Timezone: »
Multiple-Instance learning has been long known as a hard non-convex problem. In this work, we propose an approach that recasts it as a convex likelihood ratio estimation problem. Firstly, the constraint in multiple-instance learning is reformulated into a convex constraint on the likelihood ratio. Then we show that a joint estimation of a likelihood ratio function and the likelihood on training instances can be learned convexly. Theoretically, we prove a quantitative relationship between the risk estimated under the 0-1 classification loss, and under a loss function for likelihood ratio estimation. It is shown that our likelihood ratio estimation is generally a good surrogate for the 0-1 loss, and separates positive and negative instances well. However with the joint estimation it tends to underestimate the likelihood of an example to be positive. We propose to use these likelihood ratio estimates as features, and learn a linear combination on them to classify the bags. Experiments on synthetic and real datasets show the superiority of the approach.
Author Information
Fuxin Li (University of Bonn)
Cristian Sminchisescu (Lund University/Google)
More from the Same Authors
-
2021 Spotlight: H-NeRF: Neural Radiance Fields for Rendering and Temporal Reconstruction of Humans in Motion »
Hongyi Xu · Thiemo Alldieck · Cristian Sminchisescu -
2021 Poster: Relative Flatness and Generalization »
Henning Petzka · Michael Kamp · Linara Adilova · Cristian Sminchisescu · Mario Boley -
2021 Poster: REMIPS: Physically Consistent 3D Reconstruction of Multiple Interacting People under Weak Supervision »
Mihai Fieraru · Mihai Zanfir · Teodor Szente · Eduard Bazavan · Vlad Olaru · Cristian Sminchisescu -
2021 Poster: H-NeRF: Neural Radiance Fields for Rendering and Temporal Reconstruction of Humans in Motion »
Hongyi Xu · Thiemo Alldieck · Cristian Sminchisescu -
2011 Poster: Probabilistic Joint Image Segmentation and Labeling »
Adrian Ion · Joao Carreira · Cristian Sminchisescu -
2011 Spotlight: Probabilistic Joint Image Segmentation and Labeling »
Adrian Ion · Joao Carreira · Cristian Sminchisescu -
2009 Poster: Efficient Match Kernel between Sets of Features for Visual Recognition »
Liefeng Bo · Cristian Sminchisescu -
2009 Spotlight: Efficient Match Kernel between Sets of Features for Visual Recognition »
Liefeng Bo · Cristian Sminchisescu -
2007 Poster: People Tracking with the Laplacian Eigenmaps Latent Variable Model »
Zhengdong Lu · Miguel A. Carreira-Perpinan · Cristian Sminchisescu