Multi-Annotator Probabilistic Active Learning.

To deal with this problem, we propose the MAC (Multi-label Active learning from Crowds) approach which incorporate the local influence of label correlations to build a probabilistic model over the multi-label ... Multi-label active learning is a hot topic in reducing the label cost by optimally choosing the most valuable instance to query its label from an oracle. ... build a probabilistic model for multi-label classifier and annotators. ...

arXiv:1508.00722v1 fatcat:rlqvdd25ebbmjby7d3p6a5qvhi

In this paper, we employ a probabilistic model for learning from multiple annotators that can also learn the annotator expertise even when their expertise may not be consistently accurate across the task ... This paradigm posits new challenges to the active learning scenario. ... As shown in all six figures, active learning combined with the probabilistic multi-labeler model (indicated as active learning+multi-labeler) maintained the best performance under both accuracy and AUC ...

dblp:conf/icml/YanRFD11 fatcat:ybjlnsguobf4dl2fmrvrh6axdm

We thus aim to model the mapping between observations and interaction classes, as well as class overlaps, towards a probabilistic multi-label classifier that emulates human annotators. ... We outper- form conventional single-label classification by 11% and 6% on the two datasets respectively, and show that learning from annotation probabilities outperforms majority voting and enables discovery ... In the second probabilistic approach, which we propose here, we not only allow multiple annotation verbs but wish to learn the probability of a human using that verb to annotate the video. ...

arXiv:1703.08338v2 fatcat:a6rl5iglw5cdrdbszwz6kcvjpe

Multiple Versions

However, less research attention has been focused on multi-class active learning. ... Active learning is an effective way to relieve the tedious work of manual annotation in many applications of visual recognition. ... Multi-class active learning. The existing multi-class active learning approaches can be divided to two categories. ...

doi:10.1109/iccv.2015.325 dblp:conf/iccv/LongH15 fatcat:fq7wgh3aevhmtc4grqj7d3yhqm

In this paper, we address multi-labeler active learning, where data labels can be acquired from multiple labelers with various levels of expertise. ... To solve this problem, we propose a new probabilistic model that transfers knowledge from a rich set of labeled instances in some auxiliary domains to help model labelers' expertise for active learning ... Knowledge Transfer for Active Learning Based on our probabilistic model, multi-labeler active learning seeks to select the most informative instance and the most appropriate labeler, with respect to the ...

doi:10.1007/978-3-642-40988-2_18 fatcat:dfgsjkoisjb5zmzsm62o3z5kh4

However, computational time complexity limits prevalent probabilistic methods from effectively supporting active learning. ... Active learning methods, like uncertainty sampling, combined with probabilistic prediction techniques have achieved success in various problems like image classification and text classification. ... Active learning (Settles, 2008; seeks to reduce this annotation burden by requesting the most useful annotations for learning. ...

dblp:conf/icml/BehpourLZ19 fatcat:3kienvxtwnfflgmlpa7mtuz744

The goal of this article is to propose an optimization framework by acknowledging human factors to enable label acquisition through active learning . ... In particular, we are interested to investigate tasks, such as, providing (collecting or acquiring) and validating labels, or comparing data using active learning techniques. ... Multi-label active learning methods simply extend this binary uncertainty concept into the multi-label learning scenarios by integrating the binary uncertainty measures associated with each individual ...

dblp:journals/debu/Roy19 fatcat:3rkclfrdvfablhpjiftguqtqx4

We further show how the two separate tasks of location prediction and activity recognition can be fused into a single model, which allows for simultaneous learning an prediction for both tasks. ... activity. ... The project is actively working towards releasing high-quality data sets to encourage community participation in tackling the issues outlined here. ...

arXiv:1702.01209v1 fatcat:kyj2x6ei4vhqzdsdbgf5tkdyzu

Open Access

Building on prior work on tractable probabilistic circuits, we illustrate how CRISPs enable robust and efficient active and skeptical learning in large structured output spaces. ... In this position paper, we study interactive learning for structured output spaces, with a focus on active learning, in which labels are unknown and must be acquired, and on skeptical learning, in which ... Fine-grained active learning. In applications with large output spaces, the cost of annotating a full label y can be excessive. ...

