`Travelled to:`

1 × China

1 × Germany

3 × Canada

5 × USA

`Collaborated with:`

J.Langford C.Elkan N.Abe A.Beygelzimer C.N.d.Santos ∅ E.P.D.Pednault C.Perlich S.Rosset R.D.Lawrence V.Dani T.P.Hayes

`Talks about:`

learn (8) classifi (3) probabl (3) estim (3) decis (3) cost (3) reinforc (2) perform (2) classif (2) sensit (2)

## Person: Bianca Zadrozny

### DBLP: Zadrozny:Bianca

### Contributed to:

### Wrote 12 papers:

- ICML-c2-2014-SantosZ #learning
- Learning Character-level Representations for Part-of-Speech Tagging (CNdS, BZ), pp. 1818–1826.
- ICML-2009-BeygelzimerLZ #machine learning #reduction #summary #tutorial
- Tutorial summary: Reductions in machine learning (AB, JL, BZ), p. 12.
- KDD-2007-PerlichRLZ #estimation #modelling
- High-quantile modeling for customer wallet estimation and other applications (CP, SR, RDL, BZ), pp. 977–985.
- KDD-2006-AbeZL #detection #learning
- Outlier detection by active learning (NA, BZ, JL), pp. 504–509.
- ICML-2005-BeygelzimerDHLZ #classification #fault #reduction
- Error limiting reductions between classification tasks (AB, VD, TPH, JL, BZ), pp. 49–56.
- ICML-2005-LangfordZ #classification #learning #performance
- Relating reinforcement learning performance to classification performance (JL, BZ), pp. 473–480.
- ICML-2004-Zadrozny #bias #classification #learning
- Learning and evaluating classifiers under sample selection bias (BZ).
- KDD-2004-AbeZL #learning #multi
- An iterative method for multi-class cost-sensitive learning (NA, BZ, JL), pp. 3–11.
- KDD-2002-PednaultAZ #learning
- Sequential cost-sensitive decision making with reinforcement learning (EPDP, NA, BZ), pp. 259–268.
- KDD-2002-ZadroznyE #classification #multi #probability
- Transforming classifier scores into accurate multiclass probability estimates (BZ, CE), pp. 694–699.
- ICML-2001-ZadroznyE #classification #naive bayes #probability
- Obtaining calibrated probability estimates from decision trees and naive Bayesian classifiers (BZ, CE), pp. 609–616.
- KDD-2001-ZadroznyE #learning
- Learning and making decisions when costs and probabilities are both unknown (BZ, CE), pp. 204–213.