Here are some talks I have given over the years.
- One-class logistic regression & friends: probabilistic anomaly detection as loss minimisation. Data61 ML seminar, 2018.
[pdf]
- Three faces of binary classification. ANU ISML guest lecture, 2018.
[pdf]
- The hidden talents of logistic regression: from noisy labels to point processes. RIKEN AIP, 2017.
[pdf]
- Learning from noisy binary labels: a tale of two approaches. Amazon, 2016.
[pdf]
- Learning from corrupted binary labels via class-probability estimation. NICTA ML seminar, 2015.
[pdf]
- Bipartite ranking: risk, optimality, and equivalences. NICTA ML seminar, 2014.
[pdf]
- Latent feature models for dyadic prediction: beyond collaborative filtering and link prediction. NICTA ML seminar, 2013.
[pdf]
- Learning to make predictions in graphs. ID Analytics, 2012.
[pdf]
- Nash equilibria: complexity and computation. USyd INFO4011 guest lecture, 2007.
[pdf]