Learning from Ambiguously Labeled Examples.

Inducing a classification function from a set of examples in the form of labeled instances is a standard problem in supervised machine learning. ... the information contained in ambiguously labeled examples. ... Learning from Ambiguous Examples Ambiguous data may comprise important information. ...

doi:10.1007/11552253_16 fatcat:k6h2ljx2svg2ne67qfdwr7gxkm

We formulate the learning problem in this setting as partially-supervised multiclass classification where each instance is labeled ambiguously with more than one label. ... Motivated by the analysis, we propose a general convex learning formulation based on minimization of a surrogate loss appropriate for the ambiguous label setting. ... The goal is to correctly label faces from examples that have multiple potential labels (transductive case), as well as learn a model from ambiguous data that generalizes to other unlabeled examples (inductive ...

doi:10.1109/cvprw.2009.5206667 fatcat:4uyko5gsfjbitkin3ypniy3wgi

We formulate the learning problem in this setting as partially-supervised multiclass classification where each instance is labeled ambiguously with more than one label. ... Motivated by the analysis, we propose a general convex learning formulation based on minimization of a surrogate loss appropriate for the ambiguous label setting. ... The goal is to correctly label faces from examples that have multiple potential labels (transductive case), as well as learn a model from ambiguous data that generalizes to other unlabeled examples (inductive ...

doi:10.1109/cvpr.2009.5206667 dblp:conf/cvpr/CourSJT09 fatcat:csjt64zecvamli3kisffxjvngq

The study of semi-supervised category learning has generally focused on how additional unlabeled information with given labeled information might benefit category learning. ... Using an unconstrained free-sorting categorization experiment, we show that labels are useful to participants only when the category structure is ambiguous and that people's responses are driven by the ... The results from Vandist et al. (2009) suggest that labeled examples can also help in learning complex Information-Integration categories -in that case, the categories are well-separated and not ambiguous ...

doi:10.3758/s13423-015-0857-9 pmid:26106058 fatcat:bnrfiyp4xrde5idggzbtzortba

Szczepanski

Natural Language Inference (NLI) datasets contain examples with highly ambiguous labels. ... In this paper, we explore the option of training directly on the estimated label distribution of the annotators in the NLI task, using a learning loss based on this ambiguity distribution instead of the ... This ambiguity stems from the lack of proper context or differences in background knowledge between annotators, and leads to a large number of examples where the correctness of labels can be debated. ...

arXiv:2106.03020v1 fatcat:zvs5ple4rzdqhmccizecgba6pa

We therefore propose an algorithm to construct the generalized queries and improve the learning model with such ambiguous answers in active learning. ... As each generalized query is equivalent to a set of specific ones, the answers from the oracle can usually provide more information thus speeding up the learning effectively. ... In most traditional active learning studies, the learner usually regards the specific examples directly as queries, and requests the corresponding labels from the oracle. ...

doi:10.1007/978-3-642-15880-3_30 fatcat:hcdihnv2rfdd3mdtsat5bb3qru

Exploiting this property of the data, we propose a convex learning formulation based on minimization of a loss function appropriate for the partial label setting. ... labeled. ... We assume that ties are broken arbitrarily, for example, by selecting the label with smallest index a. ...

dblp:journals/jmlr/CourST11 fatcat:i75yw4lwafdhzpi7sxbfcrhitq

Szczepanski

A separate data sets using five ambiguous words were prepared for the development of this Amharic WSD prototype. ... The final classification task was done on fully labelled training set using Adaboost, bagging, and AD tree classification algorithms on WEKA package. ... Generally, semi-supervised learning is to automatically label the unlabelled examples using a small number of manually labelled examples as seeds. ...

doi:10.4314/star.v3i3.25 fatcat:ecwq3pgozbfejn2wh7ebnfpiaq

Szczepanski

Learning visual classifiers for object recognition from weakly labeled data requires determining correspondence between image regions and semantic object classes. ... We further constrain the correspondence problem by exploiting additional language constructs to improve the learning process from weakly labeled data. ... These classifiers are learned on positive and negative examples generated from captions. ...

doi:10.1007/978-3-540-88682-2_3 fatcat:p74arxhaynfjfgtvg4rodgiy3u

To learn from MIML examples we have taken a survey on MIML Boost, MIMLSVM, D-MIMLSVM, InsDif and SubCod algorithms. MIML Boost and MIML SVM are based on a simple degeneration strategy. ... InsDif and SubCod algorithms works by transforming single-instances into the MIML representation for learning, while SubCod works by transforming single-label examples into the MIML representation for ... In fact, the multi-learning frameworks are resulted from the ambiguities in representing real-world objects. ...

