Uncovering shared structures in multiclass classification

Amit, Yonatan; Fink, Michael; Srebro, Nathan; Ullman, Shimon

doi:10.1145/1273496.1273499

Cited by 229 publications

(269 citation statements)

References 12 publications

Supporting

Mentioning

267

Contrasting

Order By: Relevance

“…Our word prediction formulation of the loss is different from [21] (a pure collaborative filtering model) and [1] (a multi-class classifier), even though our tracenorm regularization term is similar to theirs. Our formulation is, to the best of our knowledge, the first application of the tracenorm regularization to a problem of these characteristics.…”

Section: A Max-margin Factorization Modelmentioning

confidence: 99%

“…Our formulation is, to the best of our knowledge, the first application of the tracenorm regularization to a problem of these characteristics. From [1] we took the optimization framework, although we are using different losses and approximations and we are using BFGS to perform the minimization. Finally, we introduce a unsupervised model on top of the internal representation this formulation produces to discover scenes.…”

Section: A Max-margin Factorization Modelmentioning

confidence: 99%

“…Theorem 1 in [1] can be applied with small modifications to prove that there exists a matrix α ∈ R M×N so that the minimizer of (3) is W = Xα. But instead of solving the dual Lagrangian problem we will use this representation of W to minimize the primal problem (actually, it's smoothed version) using gradient descent.…”

Section: Kernelizationmentioning

confidence: 99%

“…We will achieve this goal by using matrix factorization [21,1] to learn a word classifier. Let Y be a matrix of word annotations per image, X the matrix of image features per image, and W a linear classifier matrix, we will look for W to minimize…”

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

Scene Discovery by Matrix Factorization

Loeff

Farhadi

2008

Lecture Notes in Computer Science

View full text Add to dashboard Cite

Abstract. What constitutes a scene? Defining a meaningful vocabulary for scene discovery is a challenging problem that has important consequences for object recognition. We consider scenes to depict correlated objects and present visual similarity. We introduce a max-margin factorization model that finds a low dimensional subspace with high discriminative power for correlated annotations. We postulate this space should allow us to discover a large number of scenes in unsupervised data; we show scene discrimination results on par with supervised approaches. This model also produces state of the art word prediction results including good annotation completion.

show abstract

Section: A Max-margin Factorization Modelmentioning

confidence: 99%

Section: A Max-margin Factorization Modelmentioning

confidence: 99%

Section: Kernelizationmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Scene Discovery by Matrix Factorization

Loeff

Farhadi

2008

Lecture Notes in Computer Science

View full text Add to dashboard Cite

show abstract

“…where [15,24], we use the trace norm regularization for simultaneous feature extraction and model learning. As the prediction are dependent on F E H and F G G, we can add the Frobenius norm regularizer in the above EM algorithm to control the magnitude of F E , Fg, H and G. Adding the regularization terms will not change the posterior distribution of the latent variable z.…”

Section: Mixture Of Experts Model With the Trace Norm Regularizationmentioning

confidence: 99%

Multi-task feature learning by using trace norm regularization

Zhang

et al. 2017

Open Physics

View full text Add to dashboard Cite

Abstract:Multi-task learning can extract the correlation of multiple related machine learning problems to improve performance. This paper considers applying the multi-task learning method to learn a single task. We propose a new learning approach, which employs the mixture of expert model to divide a learning task into several related subtasks, and then uses the trace norm regularization to extract common feature representation of these sub-tasks. A nonlinear extension of this approach by using kernel is also provided. Experiments conducted on both simulated and real data sets demonstrate the advantage of the proposed approach.

show abstract

Protein structure determination using a Riemannian approach

Wei

et al. 2019

FEBS Letters

View full text Add to dashboard Cite

Protein NMR structure determination is one of the most extensively studied problems. Here, we adopt a novel method based on a matrix completion technique – the Riemannian approach – to rebuild the protein structure from the nuclear Overhauser effect distance restraints and the dihedral angle restraints. In comparison with the cyana method, the results generated via the Riemannian approach are more similar to the standard X‐ray crystallographic structures as a result of the simple but powerful internal calculation processing function. In addition, our results demonstrate that the Riemannian approach has a comparable or even better performance than the cyana method on other structural assessment metrics, including the stereochemical quality and restraint violations. The Riemannian approach software is available at: https://github.com/xubiaopeng/Protein_Recon_MCRiemman.

show abstract

Uncovering shared structures in multiclass classification

Cited by 229 publications

References 12 publications

Scene Discovery by Matrix Factorization

Scene Discovery by Matrix Factorization

Multi-task feature learning by using trace norm regularization

Protein structure determination using a Riemannian approach

Contact Info

Product

Resources

About