Plastic and Stable Gated Classifiers for Continual Learning

Kuo, Nicholas I-Hsien; Harandi, Mehrtash; Fourrier, Nicolas; Walder, Christian; Ferraro, Gabriela; Suominen, Hanna

doi:10.1109/cvprw53098.2021.00394

Cited by 4 publications

(3 citation statements)

References 11 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Another limitation is the imbalanced nature of the data set, which is common in many medical research data sets. Future studies may utilise weighted sampling of input data, bidirectional encoder representations from transformers and continual learning to improve performance 16,17 . Furthermore, all analyses in this study were conducted with English‐language text and all EHR data were from within South Australia.…”

Section: Discussionmentioning

confidence: 99%

Machine learning models automate classification of penicillin adverse drug reaction labels

Inglis,

Bacchi,

Troelnikov

et al. 2023

Internal Medicine Journal

View full text Add to dashboard Cite

There is a growing interest in the appropriate evaluation of penicillin adverse drug reaction (ADR) labels. We have developed machine learning models for classifying penicillin ADR labels using free‐text reaction descriptions, and here report external and practical validation. The models performed comparably with expert criteria for the categorisation of allergy or intolerance and identification of high‐risk allergies. These models have practical applications in detecting individuals suitable for penicillin ADR evaluation. Implementation studies are required.

show abstract

Section: Discussionmentioning

confidence: 99%

Machine learning models automate classification of penicillin adverse drug reaction labels

Inglis,

Bacchi,

Troelnikov

et al. 2023

Internal Medicine Journal

View full text Add to dashboard Cite

show abstract

“…The concluding phase involves refining assumptions and model performance, incorporating multiple tests to identify optimal hyperparameters [76]. Here, students peek into the "black box" nature of ML and gain an intuition for effective module combinations [77][78][79]. This step becomes critical for causal inference tasks that necessitate rigorous input data validation [80].…”

Section: Worked Examplementioning

confidence: 99%

Enriching Data Science and Health Care Education: Application and Impact of Synthetic Data Sets Through the Health Gym Project

Kuo,

Perez-Concha,

Hanly

et al. 2024

JMIR Med Educ

Self Cite

View full text Add to dashboard Cite

Large-scale medical data sets are vital for hands-on education in health data science but are often inaccessible due to privacy concerns. Addressing this gap, we developed the Health Gym project, a free and open-source platform designed to generate synthetic health data sets applicable to various areas of data science education, including machine learning, data visualization, and traditional statistical models. Initially, we generated 3 synthetic data sets for sepsis, acute hypotension, and antiretroviral therapy for HIV infection. This paper discusses the educational applications of Health Gym’s synthetic data sets. We illustrate this through their use in postgraduate health data science courses delivered by the University of New South Wales, Australia, and a Datathon event, involving academics, students, clinicians, and local health district professionals. We also include adaptable worked examples using our synthetic data sets, designed to enrich hands-on tutorial and workshop experiences. Although we highlight the potential of these data sets in advancing data science education and health care artificial intelligence, we also emphasize the need for continued research into the inherent limitations of synthetic data.

show abstract

“…Thus, new categories cannot be handled by a model with static architecture. In some CL approaches [49], [50], the network is usually constructed by two modules: a feature extractor and a classifier. Currently, to handle new classes, many works assign a new classifier to the network when a new class emerges [51], [46].…”

Section: B Prototype Library For Dynamic Label Spacementioning

confidence: 99%

Continual Local Training For Better Initialization Of Federated Models

Yao

Sun

2020

2020 IEEE International Conference on Image Processing (ICIP)

View full text Add to dashboard Cite

Federated learning (FL) refers to the learning paradigm that trains machine learning models directly in the decentralized systems consisting of smart edge devices without transmitting the raw data, which avoids the heavy communication costs and privacy concerns. Given the typical heterogeneous data distributions in such situations, the popular FL algorithm Federated Averaging (FedAvg) suffers from weight divergence and thus cannot achieve a competitive performance for the global model (denoted as the initial performance in FL) compared to centralized methods. In this paper, we propose the local continual training strategy to address this problem. Importance weights are evaluated on a small proxy dataset on the central server and then used to constrain the local training. With this additional term, we alleviate the weight divergence and continually integrate the knowledge on different local clients into the global model, which ensures a better generalization ability. Experiments on various FL settings demonstrate that our method significantly improves the initial performance of federated models with few extra communication costs.

show abstract

Plastic and Stable Gated Classifiers for Continual Learning

Cited by 4 publications

References 11 publications

Machine learning models automate classification of penicillin adverse drug reaction labels

Machine learning models automate classification of penicillin adverse drug reaction labels

Enriching Data Science and Health Care Education: Application and Impact of Synthetic Data Sets Through the Health Gym Project

Continual Local Training For Better Initialization Of Federated Models

Contact Info

Product

Resources

About