human-readable insights into user decision-making.
The paper presents a method for fine‑tuning a large language model (LLM) recommender to generate personalized, non‑harmful explanations for its recommendations. By training two LLM‑judge reward models and using constrained GRPO, the authors achieve a significant increase in the PASS rate for all three criteria, from 0.649 to 0.956 on their own judges and from 0.677 to 0.931 on an independent judge. The fine‑tuned model maintains its original recommendation performance, demonstrating that LLM‑based recommenders can be adapted to complex tasks without loss of effectiveness.
CausalDreamer enhances generative world models by freezing the original video tokenizer and re-encoding its latent space into a factored representation that separates controllable versus uncontrollable dynamics and reward-relevant versus reward-irrelevant information. The pretrained dynamics model is fine‑tuned to predict this factored representation, enabling more accurate model‑predictive planning. On the MMBench2 benchmark, CausalDreamer improves normalized scores by 14% on clean tasks and 25% on manipulated variants compared to the baseline, while both models perform poorly on entirely new environments.
arXiv:2608. 11980v2 Announce Type: replace-cross Abstract: Semantic-ID generative recommenders represent each item as a short sequence of discrete semantic tokens and predict the next item by autoregressively generating this token sequence.
arXiv:2607. 23647v1 Announce Type: cross Abstract: 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.
DeGRe is a dense‑supervised generative reranking framework designed to improve multi‑stage recommender systems by addressing label bias and credit assignment issues. It uses an offline Lookahead Evaluator with beam search to generate dense supervision signals, which are distilled into a lightweight Online Generator that can perform efficient greedy decoding at inference time. Experiments show that DeGRe outperforms baselines on public benchmarks and industrial datasets, and it has been successfully deployed on Taobao Flash Shopping to enhance online recommendations.
The paper introduces VICT, a method that leverages the internal structure of verifiable tasks to perform fine‑grained credit assignment for long‑horizon LLM agents. VICT exposes executable or evidence‑backed atoms from a task’s terminal verifier and traces them back to actions via dependency‑valid proof edges, redistributing advantage only along these edges. This approach improves performance on ALFWorld and WebShop compared to outcome‑only training and matches recent fine‑grained credit methods without requiring additional critics, labels, or inference‑time verifier access.
The paper proposes a retrieval‑grounded credit‑assignment method for generative recommenders that use Semantic IDs (SIDs). By structuring each generated trace into a history summary, a set of interest hypotheses, and a final SID, and then verifying each hypothesis with a frozen retriever, the method assigns reward at the hypothesis level rather than only at the final SID. Experiments on Amazon Reviews datasets show consistent improvements in SID recommendation, and an oracle analysis on Video Games data demonstrates that selecting target‑relevant queries among generated interests boosts recall and ranking.
arXiv:2609.38334v1 Announce Type: new Abstract: Large language models (LLMs) are increasingly deployed as agents for multi-step decision-making, yet transfer poorly to unseen environments. World-mode...
The paper introduces a retrieval‑grounded credit‑assignment method for generative recommenders that use Semantic IDs (SIDs). By structuring each autoregressive trace into a history summary, a set of interest hypotheses, and a final SID, a frozen retriever verifies each hypothesis as a catalog query. Rewards are assigned at the hypothesis level when any query retrieves the target within the top‑K, allowing distinct updates for rollouts that share the same SID reward and improving SID recommendation performance on Amazon Reviews datasets.