Abstract. In this talk I will discuss three distinct but interconnected areas of research that I have investigated in recent years, all linked by the theme of finite geometry.
In the first part, we delve into Erd˝os-Ko-Rado (EKR) problems, a classical topic in combinatorics that explores intersection properties within mathematical structures. The foundational question concerns the maximum size of a family of k-subsets of an n-set, where every pair of subsets have at least one element in common. Erd˝os, Ko, and Rado proved that for n > 2k, the optimal construction involves subsets containing a fixed element. This result, known as the EKR theorem, extends to various other settings, including multisets, permutations, and finite geometries. In this talk, I will discuss both classical and recent results in the context of finite geometries.
The second part will be about generalized Johnson and Grassmann graphs, within the field of spectral graph theory. A central question is whether a graph is uniquely determined by its spectrum (the eigenvalues of its adjacency matrix). While many graphs are known to have this property, numerous well-known graphs possess cospectral mates (non-isomorphic graphs with identical spectra). This phenomenon appears prominently in highly symmetrical graphs, including Johnson and Grassmann graphs. Using techniques such as switching, Aida Abiad, Willem H. Haemers, Robin Simoens and I found a construction of cospectral mates for several specific generalized Johnson and Grassmann graphs.
In the final part, we make the link with coding theory, specifically trifferent codes, also known as perfect q-hash codes for q = 3, which has gained much attention since the 1980s because of their connections to various topics in cryptography, information theory, and computer science. Trifferent codes are ternary codes of length n with the property that for any three distinct codewords there is a coordinate where they all have distinct values. Over the finite field F
3, Anurag Bishnoi, Dion Gijswijt, Aditya Potukuchi and I could prove that minimal codes are equivalent to linear trifferent codes, which in turn are equivalent to strong blocking sets in the corresponding projective space. Using this equivalence, we could improve the known upper and lower bounds on the size of linear trifferent codes of length n.
