English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

In this paper we introduce the ice-start problem, i.e., the challenge ofdeploying machine learning models when only little or no training data isinitially available, and acquiring each feature element of data is associatedwith costs. This setting is representative for the real-world machine learningapplications. For instance, in the health-care domain, when training an AIsystem for predicting patient metrics from lab tests, obtaining every singlemeasurement comes with a high cost. Active learning, where only the label isassociated with a cost does not apply to such problem, because performing allpossible lab tests to acquire a new training datum would be costly, as well asunnecessary due to redundancy. We propose Icebreaker, a principled framework toapproach the ice-start problem. Icebreaker uses a full Bayesian Deep LatentGaussian Model (BELGAM) with a novel inference method. Our proposed methodcombines recent advances in amortized inference and stochastic gradient MCMC toenable fast and accurate posterior inference. By utilizing BELGAM's ability tofully quantify model uncertainty, we also propose two information acquisitionfunctions for imputation and active prediction problems. We demonstrate thatBELGAM performs significantly better than the previous VAE (Variationalautoencoder) based models, when the data set size is small, using both machinelearning benchmarks and real-world recommender systems and health-careapplications. Moreover, based on BELGAM, Icebreaker further improves theperformance and demonstrate the ability to use minimum amount of the trainingdata to obtain the highest test time performance.

Icebreaker: Element-wise Active Information Acquisition with Bayesian Deep Latent Gaussian Model