Active learning

By David Cohn. 2013

2,217 citations recorded by the source corpus.

Subfields: Artificial Intelligence, Management Science and Operations Research, Representation & Self-Supervised Learning

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 Management Science and Operations Research

In Representation & Self-Supervised Learning

Citations in the corpus

Snorkel: Rapid Training Data Creation with Weak Supervision (2017, background): “Work on active learning aims to automatically estimate which data points are optimal to label, thereby hopefully reducing the total number of examples that need to be manually annotated [45].”
Scalable agent alignment via reward modeling: a research direction (2018, methodology): “During training, it can help automate the process of soliciting feedback about the most informative states (Krueger et al., 2016; Schulze & Evans, 2018) using active learning (Settles, 2012).”
Fast and Flexible Multi-Task Classification Using Conditional Neural Adaptive Processes (2019, background): “Active learning [46, 47] requires accurate dataefficient learning that returns well-calibrated uncertainty estimates.”
Prediction-Based Decisions and Fairness: A Catalogue of Choices, Assumptions, and Definitions (2018, background): “Measurement error in Y is sometimes called noisy labels [127], mislabeled data [17, 16], or unreliable oracles [154].”
Machine Teaching: An Inverse Problem to Machine Learning and an Approach Toward Optimal Education (2015, background): “Recall active learning allows the learner to “ask questions” by selecting items x and asking an oracle for its label y (Settles 2012).”
COBRA: Data-Efficient Model-Based RL through Unsupervised Object Discovery and Curiosity-Driven Exploration (2019, background): “Drawing inspiration from self-supervised “play” in human development (Gopnik et al., 1999; Settles, 2011), we introduce an agent that learns object-centric representations of its environment without supervision and subsequently harnesses these to learn policies eﬃciency and robustly.”
ERASER: A Benchmark to Evaluate Rationalized NLP Models (2019, background): “Active learning (AL) (Settles, 2012) is a complementary strategy for reducing annotator effort that entails the model selecting the examples with which it is to be trained.”
Meta-Learning in Neural Networks: A Survey (2020, background): “AL is a well studied problem with numerous hand designed algorithms [284].”
Trial without Error: Towards Safe Reinforcement Learning via Human Intervention (2017, background): “Techniques from active learning and anomaly detection can be used to detect unfamiliar states [25, 17, 9].”
Interactive machine learning for health informatics: when do we need the human-in-the-loop? (2016, background): “Active learning is well-motivated in many modern machine learning problems, where unlabeled data may be abundant or easily obtained, but labels are difﬁcult, time-consuming, or expensive to obtain [73].”
Taking Human out of Learning Applications: A Survey on Automated Machine Learning (2018, background): “For example, graph embedding [162], active learning [163], and semi-supervised learning [164].”
Towards better analysis of machine learning models: A visual analytics perspective (2017, background): “From the machine learning community, researchers have developed a wide array of techniques for system initiated guidance [45, 46, 47, 48], where the system plays a more active role, for example, by making suggestions about appropriate views or next steps in the iterative and progressive analysis pro”
Open challenges for data stream mining research (2014, background): “A related problem discussed below is active learning (AL, see [38]).”
The power of localization for efficiently learning linear separators with noise (2013, methodology): “Learning linear separators from correctly labeled (non-noisy) examples is a very well understood problem with simple efﬁcient algorithms like Perceptron being effective both in the classic passive learning setting [33, 42] and in the more modern active learning framework [21].”
Snorkel: rapid training data creation with weak supervision (2019, background): “examples that need to be manually annotated [54].”
Towards an Intelligent Edge: Wireless Communication Meets Machine Learning (2018, background): “Principally, active learning is to select important samples from a large unlabelled dataset for labelling (by querying an oracle) so as to accelerate model training with a labelling budget [7].”
Decomposition of Uncertainty in Bayesian Deep Learning for Efficient and Risk-sensitive Learning (2017, background): “Active learning is the problem of iteratively collecting data so that the final gains in predictive performance are as high as possible (Settles, 2012).”