Learning Guarantee
Jump to navigation
Jump to search
A Learning Guarantee is a theoretical guarantee for a learning algorithm.
- See: PAC Learning.
References
2008
- (Balcan et al., 2008) ⇒ Maria-Florina Balcan, Avrim Blum, and Nathan Srebro. (2008). “Improved Guarantees for Learning via Similarity Functions." Computer Science Department
- QUOTE: … Thus, all we need now to learn well is to draw a new fresh sample ˜S, map it into the transformed space using φS, and then apply a good algorithm for learning ... The Corollary now follows from the L1-margin learning guarantee in the separable case, discussed earlier in the...