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

Recent work has shown how predictive modeling can endow agents with richknowledge of their surroundings, improving their ability to act in complexenvironments. We propose question-answering as a general paradigm to decode andunderstand the representations that such agents develop, applying our method totwo recent approaches to predictive modeling -action-conditional CPC (Guo etal., 2018) and SimCore (Gregor et al., 2019). After training agents with thesepredictive objectives in a visually-rich, 3D environment with an assortment ofobjects, colors, shapes, and spatial configurations, we probe their internalstate representations with synthetic (English) questions, withoutbackpropagating gradients from the question-answering decoder into the agent.The performance of different agents when probed this way reveals that theylearn to encode factual, and seemingly compositional, information aboutobjects, properties and spatial relations from their physical environment. Ourapproach is intuitive, i.e. humans can easily interpret responses of the modelas opposed to inspecting continuous vectors, and model-agnostic, i.e.applicable to any modeling approach. By revealing the implicit knowledge ofobjects, quantities, properties and relations acquired by agents as they learn,question-conditional agent probing can stimulate the design and development ofstronger predictive learning objectives.

Probing Emergent Semantics in Predictive Agents via Question Answering