The abstract is read at the publisher; the summary is JClub's.
GENESIS-Value constructs value on the fly from episodic memories, using a flexible retrieval mechanism to support multiple value-based decision scenarios across five simulations.
- Why it matters: It addresses how a shared episodic mechanism can flexibly support diverse valuation and choice demands, beyond relying on stable, precomputed value representations.
- What they did: The framework combines generative stimulus representations with query-key-value retrieval, then tests episodic sampling for reconstructing experienced values, generalizing to novel stimuli, estimating feature values, and guiding individual and sequential choices.
- The result: Retrieval can be steered by stimulus features, temporal proximity, reminders, reward outcomes, or their combinations, enabling adaptation across different choice demands without separate value functions.
The findingWhy it mattersWhat they didThe result
AI summary of the abstract · may contain errors · not medical advice · read the paper.