Label Propagation for Learning With Label Proportions

Poyiadzi, Rafael; Santos-Rodríguez, Raúl; Twomey, Niall

doi:10.1109/mlsp.2018.8517083

Cited by 5 publications

(6 citation statements)

References 11 publications

(9 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In this section we present our algorithm for performing Active Learning with an LLP-oracle. Our approach is based on the method of Label Propagation (LP) which was developed for semi-supervised learning [21] and which was later adopted to cope with the Label Proportions level of supervision in [11]. We start with a very brief overview of LP, then proceed to present how it is adopted to learn with label proportions and finally how it could be employed within the Active Learning setting.…”

Section: Active Learning With Bag Proportionsmentioning

confidence: 99%

“…Although this relaxes the assumption of infallibility of the oracle, we argue that exploring varying degrees of supervision can lead to an easier and simpler interaction in between the algorithm and the oracle. In this paper, we cast our problem as an instance of the Learning from Label Proportions (LLP) setting [8,9,10,11,12]. Figure 1 illustrates this idea.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Active Learning with Label Proportions

Poyiadzis

Santos-Rodríguez

Twomey

2019

ICASSP 2019 - 2019 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)

Self Cite

View full text Add to dashboard Cite

Active Learning (AL) refers to the setting where the learner has the ability to perform queries to an oracle to acquire the true label of an instance or, sometimes, a set of instances. Even though Active Learning has been studied extensively, the setting is usually restricted to assume that the oracle is trustworthy and will provide the actual label. We argue that, while common, this approach can be made more flexible to account for different forms of supervision. In this paper, we propose a new framework that allows the algorithm to request the label for a bag of samples at a time. Although this label will come in the form of proportions of class labels in the bags and therefore encode less information, we demonstrate that we can still learn effectively.

show abstract

Section: Active Learning With Bag Proportionsmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Active Learning with Label Proportions

Poyiadzis

Santos-Rodríguez

Twomey

2019

ICASSP 2019 - 2019 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)

Self Cite

View full text Add to dashboard Cite

show abstract

“…However, annotating such large datasets promptly became a bottleneck in supervised learning, as it is a time-consuming and labor-intensive task. Additionally, various applied areas such as healthcare or democratic elections struggle with labels, which are often not available (Qi et al 2016). In many scenarios, despite the unavailability of instance-level annotations, approximate group-level labels like class proportions are readily obtainable from other sources, like the census or even common knowledge.…”

Section: Introductionmentioning

confidence: 99%

“…In many scenarios, despite the unavailability of instance-level annotations, approximate group-level labels like class proportions are readily obtainable from other sources, like the census or even common knowledge. In this sense, efficient learning from group-level labels would have an important impact in many real-life applications, such as demographic classification (Ardehaly and Culotta 2017), presidential elections (Sun, Sheldon, and O'Connor 2017;Qi et al 2016), remote sensing (Ding, Li, and Yu 2017), image analysis in medicine (Bortsova et al 2018), activity recognition (Poyiadzi, Santos-Rodriguez, and Twomey 2018), and others.…”

Section: Introductionmentioning

confidence: 99%

Learning from Label Proportions with Prototypical Contrastive Clustering

Rosa

Oliveira

2022

AAAI

View full text Add to dashboard Cite

The use of priors to avoid manual labeling for training machine learning methods has received much attention in the last few years. One of the critical subthemes in this regard is Learning from Label Proportions (LLP), where only the information about class proportions is available for training the models. While various LLP training settings verse in the literature, most approaches focus on bag-level label proportions errors, often leading to suboptimal solutions. This paper proposes a new model that jointly uses prototypical contrastive learning and bag-level cluster proportions to implement efficient LLP classification. Our proposal explicitly relaxes the equipartition constraint commonly used in prototypical contrastive learning methods and incorporates the exact cluster proportions into the optimal transport algorithm used for cluster assignments. At inference time, we compute the clusters' assignment, delivering instance-level classification. We experimented with our method on two widely used image classification benchmarks and report a new state-of-art LLP performance, achieving results close to fully supervised methods.

show abstract

“…Each bag contains unlabeled instances and is annotated with the proportion of instances arising from each class. Various methods for LLP have been developed, including those based on support vector machines and related models [32,44,43,30,9,19,36], Bayesian and graphical models [18,14,40,29,15], deep learning [21,1,12,22,41], clustering [7,39], and random forests [37]. In addition, LLP has found various applications including image and video analysis [8,19], high energy physics [10], vote prediction [40], remote sensing [21,11], medical image analysis [5], activity recognition [29], and reproductive medicine [15].…”

Section: Introductionmentioning

confidence: 99%

Learning from Label Proportions: A Mutual Contamination Framework

Scott,

Zhang

2020

Preprint

View full text Add to dashboard Cite

Learning from label proportions (LLP) is a weakly supervised setting for classification in which unlabeled training instances are grouped into bags, and each bag is annotated with the proportion of each class occurring in that bag. Prior work on LLP has yet to establish a consistent learning procedure, nor does there exist a theoretically justified, general purpose training criterion. In this work we address these two issues by posing LLP in terms of mutual contamination models (MCMs), which have recently been applied successfully to study various other weak supervision settings. In the process, we establish several novel technical results for MCMs, including unbiased losses and generalization error bounds under non-iid sampling plans. We also point out the limitations of a common experimental setting for LLP, and propose a new one based on our MCM framework.

show abstract

Label Propagation for Learning With Label Proportions

Cited by 5 publications

References 11 publications

Active Learning with Label Proportions

Active Learning with Label Proportions

Learning from Label Proportions with Prototypical Contrastive Clustering

Learning from Label Proportions: A Mutual Contamination Framework

Contact Info

Product

Resources

About