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

With the advent of large language models and large-scale robotic datasets,there has been tremendous progress in high-level decision-making for objectmanipulation. These generic models are able to interpret complex tasks usinglanguage commands, but they often have difficulties generalizing toout-of-distribution objects due to the inability of low-level actionprimitives. In contrast, existing task-specific models excel in low-levelmanipulation of unknown objects, but only work for a single type of action. Tobridge this gap, we present M2T2, a single model that supplies different typesof low-level actions that work robustly on arbitrary objects in clutteredscenes. M2T2 is a transformer model which reasons about contact points andpredicts valid gripper poses for different action modes given a raw point cloudof the scene. Trained on a large-scale synthetic dataset with 128K scenes, M2T2achieves zero-shot sim2real transfer on the real robot, outperforming thebaseline system with state-of-the-art task-specific models by about 19% inoverall performance and 37.5% in challenging scenes where the object needs tobe re-oriented for collision-free placement. M2T2 also achievesstate-of-the-art results on a subset of language conditioned tasks in RLBench.Videos of robot experiments on unseen objects in both real world and simulationare available on our project website https://m2-t2.github.io.

M2T2: Multi-Task Masked Transformer for Object-centric Pick and Place