Proceedings of the Twenty-Second International Joint Conference on Artificial Intelligence Large Linear Classification When Data Cannot Fit in Memory

Citations in the corpus

• CoCoA: A General Framework for Communication-Efficient Distributed Optimization (2016, methodology): “By making use of the primal-dual structure in the line of work of Yu et al. (2012); Pechyony et al. (2011); Yang (2013); Yang et al. (2013) and Lee and Roth (2015), the CoCoA-v1 and CoCoA+ frameworks (which are special cases of the presented framework, CoCoA) are the first to allow the use of any…”
• Communication-Efficient Distributed Dual Coordinate Ascent (2014, methodology): “While no convergence rate was given, the empirical results in the journal paper [12] suggest that running LibLinear for just one pass through the local data performs well in practice.”
• Least Squares Revisited: Scalable Approaches for Multi-class Prediction (2013, methodology): “There are also related approaches for training SVMs that extract the most information out of a small subset of data before moving to the next batch (Chapelle, 2007; Matsushima et al., 2012; Yu et al., 2012).”
• Large-scale image classification: Fast feature extraction and SVM training (2011, background): “Indeed, LIBLINEAR recently released an extended version that explicitly considered the memory issue [25].”
• Distributed Box-Constrained Quadratic Optimization for Dual Linear SVM (2015, unknown): “For solving each sub-problem, we use the cyclic approach, but at each inner iteration 1 , we follow the approach in Hsieh et al. (2008); Yu et al. (2012); Chang & Roth (2011) to randomly shuffle the instance order, to have faster empirical convergence.”
• Object Recognition by Sequential Figure-Ground Ranking (2011, background): “Some progress has been made in designing large-scale SVM classifiers (Yu et al. 2010), but those generally require loading the data into memory multiple times and are extremely time-consuming.”
• A Dual Augmented Block Minimization Framework for Learning with Limited Memory (2015, methodology): “Therefore, balancing the time spent on I/O and computation becomes the key issue [26].”
• Solving Large Scale Linear SVM with Distributed Block Minimization (2011, methodology): “[19] showed that (1) can be solved using sequential block minimization (SBM): at each iteration we consider a block Bi and solve (1) only for the variables in Bi.”
• Efficient Distributed Linear Classification Algorithms via the Alternating Direction Method of Multipliers (2012, methodology): “Now, we study how our distributed algorithms compare against Block LIBLINEAR running in a single machine in terms of training time.”
• Efficient Use of Limited-Memory Accelerators for Linear Learning on Heterogeneous Systems (2017, methodology): “In contrast to the related line of research [16, 2, 7], our scheme is designed to take full advantage of both compute resources A and B for training, by systematically splitting the workload amongA and B in order to adapt to their specific properties and to the available bandwidth between them.”