{"slug": "low-power-analogue-neural-networks-with-trainable-nonlinear-connections-for", "title": "Low-power analogue neural networks with trainable nonlinear connections for continuous control", "summary": "Researchers developed low-power analogue neural networks with trainable nonlinear connections, inspired by Kolmogorov-Arnold networks, achieving efficient continuous control tasks like robotic kinematics and photovoltaic tracking with fewer nodes than traditional MLPs. The approach, demonstrated on field-programmable analogue arrays and memristive simulations, projects a dedicated CMOS implementation operating at ~30 microwatts.", "body_md": "arXiv:2606.23742v1 Announce Type: new\nAbstract: Physical neural networks promise low-power machine learning by computing directly with analogue device physics, but most architectures force nonlinear device responses to act as scalar weights. Inspired by Kolmogorov-Arnold networks, we place trainable nonlinear functions on the connections, making each physical connection a learnable computational element. Realising these functions as analogue band-pass filters on field-programmable analogue arrays, we find that the benefit is task-dependent and follows from the smoothness of the physical basis: the networks represent smooth, continuously valued targets, including robotic kinematics, continuous control, and photovoltaic maximum-power-point tracking, with far fewer nodes and connections than multilayer perceptrons, but offer no parameter-efficiency advantage on classification-like decision boundaries. Trained networks transfer to hardware across approximately 35,000 connections with quantified fidelity, and a dedicated CMOS implementation is projected to operate at approximately 30 microwatts. A memristive realisation reproduces the same behaviour in simulation, indicating that the advantage comes from placing trainable nonlinearity on connections, rather than from a particular device.", "url": "https://wpnews.pro/news/low-power-analogue-neural-networks-with-trainable-nonlinear-connections-for", "canonical_source": "https://arxiv.org/abs/2606.23742", "published_at": "2026-06-24 04:00:00+00:00", "updated_at": "2026-06-24 04:28:37.232252+00:00", "lang": "en", "topics": ["neural-networks", "machine-learning"], "entities": ["Kolmogorov-Arnold networks", "CMOS", "field-programmable analogue arrays"], "alternates": {"html": "https://wpnews.pro/news/low-power-analogue-neural-networks-with-trainable-nonlinear-connections-for", "markdown": "https://wpnews.pro/news/low-power-analogue-neural-networks-with-trainable-nonlinear-connections-for.md", "text": "https://wpnews.pro/news/low-power-analogue-neural-networks-with-trainable-nonlinear-connections-for.txt", "jsonld": "https://wpnews.pro/news/low-power-analogue-neural-networks-with-trainable-nonlinear-connections-for.jsonld"}}