REVIEW 11 cited by
DigiRL: Training In-The-Wild Device-Control Agents with Autonomous Reinforcement Learning
Not yet reviewed by Pith; the record is open.
This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.
SPECIMEN: schema-true, not a live event
T0 review · schema-true
One-sentence machine reading of the paper's core claim.
pith:XXXXXXXX · record.json · timestamp
read the original abstract
Training corpuses for vision language models (VLMs) typically lack sufficient amounts of decision-centric data. This renders off-the-shelf VLMs sub-optimal for decision-making tasks such as in-the-wild device control through graphical user interfaces (GUIs). While training with static demonstrations has shown some promise, we show that such methods fall short for controlling real GUIs due to their failure to deal with real-world stochasticity and non-stationarity not captured in static observational data. This paper introduces a novel autonomous RL approach, called DigiRL, for training in-the-wild device control agents through fine-tuning a pre-trained VLM in two stages: offline RL to initialize the model, followed by offline-to-online RL. To do this, we build a scalable and parallelizable Android learning environment equipped with a VLM-based evaluator and develop a simple yet effective RL approach for learning in this domain. Our approach runs advantage-weighted RL with advantage estimators enhanced to account for stochasticity along with an automatic curriculum for deriving maximal learning signal. We demonstrate the effectiveness of DigiRL using the Android-in-the-Wild (AitW) dataset, where our 1.3B VLM trained with RL achieves a 49.5% absolute improvement -- from 17.7 to 67.2% success rate -- over supervised fine-tuning with static human demonstration data. These results significantly surpass not only the prior best agents, including AppAgent with GPT-4V (8.3% success rate) and the 17B CogAgent trained with AitW data (38.5%), but also the prior best autonomous RL approach based on filtered behavior cloning (57.8%), thereby establishing a new state-of-the-art for digital agents for in-the-wild device control.
Forward citations
Cited by 11 Pith papers
-
OSWorld 2.0: Benchmarking Computer Use Agents on Long-Horizon Real-World Tasks
OSWorld 2.0 is a benchmark of 108 realistic long-horizon computer-use tasks where current agents achieve only 20.6% binary completion, struggling with state inference and constraint tracking.
-
OSReward: Instituting Standardized Evaluation for Cross-Platform Computer-Use Reward Models
VLM judges of CUA trajectories are systematically lenient; OSReward measures this with human gold, and OS-Shepherd open models close most of the reliability gap cheaply.
-
GUI-Libra: Training Native GUI Agents to Reason and Act with Action-aware Supervision and Partially Verifiable RL
Action-weighted SFT plus KL-regularized GRPO on a curated 81K reasoning dataset lifts open-source GUI agents by 11–22 points on online task-completion benchmarks.
-
Morae: Proactively Pausing UI Agents for User Choices
Morae, a UI agent that proactively pauses at ambiguous decision points, helps blind and low-vision users complete more tasks and express preferences better than fully autonomous agents.
-
Thinking vs. Doing: Agents that Reason by Scaling Test-Time Interaction
Scaling the number of interaction steps, trained via a curriculum over rollout horizon, improves web-agent task success and outperforms scaling per-step reasoning under fixed token budgets.
-
Self-Challenging Language Model Agents
A language model agent can generate its own verifiable training tasks and improve its tool-use success rate by about 2x without human-annotated data.
-
RoboGPT-R1: Enhancing Robot Task Planning with Reinforcement Learning
A 3B VLM trained with SFT plus GRPO and an LCS-based reward reaches 55.3% on EmbodiedBench's EB-ALFRED, beating GPT-4o-mini and the 7B REBP planner.
-
Atomic-to-Compositional Generalization for Mobile Agents with A New Benchmark and Scheduling System
AI phone agents succeed on single-step commands but fail on multi-step compositions; a new benchmark measures this gap and a scheduler that decomposes tasks recovers most of it.
-
GUI-Reflection: Empowering Multimodal GUI Models with Self-Reflection Behavior
GUI-Reflection trains an 8B multimodal GUI agent to recognize mistakes, undo incorrect actions, and retry, improving AndroidWorld success rate from 14.58% (filtered BC baseline) to 34.72% with reflection data and onli...
-
SWE-Dev: Building Software Engineering Agents with Training and Inference Scaling
SWE-Dev trains open-weight software engineering agents on LLM-synthesized test cases and finds that longer agent interaction budgets raise SWE-bench-Verified resolve rates to 23.4% (7B) and 36.6% (32B).
-
Truly Self-Improving Agents Require Intrinsic Metacognitive Learning
The paper proposes that self-improving agents must learn to manage their own learning processes, framing this as intrinsic metacognitive learning, and argues it is necessary for sustained and generalized improvement.
Discussion (0). Sign in to comment.