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

Here we propose using the successor representation (SR) to acceleratelearning in a constructive knowledge system based on general value functions(GVFs). In real-world settings like robotics for unstructured and dynamicenvironments, it is infeasible to model all meaningful aspects of a system andits environment by hand due to both complexity and size. Instead, robots mustbe capable of learning and adapting to changes in their environment and task,incrementally constructing models from their own experience. GVFs, taken fromthe field of reinforcement learning (RL), are a way of modeling the world aspredictive questions. One approach to such models proposes a massive network ofinterconnected and interdependent GVFs, which are incrementally added overtime. It is reasonable to expect that new, incrementally added predictions canbe learned more swiftly if the learning process leverages knowledge gained frompast experience. The SR provides such a means of separating the dynamics of theworld from the prediction targets and thus capturing regularities that can bereused across multiple GVFs. As a primary contribution of this work, we showthat using SR-based predictions can improve sample efficiency and learningspeed in a continual learning setting where new predictions are incrementallyadded and learned over time. We analyze our approach in a grid-world and thendemonstrate its potential on data from a physical robot arm.

Accelerating Learning in Constructive Predictive Frameworks with the Successor Representation