doi:10.9756/bijaip.1001 fatcat:hzdw53v42fhtvckwsghnv24ura

Particularly, we first unify those problems into a general ambiguous problem in which we simultaneously learn a robust classifier as well as find the optimal training labels from a finite label candidate ... We propose a multi-view learning approach called co-labeling which is applicable for several machine learning problems where the labels of training samples are uncertain, including semi-supervised learning ... AMBIGUOUS LEARNING FROM LABEL CANDIDATES PERSPECTIVE In ambiguous learning, the training samples have uncertain labels that satisfy some constraints (for example, the bag constraints in MIL and the balance ...

doi:10.1109/icdm.2012.78 dblp:conf/icdm/LiDTX12 fatcat:dkali5tkuzdztmqhb2k2ewahty

In Stage 2 (RESOLVE), the requester selects one or more of these ambiguous examples to label (resolving ambiguity). ... Stage 1 (FIND) asks the crowd to find examples whose correct label seems ambiguous given task instructions. ... Despite collecting labels from nine different workers, the majority is still wrong, with majority vote accuracy on ambiguous examples falling below 50%. ...

doi:10.3389/frai.2022.828187 pmid:35664506 pmcid:PMC9159300 fatcat:2jaq6gwk4zfsve3lqqlzhcz7ta

DOAJ

Finally, data maps uncover a region with instances that the model finds "hard to learn"; these often correspond to labeling errors. ... Our results indicate that a shift in focus from quantity to quality of data could lead to robust models and improved out-of-distribution generalization. ... We thank the anonymous reviewers, and our colleagues from AI2 and UWNLP, especially Ana Marasović, and Suchin Gururangan, for their helpful feedback. ...

arXiv:2009.10795v2 fatcat:xmdge47t6nfwjo3aqjbeqjoet4

Multiple Versions

In Stage 2 (RESOLVE), the requester selects one or more of these ambiguous examples to label (resolving ambiguity). ... Stage 1 (FIND) asks the crowd to find examples whose correct label seems ambiguous given task instructions. ... Learning from crowds in the presence of schools of thought. ...

arXiv:2112.02255v1 fatcat:xpvsuclbi5dfrmfm2ejx54qdii

We propose a novel dictionary-based learning method for ambiguously labeled multiclass classification, where each training sample has multiple labels and only one of them is the correct label. ... Extensive evaluations on existing datasets demonstrate that the proposed method performs significantly better than state-of-the-art ambiguously labeled learning approaches. ... Acknowledgment This work was partially supported by a Co−operative Agreement from the National Institute of Standards and Technology (NIST) under the Grant 70NANB11H023. ...

doi:10.1109/cvpr.2013.52 dblp:conf/cvpr/ChenPPCP13 fatcat:f4ctsdmapbatpecpcpv6wxk3mu

Learning from Ambiguously Labeled Examples [chapter]

Preserved Fulltext

Learning from ambiguously labeled images

Preserved Fulltext

Learning from ambiguously labeled images

Preserved Fulltext

The helpfulness of category labels in semi-supervised learning depends on category structure

Preserved Fulltext

Embracing Ambiguity: Shifting the Training Target of NLI Models [article]

Preserved Fulltext

Asking Generalized Queries to Ambiguous Oracle [chapter]

Preserved Fulltext

Learning from Partial Labels

Preserved Fulltext

A Word Sense Disambiguation Model for Amharic Words using Semi-Supervised Learning Paradigm

Preserved Fulltext

Beyond Nouns: Exploiting Prepositions and Comparative Adjectives for Learning Visual Classifiers [chapter]

Preserved Fulltext

Automatic Image Annotation and Retrieval using Multi-Instance Multi-Label Learning

Preserved Fulltext

Co-labeling: A New Multi-view Learning Approach for Ambiguous Problems

Preserved Fulltext

In Search of Ambiguity: A Three-Stage Workflow Design to Clarify Annotation Guidelines for Crowd Workers

Preserved Fulltext

Dataset Cartography: Mapping and Diagnosing Datasets with Training Dynamics [article]

Preserved Fulltext

Other Versions

In Search of Ambiguity: A Three-Stage Workflow Design to Clarify Annotation Guidelines for Crowd Workers [article]

Preserved Fulltext

Dictionary Learning from Ambiguously Labeled Data

Preserved Fulltext