Equality of convex functions from equal minimal subgradient norms

About 9 years old · traced to

Let H\mathcal{H} be the underlying Hilbert space, and let ψ1,ψ2:H→R\psi_1,\psi_2:\mathcal{H}\rightarrow\mathbb{R} be finite convex functions bounded from below. For each x∈Hx\in\mathcal{H}, let ∂ψi(x)\partial\psi_i(x) denote the subdifferential of ψi\psi_i at xx.

Equal minimal subgradient norm conjecture. If

inf⁡p∈∂ψ1(x)∥p∥=inf⁡q∈∂ψ2(x)∥q∥for all x∈H,\inf_{p\in\partial\psi_1(x)}\|p\|=\inf_{q\in\partial\psi_2(x)}\|q\|\quad\text{for all }x\in\mathcal{H},

then ψ1=ψ2+c\psi_1=\psi_2+c for some constant c>0c>0.

This conjecture would extend the preceding uniqueness theorem for twice continuously differentiable convex functions to finite, possibly nonsmooth convex functions. The source presents it as plausible but does not provide a resolution.

References

Primary source

Tahar Boulmezaoud, Philippe Cieutat and Aris Daniilidis, “Gradient flows, second order gradient systems and convexity”, arXiv:1710.07858 (2018).

Progress summary

Never refreshed

Nothing recorded yet. Refresh searches the literature and the public web for attempts on this problem, and writes the first summary here.

Solutions 0

No solutions have been posted yet.