Inductive conformal predictor for convolutional neural networks: Applications to active learning for image classification

@article{Matiz2019InductiveCP,
  title={Inductive conformal predictor for convolutional neural networks: Applications to active learning for image classification},
  author={Sergio Matiz and Kenneth E. Barner},
  journal={Pattern Recognit.},
  year={2019},
  volume={90},
  pages={172-182}
}
Conformal prediction based active learning
TLDR
Diversity is used to avoid the selection of similar instances, whereas representativeness is used for outlier detection, and novel active learning approaches are proposed, within the conformal prediction framework, considering uncertainty, diversity, and represent ativeness, as the selection criteria.
Batch mode active learning for mitotic phenotypes using conformal prediction
TLDR
This work introduces a batch-based active learning framework in a pooled setting based around conformal predictors that facilitates efficient discovery of rare and novel classes within large screening datasets.
Automatic Open-World Reliability Assessment
TLDR
The open-world recognition reliability problem is formalized and multiple automatic reliability assessment policies are proposed to address this new problem using only the distribution of reported scores/probability data.
Semantic segmentation of high-resolution satellite images using deep learning
TLDR
The developed model can be useful to the authorities for smart city planning and landuse mapping and can further be improved by providing homogeneous features to train the model, balance classes and by incorporating more data set for semantic segmentation.
A Survey on Active Deep Learning: From Model-Driven to Data-Driven
TLDR
This survey categorizes Active Deep Learning into model-driven ADL and data- driven ADL, by whether its selector is model- driven or data-driven, and points out that, with the development of deep learning, the selector in ADL also is experiencing the stage frommodel-driven toData-driven.
From Model-driven to Data-driven: A Survey on Active Deep Learning
...
...

References

SHOWING 1-10 OF 55 REFERENCES
Effective utilization of data in inductive conformal prediction using ensembles of neural networks
TLDR
Experiments on 34 publicly available data sets conclusively show that the use of out-of-bag estimates produced the most efficient conformal predictors, making it the obvious preferred choice for ensembles in the conformal prediction framework.
Cost-Effective Active Learning for Deep Image Classification
TLDR
This paper proposes a novel active learning (AL) framework, which is capable of building a competitive classifier with optimal feature representation via a limited amount of labeled training instances in an incremental learning manner and incorporates deep convolutional neural networks into AL.
Model-agnostic nonconformity functions for conformal classification
TLDR
Three model-agnostic nonconformity functions, based on well-known loss functions, are evaluated with regard to how they affect efficiency on 21 publicly available multi-class data sets, showing that the choice of non Conformity function has a major impact on the efficiency, but also that different nonconforming functions should be used depending on the exact efficiency metric.
Training Deep Convolutional Neural Networks with Active Learning for Exudate Classification in Eye Fundus Images
TLDR
This work presents a label-efficient CNN model using the expected gradient length, an active learning algorithm to select the most informative patches and images, converging earlier and to a better local optimum than the usual SGD (Stochastic Gradient Descent) strategy.
Conformal Prediction for Reliable Machine Learning: Theory, Adaptations and Applications
The conformal predictions framework is a recent development in machine learning that can associate a reliable measure of confidence with a prediction in any real-world pattern recognition
Generalized Query by Transduction for online active learning
TLDR
The proposed Generalized Query by Transduction approach can be used along with any existing pattern classification algorithm, and can also be used to combine multiple criteria in selecting an unlabeled example appropriately in the active learning process.
Conformal Prediction with Neural Networks
TLDR
A modification of the original CP method, called inductive conformal prediction (ICP), is used, which allows for a neural network confidence predictor without the massive computational overhead of CP.
Active learning combining uncertainty and diversity for multi-class image classification
TLDR
A novel active learning algorithm is proposed here based on one-versus-one strategy support vector machine (SVM) to solve multi-class image classification and a new uncertainty measure is proposed based on some binary SVM classifiers and some of the most uncertain examples are selected from SVM output.
Visualizing and Understanding Convolutional Networks
TLDR
A novel visualization technique is introduced that gives insight into the function of intermediate feature layers and the operation of the classifier in large Convolutional Network models, used in a diagnostic role to find model architectures that outperform Krizhevsky et al on the ImageNet classification benchmark.
...
...