Saved in:
Bibliographic Details
Main Authors: Maass, Wolfgang, Janzen, Sabine, Saxena, Prajvi, Mukherjee, Sach
Format: Preprint
Published: 2026
Subjects:
Online Access:https://arxiv.org/abs/2602.04807
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866908813517389824
author Maass, Wolfgang
Janzen, Sabine
Saxena, Prajvi
Mukherjee, Sach
author_facet Maass, Wolfgang
Janzen, Sabine
Saxena, Prajvi
Mukherjee, Sach
contents We introduce Afferent Learning, a framework that produces Computational Afferent Traces (CATs) as adaptive, internal risk signals for damage-avoidance learning. Inspired by biological systems, the framework uses a two-level architecture: evolutionary optimization (outer loop) discovers afferent sensing architectures that enable effective policy learning, while reinforcement learning (inner loop) trains damage-avoidance policies using these signals. This formalizes afferent sensing as providing an inductive bias for efficient learning: architectures are selected based on their ability to enable effective learning (rather than directly minimizing damage). We provide theoretical convergence guarantees under smoothness and bounded-noise assumptions. We illustrate the general approach in the challenging context of biomechanical digital twins operating over long time horizons (multiple decades of the life-course). Here, we find that CAT-based evolved architectures achieve significantly higher efficiency and better age-robustness than hand-designed baselines, enabling policies that exhibit age-dependent behavioral adaptation (23% reduction in high-risk actions). Ablation studies validate CAT signals, evolution, and predictive discrepancy as essential. We release code and data for reproducibility.
format Preprint
id arxiv_https___arxiv_org_abs_2602_04807
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle Evolving Afferent Architectures: Biologically-inspired Models for Damage-Avoidance Learning
Maass, Wolfgang
Janzen, Sabine
Saxena, Prajvi
Mukherjee, Sach
Machine Learning
We introduce Afferent Learning, a framework that produces Computational Afferent Traces (CATs) as adaptive, internal risk signals for damage-avoidance learning. Inspired by biological systems, the framework uses a two-level architecture: evolutionary optimization (outer loop) discovers afferent sensing architectures that enable effective policy learning, while reinforcement learning (inner loop) trains damage-avoidance policies using these signals. This formalizes afferent sensing as providing an inductive bias for efficient learning: architectures are selected based on their ability to enable effective learning (rather than directly minimizing damage). We provide theoretical convergence guarantees under smoothness and bounded-noise assumptions. We illustrate the general approach in the challenging context of biomechanical digital twins operating over long time horizons (multiple decades of the life-course). Here, we find that CAT-based evolved architectures achieve significantly higher efficiency and better age-robustness than hand-designed baselines, enabling policies that exhibit age-dependent behavioral adaptation (23% reduction in high-risk actions). Ablation studies validate CAT signals, evolution, and predictive discrepancy as essential. We release code and data for reproducibility.
title Evolving Afferent Architectures: Biologically-inspired Models for Damage-Avoidance Learning
topic Machine Learning
url https://arxiv.org/abs/2602.04807