ProbXiv
Explorersign in
Problem archiveProblem record

Statement

Let QQ and KK be arbitrary matrices, and let VV be a matrix whose largest eigenvalue λmax⁡>0\lambda_{\max}>0 is real, with eigenspace LL satisfying dimL≥2\mathop{\rm dim} L\geq2. Assume that Vz∈L⊥Vz\in L^{\perp} and ⟨Vz,z⟩<λmax⁡∣z∣2\langle Vz,z\rangle<\lambda_{\max}|z|^2 for every z∈L⊥z\in L^{\perp}. For an initialization (x1(0),…,xn(0))(x_1(0),\ldots,x_n(0)), let PL⊥\mathbf{P}_{L^{\perp}} denote orthogonal projection onto L⊥L^{\perp}, set y1:=PL⊥(x1(0))y_1:=\mathbf{P}_{L^{\perp}}(x_1(0)), and define ξ:=y1/∣y1∣\xi:=y_1/|y_1|. One-cluster convergence conjecture. For almost every initialization, the causal attention dynamics converge to one cluster, specifically

∀k∈[n],lim⁡t→∞xk(t)=ξ.\forall k\in[n],\qquad \lim_{t\to\infty}x_k(t)=\xi.

This is the proposed description of the positive-dominant-eigenvalue case when the dominant eigenspace has dimension at least two. The paper presents it as a conjectural generalization because critical manifolds make the corresponding stable-manifold analysis technically difficult; no resolution is given.

Record

Comments

Nothing has been published against this problem yet, and nothing has been said about it.

Sign in with an institutional address to take part in the discussion. Reading every thread stays open to everyone.

Sign in

Solve with an agent

Open the statement in a chat, with the problem and the ground rules already written into the prompt.

This opens a third-party site. Nothing is posted back to ProbXiv and nothing you write there is recorded here — what a model gives you is an attempt, which a person still has to check.