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 Tools annual

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 Tools monthly

Global: Zendy Plus monthly

Presents the access of premium content as premium feature

Premium Content

Global: Zendy Plus annual

Global: Zendy Free Plan

Learning transferable knowledge across similar but different settings is afundamental component of generalized intelligence. In this paper, we approachthe transfer learning challenge from a causal theory perspective. Our agent isendowed with two basic yet general theories for transfer learning: (i) a taskshares a common abstract structure that is invariant across domains, and (ii)the behavior of specific features of the environment remain constant acrossdomains. We adopt a Bayesian perspective of causal theory induction and usethese theories to transfer knowledge between environments. Given these generaltheories, the goal is to train an agent by interactively exploring the problemspace to (i) discover, form, and transfer useful abstract and structuralknowledge, and (ii) induce useful knowledge from the instance-level attributesobserved in the environment. A hierarchy of Bayesian structures is used tomodel abstract-level structural causal knowledge, and an instance-levelassociative learning scheme learns which specific objects can be used to inducestate changes through interaction. This model-learning scheme is thenintegrated with a model-based planner to achieve a task in the OpenLockenvironment, a virtual ``escape room'' with a complex hierarchy that requiresagents to reason about an abstract, generalized causal structure. We compareperformances against a set of predominate model-free reinforcement learning(RL)algorithms. RL agents showed poor ability transferring learned knowledge acrossdifferent trials. Whereas the proposed model revealed similar performancetrends as human learners, and more importantly, demonstrated transfer behavioracross trials and learning situations.

Theory-based Causal Transfer: Integrating Instance-level Induction and Abstract-level Structure Learning