Cutting-plane training of structural SVMs

By Thorsten Joachims, Thomas Finley, Chun-Nam John Yu. 2009

1,126 citations recorded by the source corpus.

Subfields: Artificial Intelligence, Molecular Biology, Optimization & Training

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In Molecular Biology

In Optimization & Training

Citations in the corpus

Asymmetric LSH (ALSH) for Sublinear Time Maximum Inner Product Search (MIPS) (2014, methodology): “Latent factor modeling based on matrix factorization [14] is a popular approach for solving collaborative filtering.”
Learning Trajectory Preferences for Manipulators via Iterative Improvement (2013, methodology): “We quantify closeness of trajectories by the l2−norm of difference in their feature representations, and choose the regularization parameter C for training SSVM in hindsight, to give an unfair advantage to MMP-online.”
Minding the Gaps for Block Frank-Wolfe Optimization of Structured SVMs (2016, methodology): “Problem (47) is often referred to as the n-slack structured SVM with margin-rescaling (Joachims et al., 2009, Optimization Problem 2).”
Block-Coordinate Frank-Wolfe Optimization for Structural SVMs (2012, methodology): “A MATLAB wrapper of SVMstruct. http://www.vlfeat.org/~vedaldi/code/ svm-struct-matlab.html, 2011.”
Learning Deep Structure-Preserving Image-Text Embeddings (2015, positive): “Inspired by [21, 40], instead of choosing the most violating negative match in all instance space, we select top K most violated matches in each mini-batch.”
Learning Optimized Risk Scores (2016, background): “Cutting plane algorithms have an impressive track record on large-scale supervised learning problems, as they scale linearly with the number of samples and provide precise control over data-related computation (see e.g., Teo et al., 2009; Franc and Sonnenburg, 2009; Joachims et al., 2009).”
Learning human activities and object affordances from RGB-D videos (2012, methodology): “…2009): min w,ξ 1 2 wTw + Cξ (14) s.t. ∀ȳ1, ..., ȳM ∈ {0, 0.5, 1}N·K : 1 M wT M∑ m=1 [Ψ(xm,ym)−Ψ(xm, ȳm)] ≥ ∆(ym, ȳm)− ξ While the number of constraints in this QP is exponential in M , N and K, it can nevertheless be solved efficiently using the cutting-plane algorithm (Joachims et al., 2009).”
Deep Metric Learning via Lifted Structured Feature Embedding (2015, neutral): “The experiments are performed with GoogLeNet [33].”
Structural-RNN: Deep Learning on Spatio-Temporal Graphs (2015, methodology): “This has been the incentive for many efforts in computer vision and machine learning, such as Logic Networks [46], Graphical Models [28], and Structured SVMs [26].”
Learning Two-Branch Neural Networks for Image-Text Matching Tasks (2017, methodology): “Similarly to [1], [32], [37], we use SGD to optimize the loss function and sample triplets within each mini-batch and our sampling strategy is loosely inspired by [40], [55].”
CRAFT Objects from Images (2016, background): “In order to get an end-to-end system, Fast R-CNN drops the one-vs-rest SVM in R-CNN, which creates the gap between the resulting solution and the task demand.”
A Hierarchical Pose-Based Approach to Complex Action Understanding Using Dictionaries of Actionlets and Motion Poselets (2016, methodology): “Then, we solve for W via 1-slack formulation using Cutting Plane algorithm [12].”
A Survey on Metric Learning for Feature Vectors and Structured Data (2013, background): “…the number of constraints is super-exponential in the number of training instances, the authors solve (11) using a 1-slack cutting-plane approach (Joachims et al., 2009) which essentially iteratively optimizes over a small set of active constraints (adding the most violated ones at each step)…”
Multi-agent Diverse Generative Adversarial Networks (2017, methodology): “The intuition behind using these constraints comes from the well know 1 -slack formulation of the structured SVM framework [12, 24].”
Inferring the Why in Images (2014, methodology): “This optimization is equivalent to a structured SVM and can be solved by efficient off-the-shelf solvers [13].”
Learning 3D Part Detection from Sparsely Labeled Data (2014, methodology): “11 we utilize the cutting plane method [11], which during optimization requires the computation of the “most violated” constraint:”