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

We introduce NIFT, Neural Interaction Field and Template, a descriptive androbust interaction representation of object manipulations to facilitateimitation learning. Given a few object manipulation demos, NIFT guides thegeneration of the interaction imitation for a new object instance by matchingthe Neural Interaction Template (NIT) extracted from the demos in the targetNeural Interaction Field (NIF) defined for the new object. Specifically, theNIF is a neural field that encodes the relationship between each spatial pointand a given object, where the relative position is defined by a sphericaldistance function rather than occupancies or signed distances, which arecommonly adopted by conventional neural fields but less informative. For agiven demo interaction, the corresponding NIT is defined by a set of spatialpoints sampled in the demo NIF with associated neural features. To bettercapture the interaction, the points are sampled on the Interaction BisectorSurface (IBS), which consists of points that are equidistant to the twointeracting objects and has been used extensively for interactionrepresentation. With both point selection and pointwise features defined forbetter interaction encoding, NIT effectively guides the feature matching in theNIFs of the new object instances such that the relative poses are optimized torealize the manipulation while imitating the demo interactions. Experimentsshow that our NIFT solution outperforms state-of-the-art imitation learningmethods for object manipulation and generalizes better to objects from newcategories.

NIFT: Neural Interaction Field and Template for Object Manipulation