{"ID":2837243,"CreatedAt":"2026-06-01T04:54:23.091178241Z","UpdatedAt":"2026-06-01T04:54:23.091178241Z","DeletedAt":null,"paper_url":"https://arxiv.org/abs/2511.18667","arxiv_id":"2511.18667","title":"Equivariant Deep Equilibrium Models for Imaging Inverse Problems","abstract":"Equivariant imaging (EI) enables training signal reconstruction models without requiring ground truth data by leveraging signal symmetries. Deep equilibrium models (DEQs) are a powerful class of neural networks where the output is a fixed point of a learned operator. However, training DEQs with complex EI losses requires implicit differentiation through fixed-point computations, whose implementation can be challenging. We show that backpropagation can be implemented modularly, simplifying training. Experiments demonstrate that DEQs trained with implicit differentiation outperform those trained with Jacobian-free backpropagation and other baseline methods. Additionally, we find evidence that EI-trained DEQs approximate the proximal map of an invariant prior.","short_abstract":"Equivariant imaging (EI) enables training signal reconstruction models without requiring ground truth data by leveraging signal symmetries. Deep equilibrium models (DEQs) are a powerful class of neural networks where the output is a fixed point of a learned operator. However, training DEQs with complex EI losses requir...","url_abs":"https://arxiv.org/abs/2511.18667","url_pdf":"https://arxiv.org/pdf/2511.18667v1","authors":"[\"Alexander Mehta\",\"Ruangrawee Kitichotkul\",\"Vivek K Goyal\",\"Julián Tachella\"]","published":"2025-11-24T00:43:54Z","proceeding":"eess.IV","tasks":"[\"eess.IV\",\"cs.LG\",\"eess.SP\"]","methods":"[]","has_code":false}