Tikhonov, Ivanov and Morozov regularization for support vector machine learning

69 citations recorded by the source corpus.

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

Efficient Data Representation by Selecting Prototypes with Importance Weights (2017, methodology): “The P-Lasso objective in our setting is the equivalent Ivanov formulation of standard Lasso [22] where we maximize l(w) subject to ||w||1 ≤ ,w ≥ 0 where for every choice of λ in Lasso, there is an equivalent in the Ivanov formulation [23] 1.”
Probably Approximately Metric-Fair Learning (2018, unknown): “lem: 12The two formulations are equivalent in the sense that any solution to the ﬁrst program (for some desired B) can be obtained by solving Program 2 (for an appropriate value λB), see Theorem 1 in [ORA16]. 30 minimize ξ,β Xm i=1 Xm ℓ=1 βℓK(xℓ,xi) −yi +λB Xm i,j=1 βiβjK(xi,xj) subject to Xm ℓ=1 βℓ”
Fairness in Machine Learning (2020, background): “However, in a large number of real world applications [9, 33, 142, 170] using the same model or part of it over diﬀerent tasks might be desirable.”