JMLR2019.md

JMLR 2019 and before

Year	Title	Author	Publication	Code	Tasks	Notes	Datasets	Notions
2019	Active Learning for Cost-Sensitive Classification	Krishnamurthy et al.	JMLR	-	Cost-Sensitive Classification	Uncertainty, MLP, None, Tra, Hard	ImageNet 20, Imagenet 40, RCV1-v2, POS, NER, NER-wiki	COAL assumes access to a set of regression functions, and, when processing an example x, it uses the functions with good past performance to compute the range of possible costs that each label might take. Naturally, COAL only queries labels with large cost range, akin to uncertainty-based approaches in active regres- sion (Castro et al., 2005), but furthermore, it only queries labels that could possibly have the smallest cost, avoiding the uncertain, but surely suboptimal labels. The key algorith- mic innovation is an efficient way to compute the cost range realized by good regressors.
2019	SMART: An Open Source Data Labeling Platform for Supervised Learning	Chew et al.	JMLR	code	human annotation platform	SMART
2019	The Relationship Between Agnostic Selective Classification, Active Learning and the Disagreement Coefficient	Gelbhart and El-Yaniv	JMLR	-	stream-based	Disagreement Coefficient,		ILESS. Active-ILESS is constructed to work in a stream-based AL model and its querying function is extremely conservative:
2018	Active Nearest-Neighbor Learning in Metric Spaces	Kontorovich et al.	JMLR	-	Thory	estimating the sample error, BNNs, None, Tra, Hard		prove that the label complexity of MARMANN is significantly lower than that of any passive learner with similar error guarantees.
2017	Asymptotic Analysis of Objectives Based on Fisher Information in Active Learning	Sourati et al.	JMLR	-	Thory	Fisher Information		we attempt to fill this gap and provide a rigorous framework for analyzing existing FIR-based active learning methods. In particular, we show that FIR can be asymptotically viewed as an upper bound of the expected variance of the log-likelihood ratio.