2012 IEEE Conference on Computer Vision and Pattern Recognition (2012)
Providence, RI USA
June 16, 2012 to June 21, 2012
While Boltzmann Machines have been successful at unsupervised learning and density modeling of images and speech data, they can be very sensitive to noise in the data. In this paper, we introduce a novel model, the Robust Boltzmann Machine (RoBM), which allows Boltzmann Machines to be robust to corruptions. In the domain of visual recognition, the RoBM is able to accurately deal with occlusions and noise by using multiplicative gating to induce a scale mixture of Gaussians over pixels. Image denoising and in-painting correspond to posterior inference in the RoBM. Our model is trained in an unsupervised fashion with unlabeled noisy data and can learn the spatial structure of the occluders. Compared to standard algorithms, the RoBM is significantly better at recognition and denoising on several face databases.
learning (artificial intelligence), Boltzmann machines, image denoising, image recognition, image reconstruction, inference mechanisms, face databases, robust Boltzmann machines, supervised learning, density modeling, speech data, RoBM, visual recognition, occlusions, multiplicative gating, Gaussian scale mixture, image denoising, image inpainting, posterior inference, Noise, Face, Robustness, Data models, Noise measurement, Databases, Training
G. Hinton, R. Salakhutdinov and Yichuan Tang, "Robust Boltzmann Machines for recognition and denoising," 2012 IEEE Conference on Computer Vision and Pattern Recognition(CVPR), Providence, RI USA, 2012, pp. 2264-2271.