Telogenesis: Goal Is All U Need
This addresses the problem of autonomous attention allocation in AI agents, offering a novel approach that is incremental in advancing goal-conditioned systems.
The paper tackled the problem of whether attentional priorities can emerge internally from an agent's cognitive state without external goals, proposing a priority function based on epistemic gaps like ignorance, surprise, and staleness, and found that this approach outperforms fixed strategies and recovers latent environmental structure, with advantages growing with dimensionality (d = -0.95 at N=48, p < 10^-6) and detection latency following a power law (exponent 0.55 vs. 0.40).
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.