• Corpus ID: 220936031

Classification from Ambiguity Comparisons

  title={Classification from Ambiguity Comparisons},
  author={Zhenghang Cui and Issei Sato},
Labeling data is an unavoidable pre-processing procedure for most machine learning tasks. However, it takes a considerable amount of time, money, and labor to collect accurate \textit{explicit class labels} for a large dataset. A positivity comparison oracle has been adapted to relieve this burden, where two data points are received as input and the oracle answers which one is more likely to be positive. However, when information about the classification threshold is lacking, this oracle alone… 



Active Classification with Comparison Queries

An extension of active learning in which the learning algorithm may ask the annotator to compare the distances of two examples from the boundary of their label-class is studied, and a combinatorial dimension is identified that captures the query complexity when each additional query is determined by O(1) examples.

Classification from Triplet Comparison Data

This letter proposes an unbiased estimator for the classification risk under the empirical risk minimization framework, which inherently has the advantage that any surrogate loss function and any model, including neural networks, can be easily applied.

Classification from Pairwise Similarity and Unlabeled Data

It is shown that an unbiased estimator of the classification risk can be obtained only from SU data, and the estimation error of its empirical risk minimizer achieves the optimal parametric convergence rate.

On the Minimal Supervision for Training Any Binary Classifier from Only Unlabeled Data

It is proved that it is impossible to estimate the risk of an arbitrary binary classifier in an unbiased manner given a single set of U data, but it becomes possible given two sets of UData with different class priors, which answers a fundamental question---what the minimal supervision is for training any binary classifiers from only U data.

Noise-Tolerant Interactive Learning Using Pairwise Comparisons

This paper presents an algorithm that interactively queries the label and comparison oracles and describes its query complexity under Tsybakov and adversarial noise conditions for the comparison and labeling oracles.

Theory of Disagreement-Based Active Learning

Recent advances in the understanding of the theoretical benefits of active learning are described, and implications for the design of effective active learning algorithms are described.

Classification From Pairwise Similarities/Dissimilarities and Unlabeled Data via Empirical Risk Minimization

This letter derives an unbiased estimator of the classification risk based on all of similarities and dissimilarities and unlabeled data and theoretically establish an estimation error bound and experimentally demonstrate the practical usefulness of the empirical risk minimization method.

Search Improves Label for Active Learning

It is shown that an algorithm using both oracles can provide exponentially large problem-dependent improvements over Label alone, which is stronger than Label while being natural to implement in many situations.

Fast Rates for a kNN Classifier Robust to Unknown Asymmetric Label Noise

A key idea in this analysis is a simple kNN based method for estimating the maximum of a function that requires far less assumptions than existing mode estimators do, and which may be of independent interest for noise proportion estimation and randomised optimisation problems.

Active Learning for Top-K Rank Aggregation from Noisy Comparisons

It is demonstrated that active ranking can offer significant multiplicative gains in sample complexity over passive ranking, depending on the underlying stochastic noise model.