Estimating Accuracy from Unlabeled Data
By Emmanouil Antonios Platanios, Avrim Blum, Tom Michael Mitchell. 2014
73 citations recorded by the source corpus.
Subfields: 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 Evaluation & Benchmarks
- Abhinav Kumar Gupta
- Jacob Steinhardt
- Matt Gardner
- Tom Michael Mitchell
- Estevam R. Hruschka
- Andrew Carlson
- Justin Betteridge
- Burr Settles
Citations in the corpus
- Concrete Problems in AI Safety (2016, background): “areas that are potentially relevant to this problem, including change detection and anomaly detection [21, 80, 91], hypothesis testing [145], transfer learning [138, 124, 125, 25], and several others [136, 87, 18, 122, 121, 74, 147]. Rather than fully reviewing all of this work in detail (which woul”
- Never-Ending Learning (2015, methodology): “…coupling constraints C = { 〈 φ k , V k 〉 } among the solutions to these learning tasks, where φ k is a real-valued function over two or more learning tasks, specifying the degree of satisfaction of the constraint, and V k is a vector of indices over learning tasks, specifying the arguments to φ k .”
- Unsupervised Risk Estimation with only Structural Assumptions (2016, background): “…et al. (2011) assume Gaussianity of model scores, while a line of work starting with Dawid & Skene (1979) assume multiple classifiers with known dependency structure, which specifies a complete generative distribution for the 0/1-loss (Zhang et al., 2014; Platanios, 2015; Jaffe et al., 2015).”