Pruning training sets for learning of object categories

By Anelia Angelova, Yaser S. Abu-Mostafa, Pietro Perona. 2005

179 citations recorded by the source corpus.

Subfields: Artificial Intelligence, Computer Vision and Pattern Recognition

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 Computer Vision and Pattern Recognition

Citations in the corpus

Dataset Distillation (2018, neutral): “Another way to distill knowledge is to summarize the entire dataset by a small subset, either by only using the “valuable” data for model training (Angelova et al., 2005; Felzenszwalb et al., 2010; Lapedriza et al., 2013) or by only labeling the “valuable” data via active learning (Cohn et al.,…”
Distilled One-Shot Federated Learning (2020, methodology): “These methods include dataset pruning [1] and core set construction [2, 34, 30], which keep the examples that are measured to be more useful for training and remove the rest.”
AgeDB: The First Manually Collected, In-the-Wild Age Database (2017, methodology): “Some of the most widely used databases which included images captured under controlled conditions were the XM2VTS database [30], the Multi-PIE database [13, 14], the AR face database [27], the Caltech faces database [2], the FERET database [35, 34, 38], the Yale face database [11].”
Robust Loss Functions for Learning Multi-class Classifiers (2018, methodology): “Several heuristics have been used to guess such noisy points [4]–[6].”
Classification in the Presence of Label Noise: A Survey (2014, background): “In fact, as discussed in [170], classiﬁcation ﬁltering (and data cleansing in general) suffers from a chicken-and-egg dilemma, since: 1) good classiﬁers are necessary for classiﬁcation ﬁltering and 2) learning in the presence of label noise may precisely produce poor classiﬁers.”
Soft-Label Dataset Distillation and Text Dataset Distillation (2019, methodology): “Several methods aim to ‘prune’ a dataset, or create a ‘core-set’, by leaving in only examples that are determined to be useful [13]–[16].”
Are all training examples equally valuable? (2013, methodology): “Angelova et al. [2] propose a method of pruning a training set and show its effectiveness in removing outliers and hard-to-learn examples.”
Learning with Confident Examples: Rank Pruning for Robust Classification with Noisy Labels (2017, background): “Procedural efforts have improved robustness to mislabeling in the context of machine vision (Xiao et al., 2015), neural networks (Reed et al., 2015), and face recognition (Angelova et al., 2005).”
Complete Random Forest Based Class Noise Filtering Learning for Improving the Generalizability of Classifiers (2019, background): “However, as previously discussed in [31], the misclassified samples contain the class noise, but the classification results of distinct classifiers are different, i.”
Flexible Dataset Distillation: Learn Labels Instead of Images (2020, background): “In this sense it is related to other endeavours such as dataset pruning [2, 10, 16], core-set construction [29, 3, 26] and instance selection [22] that focuses on dataset summarization through a small number of examples.”
Robust Loss Functions under Label Noise for Deep Neural Networks (2017, methodology): “Several heuristics have been used to guess such noisy points [1, 6, 31].”
Labeled Faces in the Wild: A Database forStudying Face Recognition in Unconstrained Environments (2008, background): “Caltech Faces 27 450 lighting, expression, background Caltech 10000 Web Faces ≈ 10000 10000 wide variability, facial features annotated [1]”
Annotated Facial Landmarks in the Wild: A large-scale, real-world database for facial landmark localization (2011, background): “Caltech 10,000 Web Faces 10,524 4 - - color [2] CMU / VASC Frontal 734 6 - - grayscale [26] CMU / VASC Profile 590 6 to 9 - - grayscale [27] IMM 240 58 40 648x480 color/grayscale [23] MUG 401 80 26 896x896 color [22] AR Purdue 508 22 116 768x576 color [18] BioID 1,521 20 23 384x286 grayscale [15] XM”
Combating Label Noise in Deep Learning Using Abstention (2019, background): “While data cleaning (or pruning) approaches have been considered before in the context of shallow classiﬁers (Angelova et al., 2005; Brodley & Friedl, 1999; Zhu et al., 2003), to the best of our knowledge, this is the ﬁrst work to show how ab-Combating stention training can be used to identify and…”
Cascaded pose regression (2010, methodology): “For the Faces dataset we used Caltech 10,000 Web Faces [3], where each face was labeled by four points (eyes, mouth, nose).”
Global-Local GCN: Large-Scale Label Noise Cleansing for Face Recognition (2020, methodology): “[5] adopt data pruning to build a dataset.”