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

Central to the success of artificial neural networks is their ability togeneralize. But does neural network generalization primarily rely on seeinghighly similar training examples (memorization)? Or are neural networks capableof human-intelligence styled reasoning, and if so, to what extent? These remainfundamental open questions on artificial neural networks. In this paper, assteps towards answering these questions, we introduce a new benchmark, PointerValue Retrieval (PVR) to study the limits of neural network reasoning. The PVRsuite of tasks is based on reasoning about indirection, a hallmark of humanintelligence, where a first stage (task) contains instructions for solving asecond stage (task). In PVR, this is done by having one part of the task inputact as a pointer, giving instructions on a different input location, whichforms the output. We show this simple rule can be applied to create a diverseset of tasks across different input modalities and configurations. Importantly,this use of indirection enables systematically varying task difficulty throughdistribution shifts and increasing functional complexity. We conduct a detailedempirical study of different PVR tasks, discovering large variations inperformance across dataset sizes, neural network architectures and taskcomplexity. Further, by incorporating distribution shift and increasedfunctional complexity, we develop nuanced tests for reasoning, revealing subtlefailures and surprising successes, suggesting many promising directions ofexploration on this benchmark.

Pointer Value Retrieval: A new benchmark for understanding the limits of neural network generalization