Семинар за вештачку интелигенцију, 11. октобар 2023.

• 10. Октобар, 2023
• Коментари (0)

Наредни састанак Семинара биће одржан онлајн у среду, 11. октобра 2023, са почетком у 19 часова.

Предавач: Јohanna Björklund, Associate Professor, Department of Computing Science, Umeå University, Umeå, Sweden

Наслов предавања: THE IMPACT OF STATE MERGING ON PREDICTIVE ACCURACY IN PROBABILISTIC TREE AUTOMATA: DIETZE`S CONJECTURE REVISITED
 
Апстракт: Dietze`s conjecture concerns the problem of equipping a tree automaton M with weights to make it probabilistic, in such a way that the resulting automaton N predicts a given corpus C as accurately as possible. The conjecture states that the accuracy cannot increase if the states in M are first merged with respect to an equivalence relation on the state space, so that the result is a smaller automaton M`. Put differently, merging states can never improve predictions. This is under the assumption that both M and M` are bottom-up deterministic and accept every tree in corpus. We prove that the conjecture holds, using a construction that turns any probabilistic version N` of M` into a probabilistic version N of M, such that N assigns at least as great a weight to each tree in C as N` does.
 
Напомена: Регистрациона форма за учешће на Семинару је доступна на линку:
https://miteam.mi.sanu.ac.rs/asset/CW5nJWDSEZDj7p32p

Уколико желите само да пратите предавање без могућности активног учешћа, пренос је доступан на линку:
https://miteam.mi.sanu.ac.rs/asset/4LNW8WtML7rLKojoz