arXiv:2202.08566v1 fatcat:y4p27ribaje2jfgo46ybl3bmqy

We show that class-order feedback and active learning can reduce the annotation cost both individually and jointly. ... In this paper, we study the problem of learning multi-class classification models from a limited set of labeled examples obtained from human annotator. ... Active learning In active learning, model training and data instance annotation process are interleaved. ...

doi:10.1609/aaai.v33i01.33015589 fatcat:7wra52ogxzg2pjrckwus5vio7e

We show that class-order feedback and active learning can reduce the annotation cost both individually and jointly. ... In this paper, we study the problem of learning multi-class classification models from a limited set of labeled examples obtained from human annotator. ... Active learning In active learning, model training and data instance annotation process are interleaved. ...

pmid:31750011 pmcid:PMC6867686 fatcat:czi672qjmrej3e5kaaoytr6f6m

Mass spectrometry-based spatial proteomics and contemporary machine learning enable to build proteome-wide spatial maps, informing us on the location of thousands of proteins. ... Recent advances enable us to probabilistically model protein localisation as well as quantify the uncertainty in the location assignments, thus leading to better and more trustworthy biological interpretation ... Acknowledgements Mr Oliver Crook and Dr Lisa Breckels, Computational Proteomics Unit, Cambridge (machine learning, algorithms, software). Dr Sebastian Gibb and ...

doi:10.5281/zenodo.1435058 fatcat:ps4srg3fe5gbdj2oezmw6zxl5q

Open Access

This paper studies the active learning problem in crowdsourcing settings, where multiple imperfect annotators with varying levels of expertise are available for labeling the data in a given task. ... Experiments on both text and image datasets demonstrate that our proposed method outperforms other state-of-the-art active learning methods. ... Motivated by these observations, we propose a novel probabilistic model that addresses the active learning problem in crowdsourcing settings, where multiple cheap labelers work together for data annotation ...

doi:10.1609/aaai.v28i1.8993 fatcat:2ynw4ua6izboniesuebtbfsuw4

However, computational time complexity limits prevalent probabilistic methods from effectively supporting active learning. ... Active learning methods, like uncertainty sampling, combined with probabilistic prediction techniques have achieved success in various problems like image classification and text classification. ... Active learning (Settles 2012) seeks to reduce this annotation burden by only requesting the annotations that are most useful for learning. ...

arXiv:1912.12557v3 fatcat:sznqd3qrrbdwbpwt3alswbxpl4

Multiple Versions

probabilistically estimates the state at each genomic position and derives a representative chromatin state map for the group. ... CSREP uses an ensemble of multi-class logistic regression classifiers to predict the chromatin state assignment of each sample given the state maps from all other samples. ... CSREP does this by first generating probabilistic estimates of chromatin state annotations by using an ensemble of multi-class logistic regression classifiers that predict the state assignment in a sample ...

doi:10.1101/2022.05.08.491094 fatcat:ex25v7tfzbdr3b2waegfexmkuy

Multi-Label Active Learning from Crowds [article]

Preserved Fulltext

Active Learning from Crowds

Preserved Fulltext

Improving Classification by Improving Labelling: Introducing Probabilistic Multi-Label Object Interaction Recognition [article]

Preserved Fulltext

Other Versions

Multi-class Multi-annotator Active Learning with Robust Gaussian Process for Visual Recognition

Preserved Fulltext

Knowledge Transfer for Multi-labeler Active Learning [chapter]

Preserved Fulltext

Active Learning for Probabilistic Structured Prediction of Cuts and Matchings

Preserved Fulltext

Capturing Human Factors to Optimize Crowdsourced Label Acquisition through Active Learning

Preserved Fulltext

Probabilistic Sensor Fusion for Ambient Assisted Living [article]

Preserved Fulltext

Efficient and Reliable Probabilistic Interactive Learning with Structured Outputs [article]

Preserved Fulltext

Active Learning of Multi-Class Classification Models from Ordered Class Sets

Preserved Fulltext

Active Learning of Multi-class Classification Models from Ordered Class Sets

Preserved Fulltext

Probabilistic Modelling Of Protein Sub-Cellular Localisation

Preserved Fulltext

Active Learning for Crowdsourcing Using Knowledge Transfer

Preserved Fulltext

Active Learning in Video Tracking [article]

Preserved Fulltext

Other Versions

A framework for summarizing chromatin state annotations within and identifying differential annotations across groups of samples [article]

Preserved Fulltext