Self-Improvement Needs Attribution: Dependency-Aware Causal Memory for Long-Horizon Mobile GUI Agents

**Abstract of paper.** Mobile GUI agents are moving toward long-horizon tasks that span dozens or even hundreds of steps across multiple applications, where the working context explodes as execution progresses. Information the agent reads on earlier screens is gradually drowned in the expanding context and becomes hard to identify when needed by subsequent steps. Existing systems fold raw histories into task state or structured memory and recall items by recency or similarity. However, merely summarizing or retrieving past context fails to catch silent step errors, leaving active context polluted with unverified steps and further impairing downstream experience distillation. We introduce **$\mathrm{CausGUI}$**, a self-improving framework that empowers mobile GUI agents with causal context and attribution-driven evolution, alongside MobileWorld-XL, an evaluation benchmark derived from the MobileWorld ecosystem to assess the generalization of extracted experience across unseen tasks. At its core, **$\mathrm{CausGUI}$** maintains a causal ledger to log the causal dependencies across actions and observed facts, enabling online step auditing, causal-dependency context assembly, and cause-attributed experience curation. Under identical conditions on three public benchmarks with a Qwen3.5-397B-A17B backbone, **$\mathrm{CausGUI}$** consistently outperforms competitive frameworks, achieving $52.1\%$ on MobileWorld ($+16.2$ over backbone), $40.6\%$ on MemGUI-Bench ($+10.9$), and $51.9\%$ on AndroidWorld hard slice ($+14.9$). Furthermore, task success rises across five rounds of self-evolution from $52.1\%$ to $59.0\%$, with experience transferring to the unseen tasks of MobileWorld-XL. The code is available at: $\text{url}$.