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

Learning user representations is a vital technique toward effective usermodeling and personalized recommender systems. Existing approaches often derivean individual set of model parameters for each task by training on separatedata. However, the representation of the same user potentially has somecommonalities, such as preference and personality, even in different tasks. Assuch, these separately trained representations could be suboptimal inperformance as well as inefficient in terms of parameter sharing.  In this paper, we delve on research to continually learn user representationstask by task, whereby new tasks are learned while using partial parameters fromold ones. A new problem arises since when new tasks are trained, previouslylearned parameters are very likely to be modified, and as a result, anartificial neural network (ANN)-based model may lose its capacity to serve forwell-trained previous tasks forever, this issue is termed catastrophicforgetting. To address this issue, we present \emph{Conure} the first\underline{con}tinual, or lifelong, \underline{u}ser \underline{re}presentationlearner -- i.e., learning new tasks over time without forgetting old ones.Specifically, we propose iteratively removing less important weights of oldtasks in a deep user representation model, motivated by the fact that neuralnetwork models are usually over-parameterized. In this way, we could learn manytasks with a single model by reusing the important weights, and modifying theless important weights to adapt to new tasks. We conduct extensive experimentson two real-world datasets with nine tasks and show that \emph{Conure} largelyexceeds the standard model that does not purposely preserve such old"knowledge", and performs competitively or sometimes better than models whichare trained either individually for each task or simultaneously by merging alltask data.

One Person, One Model, One World: Learning Continual User Representation without Forgetting