LIBLINEAR: A Library for Large Linear Classification

By Rong-En Fan, Kai-Wei Chang, Cho-Jui Hsieh, Xiangwen Wang, Chih-Jen Lin. 2008

8,472 citations recorded by the source corpus.

Subfields: Artificial Intelligence, Evaluation & Benchmarks

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 Evaluation & Benchmarks

Citations in the corpus

VirTex: Learning Visual Representations from Textual Annotations (2020, methodology): “We use the same hyperparameters as [22]: we use scikitlearn [106] package with LIBLINEAR [107] backend, default parameters are:”
Hypercolumns for object segmentation and fine-grained localization (2014, methodology): “This training methodology does not directly optimize the loss we would encounter at test time, but allows us to use off-the-shelf code such as liblinear [14] to train the logistic regressor.”
DeepWalk: online learning of social representations (2014, methodology): “For all models we use a one-vs-rest logistic regression implemented by LibLinear [11] extended to return the most probable labels as in [39].”
Scaling and Benchmarking Self-Supervised Visual Representation Learning (2019, methodology): “VOC07 and COCO2014: For smaller datasets that fit in memory, we follow [53] and train linear SVMs [7] on the frozen feature representations using LIBLINEAR package [20].”
Unsupervised Domain Adaptation by Backpropagation (2014, methodology): “…the features, and in order to put all the compared methods on an equal footing, we retrain the last layer of the label predictor using a standard linear SVM [4] for all four compared methods (including ours; the performance on the target domain remains approximately the same after the retraining).”
Deep Learning Face Attributes in the Wild (2014, neutral): “SVMs [8] are trained to predict attribute values given each FC.”
Bag of Tricks for Efficient Text Classification (2016, methodology): “Meanwhile, linear classifiers are often considered as strong baselines for text classification problems (Joachims, 1998; McCallum and Nigam, 1998; Fan et al., 2008).”
Domain-Adversarial Training of Neural Networks (2015, methodology): “Since the SA baseline requires training a new classifier after adapting the features, and in order to put all the compared settings on an equal footing, we retrain the last layer of the label predictor using a standard linear SVM (Fan et al., 2008) for all four considered methods (including ours; th”
Learning Rich Features from RGB-D Images for Object Detection and Segmentation (2014, neutral): “0 using liblinear [12].”
Optimization Methods for Large-Scale Machine Learning (2016, background): “Several first-order [56, 55] and second-order [132] gradient projection methods have been proposed to solve (8.”
Certifying and Removing Disparate Impact (2014, methodology): “LR: Logistic Regression: Liblinear’s [4] logistic regression algorithm for L2 regularization and logistic loss.”
Learning Deep Features for Discriminative Localization (2015, methodology): “We find that GoogLeNet-GAP performs comparably to existing approaches, achieving an accuracy of 63.0% when using the full image without any bounding box annotations for both train and test.”
Detect What You Can: Detecting and Representing Objects Using Holistic Models and Body Parts (2014, methodology): “We adopt the code of LIBLINEAR [8] to learn our model parameters.”
A Fast and Accurate Dependency Parser using Neural Networks (2014, methodology): “9 In this comparison, for MaltParser, we select stackproj (arc-standard) and nivreeager (arc-eager) as parsing algorithms, and liblinear (Fan et al., 2008) for optimization.”
NTU RGB+D: A Large Scale Dataset for 3D Human Activity Analysis (2016, methodology): “For the baseline methods which use SVM as their classi-ﬁer, to be able to manage the large scale of the data, we use Libliner SVM toolbox [11].”
On the Effectiveness of Linear Models for One-Class Collaborative Filtering (2016, methodology): “We shall also choose to be the logistic loss, (y, v) = log(1+e−yv), which we favour due to the existence of efficient solvers for linear logistic regression, such as LIBLINEAR (Fan et al. 2008), and its suitability for estimating class-probabilities, which we shall see is appropriate for OC-CF probl”