How statement and proof provenance work
The first chip identifies the source of the statement or construction; the second identifies the source of its local proof or verification.
- Literature-sourced: the exact statement appears in a cited source; only wording and notation differ.
- AI-adapted: a semantically identical restatement of literature-sourced material, modulo indexing, notation, and boundary cases adopted by the library.
- AI-generated: a genuinely novel statement formulated by AI, with no source for the claim itself.
These labels describe origin, not correctness: citations and verification chips remain separate evidence.
Bipartite Ramanujan existence and the centered spectrum
Statement
For every integer there is a sequence of finite simple -regular bipartite graphs whose vertex counts tend to infinity and whose adjacency eigenvalues other than the source's trivial values and lie in . Here bipartite means that the vertices have two parts and every edge joins different parts; -regular means that each vertex has exactly neighbors. The adjacency matrix has entry one at adjacent pairs and zero elsewhere.
For each such graph, with vertices and all-ones matrix , the centered adjacency norm satisfies . Thus the existence theorem with both trivial values excluded does not itself provide the centered absolute-spectrum bound .
External input
The existence and nontrivial-eigenvalue bound in the first paragraph are imported from Marcus, Spielman and Srivastava, Section 2.1 and Theorem 5.5. Their proof starts with a complete bipartite graph and obtains graphs of twice the preceding order; the cited interlacing and lifting machinery is not reconstructed here.
Facts & Assumptions
Given: An integer and a graph supplied by the external input, with bipartition .
Goal: Establish the stated comparison with the centered adjacency norm.
The imported theorem supplies the graphs and their nontrivial spectral bound, with both and designated trivial in the bipartite case.
For a finite simple graph, adjacency is represented by its symmetric zero-diagonal – matrix; its row sums are its vertex degrees (Finite symmetric networks and loop removal).
For a symmetric nonnegative matrix with constant row sum , the centered operator agrees with on the vectors of coordinate sum zero, and its norm is the norm of that restriction (Regular adjacency spectra and normalized expansion).
Proof
Counting edges from the two sides gives , so . Define to be on and on . It is nonzero and has coordinate sum zero. Each neighbor of a vertex lies in the opposite part, and every vertex has neighbors; therefore .
By [F3], . The definition of operator norm applied to this nonzero vector gives . Since , , and hence , which gives . Together with the imported existence statement [F1], this proves the claim and the stated limitation of the bipartite convention.
Depends on
Dependency tree · 2 levels
Each result is drawn at its shortest distance from this item. Arrows run from prerequisites to the items that use them; this item has a heavier outline. External source nodes link directly to the cited text. The view shows up to two dependency steps; open a prerequisite to continue.
External dependencies
For every integer d at least 3 there is an infinite sequence of d-regular bipartite Ramanujan graphs. In the source convention the eigenvalues d and, for a bipartite graph, -d are trivial; every nontrivial eigenvalue lies between -2 sqrt(d-1) and 2 sqrt(d-1). The proof produces successive graphs with twice as many vertices.
Hypotheses: An integer d at least 3; finite d-regular bipartite graphs, with Ramanujan defined using the source's trivial-eigenvalue convention.
Reading limits: Read Section 2.1 and the statement and proof paragraph of Theorem 5.5. The interlacing-polynomial, signing and 2-lift machinery used by the source was not reconstructed or independently verified. The centered-spectrum comparison below is a local deduction, not a claim attributed to the source.
Sources
- Interlacing families I: Bipartite Ramanujan graphs of all degrees · Section 2.1, printed p. 308; Theorem 5.5 and its proof, printed p. 316