Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Deng, Zhuoran, Zhang, Yizhi, Zhang, Ziyi, Shen, Wan
Format: Preprint
Veröffentlicht: 2026
Schlagworte:
Online-Zugang:https://arxiv.org/abs/2603.09476
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
_version_ 1866911501858635776
author Deng, Zhuoran
Zhang, Yizhi
Zhang, Ziyi
Shen, Wan
author_facet Deng, Zhuoran
Zhang, Yizhi
Zhang, Ziyi
Shen, Wan
contents Goal-conditioned systems assume goals are provided externally. We ask whether attentional priorities can emerge endogenously from an agent's internal cognitive state. We propose a priority function that generates observation targets from three epistemic gaps: ignorance (posterior variance), surprise (prediction error), and staleness (temporal decay of confidence in unobserved variables). We validate this in two systems: a minimal attention-allocation environment (2,000 runs) and a modular, partially observable world (500 runs). Ablation shows each component is necessary. A key finding is metric-dependent reversal: under global prediction error, coverage-based rotation wins; under change detection latency, priority-guided allocation wins, with advantage growing monotonically with dimensionality (d = -0.95 at N=48, p < 10^-6). Detection latency follows a power law in attention budget, with a steeper exponent for priority-guided allocation (0.55 vs. 0.40). When the decay rate is made learnable per variable, the system spontaneously recovers environmental volatility structure without supervision (t = 22.5, p < 10^-6). We demonstrate that epistemic gaps alone, without external reward, suffice to generate adaptive priorities that outperform fixed strategies and recover latent environmental structure.
format Preprint
id arxiv_https___arxiv_org_abs_2603_09476
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle Telogenesis: Goal Is All U Need
Deng, Zhuoran
Zhang, Yizhi
Zhang, Ziyi
Shen, Wan
Artificial Intelligence
Goal-conditioned systems assume goals are provided externally. We ask whether attentional priorities can emerge endogenously from an agent's internal cognitive state. We propose a priority function that generates observation targets from three epistemic gaps: ignorance (posterior variance), surprise (prediction error), and staleness (temporal decay of confidence in unobserved variables). We validate this in two systems: a minimal attention-allocation environment (2,000 runs) and a modular, partially observable world (500 runs). Ablation shows each component is necessary. A key finding is metric-dependent reversal: under global prediction error, coverage-based rotation wins; under change detection latency, priority-guided allocation wins, with advantage growing monotonically with dimensionality (d = -0.95 at N=48, p < 10^-6). Detection latency follows a power law in attention budget, with a steeper exponent for priority-guided allocation (0.55 vs. 0.40). When the decay rate is made learnable per variable, the system spontaneously recovers environmental volatility structure without supervision (t = 22.5, p < 10^-6). We demonstrate that epistemic gaps alone, without external reward, suffice to generate adaptive priorities that outperform fixed strategies and recover latent environmental structure.
title Telogenesis: Goal Is All U Need
topic Artificial Intelligence
url https://arxiv.org/abs/2603.09476