Active Learning with Crowdsourcing for the Cold Start of Imbalanced Classifiers

Brangbour, Etienne; Bruneau, Pierrick; Tamisier, Thomas; Marchand-Maillet, Stéphane

doi:10.1007/978-3-030-60816-3_22

Cited by 5 publications

(6 citation statements)

References 16 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…However, with highly imbalanced and complex data sets (i.e. overlapping classes, multiple subconcepts), the likelihood that only labels from the majority class are returned for a sample of any cluster is high, which causes the proposed impurity criterion [3] to discard clusters featuring more than average minority class content too early. Alternatively, using a semi-supervised model will allow to gracefully integrate label feedback.…”

Section: Proposed Methodologymentioning

confidence: 99%

“…Observing that cold start situations remain mostly unaddressed by the active learning literature, and are critical in the context of imbalanced problems, in [3] the general idea is to use a clustering structure to guide batch sampling. The underlying hypothesis is that the label information acquired in high quality clusters would be more effectively propagated to unlabeled elements using a semi-supervised algorithm.…”

Section: Related Workmentioning

confidence: 99%

“…Let us also assume a clustering algorithm has been applied to the data set X , yielding K clusters, and a set of N cluster labels Y = {y i } i∈ 1,N with y i ∈ 1, K . Few unsupervised quality criteria can be decomposed per-cluster in a straightforward manner [3], thus enabling the computation of individual cluster quality scores. The conductance [1] offers this possibility, while having been recognized as showing good agreement to supervised criteria in the context of graph clustering [11], and no explicit requirement of spherical clusters.…”

Section: Cluster Quality Criterionmentioning

confidence: 99%

“…The proposed experimental use case focuses as detecting content posted by people who have witnessed hurricane Harvey, that hit Texas and Louisiana in August 2017 3 . The data we collected are all Tweets with location fields overlapping an area of approximately 40.000 km 2 around Houston, posted in between August the nineteenth and September the twenty-first of 2017.…”

Section: Use Case Descriptionmentioning

confidence: 99%

See 3 more Smart Citations

Cold Start Active Learning Strategies in the Context of Imbalanced Classification

Brangbour¹,

Bruneau²,

Tamisier³

et al. 2022

Preprint

Self Cite

View full text Add to dashboard Cite

We present novel active learning strategies dedicated to providing a solution to the cold start stage, i.e. initializing the classification of a large set of data with no attached labels. Moreover, proposed strategies are designed to handle an imbalanced context in which random selection is highly inefficient. Specifically, our active learning iterations address label scarcity and imbalance using element scores, combining information extracted from a clustering structure to a label propagation model. The strategy is illustrated by a case study on annotating Twitter content w.r.t. testimonies of a real flood event. We show that our method effectively copes with class imbalance, by boosting the recall of samples from the minority class.

show abstract

Section: Proposed Methodologymentioning

confidence: 99%

Section: Related Workmentioning

confidence: 99%

Section: Cluster Quality Criterionmentioning

confidence: 99%

Section: Use Case Descriptionmentioning

confidence: 99%

See 2 more Smart Citations

Cold Start Active Learning Strategies in the Context of Imbalanced Classification

Brangbour¹,

Bruneau²,

Tamisier³

et al. 2022

Preprint

Self Cite

View full text Add to dashboard Cite

show abstract

“…• New users: new users may enter the system without any interaction with the systems and without any personalized recommendations for the guests. In recent years, methods like transfer learning [19], active learning [20], and zero-shot learning [21] are often used to solve cold-start problems.…”

Section: Cold-start Problemmentioning

confidence: 99%

Accuracy improvements for cold-start recommendation problem using indirect relations in social networks

Tey

Lin

et al. 2021

J Big Data

View full text Add to dashboard Cite

Recent advances in Internet applications have facilitated information spreading and, thanks to a wide variety of mobile devices and the burgeoning 5G networks, users easily and quickly gain access to information. Great amounts of digital information moreover have contributed to the emergence of recommender systems that help to filter information. When the rise of mobile networks has pushed forward the growth of social media networks and users get used to posting whatever they do and wherever they visit on the Web, such quick social media updates already make it difficult for users to find historical data. For this reason, this paper presents a social network-based recommender system. Our purpose is to build a user-centered recommender system to exclude the products that users are disinterested in according to user preferences and their friends' shopping experiences so as to make recommendations effective. Since there might be no corresponding reference value for new products or services, we use indirect relations between friends and “friends’ friends” as well as sentinel friends to improve the recommendation accuracy. The simulation result has proven that our proposed mechanism is efficient in enhancing recommendation accuracy.

show abstract

Deep Active Learning Framework for Crowdsourcing-Enhanced Image Classification and Segmentation

Gao

Chen

2022

Lecture Notes in Computer Science

View full text Add to dashboard Cite

Active Learning with Crowdsourcing for the Cold Start of Imbalanced Classifiers

Cited by 5 publications

References 16 publications

Cold Start Active Learning Strategies in the Context of Imbalanced Classification

Cold Start Active Learning Strategies in the Context of Imbalanced Classification

Accuracy improvements for cold-start recommendation problem using indirect relations in social networks

Deep Active Learning Framework for Crowdsourcing-Enhanced Image Classification and Segmentation

Contact Info

Product

Resources

About