Invited Talk by Dr. Naresh Manwani on Learning in presence of noise
Title: Learning in presence of noise
Speaker: Dr. Naresh Manwani.
Host Faculty: Dr.Subrahmanyam Kalyanasundaram
Time: 14:30 -15:30
In many applications, the training data is corrupted with label noise. Many standard algorithms such as SVM, logistic regression etc. perform poorly in the presence of label noise. We investigate the robustness of risk minimization to label noise. We prove a sufficient condition on a loss function for the risk minimization under that loss to be tolerant to uniform label noise. We show that the 0–1 loss, sigmoid loss, ramp loss and probit loss satisfy this condition. Though none of the standard convex loss functions satisfy it. We also prove that, by choosing a sufficiently large value of a parameter in the loss function, the sigmoid loss, ramp loss and probit loss can be made tolerant to nonuniform label noise also if we can assume the classes to be separable under noise-free data distribution. Through extensive empirical studies, we show that risk minimization under the 0–1 loss, the sigmoid loss and the ramp loss has much better robustness to label noise when compared to the SVM algorithm.
Naresh Manwani did B.E. in Electronics and Communication from Rajasthan University Jaipur in 2003, M. Tech. in Information and Communication Technology from Dhirubhai Ambani Institute of Information and Communication Technology (DAIICT) Gandhinagar in 2006. He finished his Ph.D. from Indian Institute of Science, Bangalore in 2013. His main research interests are machine learning and data mining. Currently he is working as an Applied Scientist at Microsoft IDC Bangalore. before that he worked as a Research Scientist in the Data Mining Lab at GE Global Research Bangalore.
- Naresh Manwani, P. S. Sastry, Noise Tolerance under Risk Minimization, IEEE Transactions on Cybernetics, volume: 43, issue: 3, pages: 1146-1151, March, 2013.
- Aritra Ghosh, Naresh Manwani, P. S. Sastry, Making Risk Minimization Tolerant to Label Noise, Neurocomputing Elsevier, volume: 160, pages: 93-107, July 2015.
Wednesday, January 20, 2016 - 14:30 to 15:30