SOURCE-LINKED INTELLIGENCE

Iron: Intent-Aligned and Retrospective Dual Learning Framework for Enhancing Generalist Virtual Agents

arXiv · AI, language, vision and robotics · article · Aug 28, 2026 · UTC

Achieving virtual agents capable of automating tasks across diverse digital environments remains a pivotal challenge in Embodied AI. While Multimodal Large Language Models (MLLMs) offer enhanced visual perception and reasoning, their agentic deployment faces three challenges: costly data annotation, imprecise action-intent alignment, and inefficient exploration from discarded failed trajectories. To address these, we introduce Iron, an intent-aligned, self-improved, and annotation-efficient framework for training GUI agents. Iron employs a novel dual learning strategy that utilizes a stepwise

Read original source ↗ Open in workspace

recordType

paper

region

Global