Online Dense Reinforcement Learning with Planning-Grounding Agent for UI Navigation on Live IOS Apps
READ THE ORIGINAL SOURCEOpen original sourcehuggingface.co · Opens in a new tabThe paper proposes an online reinforcement learning framework that trains UI navigation agents directly on live iOS applications. The agent is split into a Planning agent and a UI Grounding model, trained with dense step-level rewards from an LLM-as-a-Judge plus a terminal reward, optimized with PPO. It introduces GiGPPO, a neighbor-conditioned value estimation method, and infrastructure running hundreds of concurrent virtual iOS devices. The authors report gains over SFT and terminal-only RL baselines on UINavBench, iOSWorld, and AndroidWorld.
One primary source; other links are context, not independent corroboration.
- Primary sourceHugging Face Daily Papers
- Discovery linkHugging Face
- Linked in sourceView arXiv page
- Linked in sourceView PDF
The paper page is linked. Peer review and findings have not been independently checked by AINewsIQ.
Open full article view →- Original paper+45
- Fetched source page+30
- No editor link review+0
- Original non-discussion link+5
This measures the stored source trail and editorial steps. It does not certify accuracy, peer review, or independent corroboration. Popularity adds no points.
View full scoring →Read the scoring rubric →What the source says
The paper proposes an online reinforcement learning framework that trains UI navigation agents directly on live iOS applications. The agent is split into a Planning agent and a UI Grounding model, trained with dense step-level rewards from an LLM-as-a-Judge plus a terminal reward, optimized with PPO. It introduces GiGPPO, a neighbor-conditioned value estimation method, and infrastructure running hundreds of concurrent virtual iOS devices. The authors report gains over SFT and terminal-only RL baselines on UINavBench, iOSWorld, and AndroidWorld.
A READER'S EVIDENCE CHECKHow to read this signalSix questions about claims, sources, risks, and open checks
This is a signal-level guide until claim-level evidence is available. “Not recorded” means the evidence has not been assessed here; it does not mean a result is false. How we check AI news →
- 01Who is making each claim?
The claimant and whether each claim was measured or asserted are not recorded in this public brief.
- 02What was actually tested?
Not recorded per claim. Check the source for sample size, tasks, model and benchmark versions, and reproducibility.
- 03Where are the sources?
Original link: huggingface.co. Brief basis: fetched source page. Primary page date and quote locations are not established in this panel.
- 04What could mislead?
Source claims not independently checked; results are self-reported by the authors on their benchmarks.
- 05What is at risk if wrong?
Not assessed per claim. Consider safety, privacy, cost, and adoption decisions before relying on the result.
- 06What would change the view?
No claim-specific check is recorded. An inspectable primary method or independent replication could change the assessment.
Key findings from the stored brief
- Online RL trains UI navigation agents on live iOS apps using dense step-level and terminal rewards from an LLM-as-a-Judge.
- GiGPPO improves critic stability by using reward signals from visually similar states across concurrent trajectories.
- The authors report 8B RL agents achieve nearly sevenfold improvement over the SFT baseline in out-of-distribution Android generalization.
- Reported results show the 8B RL agents outperform Gemini-3.5-Flash and tie Gemini-3.1-Pro on real iOS applications.
Source claims not independently checked; results are self-reported by the authors on their benchmarks.
This is a source-based abstract, not independent validation of the source's claims.
Evidence & sources
These links show where the signal came from and what the source page linked to. A linked page is context, not independent corroboration.
Related archive coverage
These relationships record a citation or topic overlap, not confirmation of the finding.
The paper page is linked. Peer review and findings have not been independently checked by AINewsIQ.
Timeline & updates
- Signal dated by source or discovery feed
This timestamp places the signal in the radar. It may be the discovery post date rather than the primary page's first publication date.
- Shared brief stored
AINewsIQ prepared one abstract from fetched source page.
A publication revision history is not stored for this signal. New engagement snapshots do not mean the source itself changed.
Why this may matter
Training UI agents online on live apps aims to address brittle, memorized behavior of smaller models and their inability to recover from errors, while keeping inference costs low enough for everyday use.
Limits and open questions
Source claims not independently checked; results are self-reported by the authors on their benchmarks.
Organization-specific financial, security, and operational effects require the organization's own context and review.
Discussion
Comments are reviewed before publication. Discuss the evidence, not the person.
No published comments yet.
Reader discussion opens with accounts. You can read the original source and published comments now.