Learning What to Recall: Adaptive Multi-Cue Episodic Memory for World Models

arXiv 2026 31.8 method

TLDR

Proposes Future-Aware Recall (FAR), a learned episodic memory retrieval approach for world models using adaptive multi-cue scoring and future-aware predictive supervision.

Reasoning

The paper addresses a meaningful gap in world models by learning which memories and cues to trust for prediction, supported by empirical evaluations across three settings. Weaknesses are that the abstract does not specify real-world deployment details or compare against a wide range of baselines.

Read-first score

Read-first score 31.8, weighted from topical fit, citation, graph, method, reproducibility, and recency signals. Original total remains 39.

Recency 5%

100

Uses a gentle age decay so recent papers surface without erasing older foundations. 2026

Topical relevance 25%

55.7

Uses existing LLM keyword relevance scores normalized to 0-100. world model,world simulator,generative world model,interactive world model,video world model,world dynamics prediction,model-based reinforcement learning world model

Methodology quality 15%

Screens visible abstract and analysis fields for experiment, dataset, baseline, metric, and limitation evidence. markers=result

Reproducibility 15%

Screens links and visible text for paper, code, dataset, artifact, and repository signals. pdf=True; code=False; dataset=False; markers=none

Graph prestige 15%

5.7

PageRank-style prestige over the local citation graph built from available OpenAlex references. local_pagerank=0.0016

Citation impact 15%

Uses OpenAlex-shaped citation metadata as a bibliometric attention signal, separate from paper quality. cited_by_count=0

Citation velocity 10%

Citation velocity estimates citations per publication-year to reduce old-paper bias. velocity=0.00

Field roles

Frontier

Rank sensitivity

Stability: volatile; rank range: 268.

Keyword Scores

world model

world dynamics prediction

generative world model

interactive world model

video world model

model-based reinforcement learning world model

world simulator

Deep Analysis

Innovations

Future-Aware Recall (FAR), a framework that learns episodic recall from future-aware predictive supervision
Adaptive multi-cue scoring that learns cue-specific relevance and automatically determines which available retrieval cues to trust per query
Retriever trained using predictive utility measured by conditional log-likelihood of the realized future given recalled context, approximated by negative diffusion prediction loss
Retrieval remains future-blind at inference despite future-aware training

Methodology

FAR trains an episodic memory retriever by measuring the predictive utility of recalled context through the conditional log-likelihood of the realized future, approximated by negative diffusion prediction loss. The retriever learns adaptive scoring over available retrieval cues such as time, pose, vision, and audio, but uses only current information at inference. The approach is evaluated across three complementary settings.

Key Results

FAR outperforms hand-designed recall even when using the same retrieval cues, automatically adapts which available cues to trust, and recalls the right history as the world changes.

Resources

Paper