The cycle double cover conjecture is a famous unsolved problem in graph theory. It asks whether every bridgeless graph (a network with no single edge whose removal disconnects the graph) can have its edges covered by a collection of cycles such that every edge appears in exactly two of those cycles. This has been an open problem for decades, and the paper claims to prove it.
The authors approach the problem through the lens of "nowhere-zero flows," which are ways of assigning numerical values to the edges of a graph such that flow is conserved at every vertex and no edge gets a value of zero. Specifically, they work with three-bit flows, meaning the flow values come from a small algebraic structure built out of combinations of three binary digits. A key insight is that the flow values at each vertex can be visualized as the differences along the sides of a triangle, where each side carries a label from this binary structure. The challenge then becomes showing that these local triangles can be assembled consistently across the entire graph, meaning neighboring vertices agree on the labels they share.
The authors frame this consistency requirement as a system of binary equations and prove the system always has a solution using three complementary arguments: a certificate that detects inconsistencies, a local identity connecting testing and parity, and a global counting argument. Once consistent labels exist across the whole graph, the labeled edges naturally organize themselves into cycles where every edge appears exactly twice, which is precisely what the cycle double cover conjecture requires. If the proof holds up to scrutiny, it would resolve one of the most prominent open questions in combinatorics and graph theory.