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

Diffusion models have demonstrated highly-expressive generative capabilitiesin vision and NLP. Recent studies in reinforcement learning (RL) have shownthat diffusion models are also powerful in modeling complex policies ortrajectories in offline datasets. However, these works have been limited tosingle-task settings where a generalist agent capable of addressing multi-taskpredicaments is absent. In this paper, we aim to investigate the effectivenessof a single diffusion model in modeling large-scale multi-task offline data,which can be challenging due to diverse and multimodal data distribution.Specifically, we propose Multi-Task Diffusion Model (\textsc{MTDiff}), adiffusion-based method that incorporates Transformer backbones and promptlearning for generative planning and data synthesis in multi-task offlinesettings. \textsc{MTDiff} leverages vast amounts of knowledge available inmulti-task data and performs implicit knowledge sharing among tasks. Forgenerative planning, we find \textsc{MTDiff} outperforms state-of-the-artalgorithms across 50 tasks on Meta-World and 8 maps on Maze2D. For datasynthesis, \textsc{MTDiff} generates high-quality data for testing tasks givena single demonstration as a prompt, which enhances the low-quality datasets foreven unseen tasks.

Diffusion Model is an Effective Planner and Data Synthesizer for Multi-Task Reinforcement Learning