{"id":"arxiv/2606.21153","url":"https://arxiv.org/abs/2606.21153","tags":["math.OC","cs.LG"],"score":0,"title":"DUET: Decentralized Bilevel Optimization without Lower-Level Strong Convexity","domain":"math.OC","source":"arxiv","authors":["Zhen Qin","Zhuqing Liu","Songtao Lu","Yingbin Liang","Jia Liu"],"reviews":0,"subject":"math.OC","summary":"Decentralized bilevel optimization (DBO) provides a powerful framework for multi-agent systems to solve local bilevel tasks in a decentralized fashion without the need for a central server. However, most existing DBO methods rely on lower-level strong convexity (LLSC) to guarantee unique solutions and a well-defined hypergradient for stationarity measure, hindering their applicability in many practical scenarios not satisfying LLSC. To overcome this limitation, we introduce a new single-loop DBO algorithm called diminishing quadratically-regularized bilevel decentralized optimization (DUET),…","category":"research","comments":0,"disputed":false,"division":"math","mentions":0,"discussed":0,"published":"2026-06-19","reactions":{}}