Weakly-Paired Maximum Covariance Analysis
نویسندگان
چکیده
We study the problem of multimodal dimensionality reduc6 6 tion assuming that data samples can be missing at training time, and 7 7 not all data modalities may be present at application time. Maximum 8 8 covariance analysis, as a generalization of PCA, has many desirable prop9 9 erties, but its application to practical problems is limited by its need for 10 10 perfectly paired data. We overcome this limitation by a latent variable 11 11 approach that allows working with weakly paired data and is still able 12 12 to efficiently process large datasets using standard numerical routines. 13 13 The resulting weakly paired maximum covariance analysis often finds 14 14 better representations than alternative methods, as we show in two ex15 15 emplary tasks: texture discrimination and transfer learning. 16 16
منابع مشابه
Weakly-Paired Maximum Covariance Analysis for Multimodal Dimensionality Reduction and Transfer Learning
We study the problem of multimodal dimensionality reduction assuming that data samples can be missing at training time, and not all data modalities may be present at application time. Maximum covariance analysis, as a generalization of PCA, has many desirable properties, but its application to practical problems is limited by its need for perfectly paired data. We overcome this limitation by a ...
متن کاملViTac: Feature Sharing between Vision and Tactile Sensing for Cloth Texture Recognition
Vision and touch are two of the important sensing modalities for humans and they offer complementary information for sensing the environment. Robots could also benefit from such multi-modal sensing ability. In this paper, addressing for the first time (to the best of our knowledge) texture recognition from tactile images and vision, we propose a new fusion method named Deep Maximum Covariance A...
متن کاملWeakly Informative Prior for Covariance Matrices 1 Running head: WEAKLY INFORMATIVE PRIOR FOR COVARIANCE MATRICES Weakly Informative Prior for Point Estimation of Covariance Matrices in Hierarchical Models
When fitting hierarchical regression models, maximum likelihood estimation has computational (and, for some users, philosophical) advantages compared with full Bayesian inference, but when the number of groups is small, estimates of the covariance matrix (Σ) of group-level varying coefficients are often degenerate. One can do better, even from a purely point-estimation perspective, by using a p...
متن کاملTesting the equality of mean vectors for paired doubly multivariate observations in blocked compound symmetric covariance matrix setup
In this article we develop a test statistic for testing the equality of mean vectors for paired doubly multivariate observations for q response variables and u sites in blocked compound symmetric covariance matrix setting. The new test is implemented with two real data sets.
متن کاملThurstonian covariance and correlation structures for multiple judgment paired comparison data
Thurstone's (1927) is not a proper model for multiple judgment paired comparison data as it assigns zero probabilities to all intransitive patterns. To obtain a proper model, Takane (1987) extended Thurstone's model by adding a vector of pair specific random errors. We investigate an unrestricted Thurstone-Takane model when (a) the variances of the paired specific errors are assumed to be equal...
متن کامل