The paper tackles a question about the internal structure of large graphs. In any graph, some vertices have more connections (neighbors) than others, and the vertex or vertices with the most connections are said to have the "maximum degree." A natural question is: can you always find a large portion of the graph where the maximum degree is achieved by many vertices simultaneously, rather than just one or two? Caro and Yuster conjectured that the answer is yes, and this paper proves it.
Specifically, the authors show that for any target number k of vertices you want to achieve the maximum degree, you can always find an induced subgraph (meaning a subset of vertices along with all the edges between them that existed in the original graph) that is nearly as large as the whole graph and in which at least k vertices all share the same maximum degree. The size of what you have to remove from the graph to achieve this is at most proportional to the square root of the maximum degree, which is a very small sacrifice when the graph is large.
The result is described as confirming the Caro-Yuster conjecture "in strong form" because it not only proves the conjecture but gives a quantitative bound on how few vertices need to be removed to get the desired structure. This kind of result belongs to a broader area of combinatorics concerned with finding well-structured or "regular" subgraphs hiding inside arbitrary, potentially messy, large graphs.