{"id":"arxiv/2605.19705","url":"https://arxiv.org/abs/2605.19705","tags":["math.OC"],"score":0,"title":"i-DEQ: A stable inertial deep equilibrium model for image restoration","domain":"math.OC","source":"arxiv","authors":["Antonin Clerc","Marien Renaud","Baudouin Denis De Senneville","Nicolas Papadakis"],"reviews":0,"subject":"math.OC","summary":"Deep Equilibrium Models (DEQs) are an established framework for image restoration that learn a problem-adapted regularization by solving a fixed-point (i.e. equilibrium) problem. While flexible and expressive, DEQs are often hindered by high computational cost and training instability. We propose an inertial DEQ (i-DEQ) that learns an explicit nonconvex regularization within the DEQ formulation. By using momentum within the fixed-point iterations, i-DEQ has convergence guarantees and accelerated rates. Moreover, we observe that i-DEQ is significantly more stable during the training and…","category":"research","comments":0,"disputed":false,"division":"math","mentions":0,"discussed":0,"published":"2026-05-19","reactions":{}}