已审阅并标注「强化学习」的研究与资产。
条目依据已审阅标签归类，同一主题不表示资源之间存在依赖关系。日期保留来源精度，注释保留原语言。
DART separates GUI reinforcement learning into asynchronous environment, rollout, data-management, and training services. It curates experience by task difficulty, successful trajectories, and step uncertainty while correcting rollout-policy mismatch, with evaluation on OSWorld.
RISK combines browser-interaction training data, a benchmark, and reinforcement fine-tuning for e-commerce risk-management workflows. Its trajectories and rewards distinguish individual actions from multi-step stateful tasks involving deeply embedded web content.
GUI-Cursor treats GUI grounding as a sequence of cursor movements guided by visible feedback and movement history. Online reinforcement learning with trajectory rewards trains the model to refine the target location across multiple steps.
Orcust combines principle-constrained reward modeling with GUI trajectory collection in instrumented virtual machines. Environment-checkable and model-derived principles guide stepwise feedback for grounding, basic operations, and complete task execution.
BTL-UI separates GUI interaction into screen-region attention, planning and executable-command generation. It adds an automated annotation pipeline and rule-based process/outcome rewards for training, with evaluation on both static understanding and dynamic interaction.
GUI-ARP uses visual attention to crop relevant screenshot regions and decides whether a target needs single-stage or multi-stage analysis. Supervised fine-tuning followed by GRPO trains this adaptive procedure for high-resolution GUI grounding.
TGPO merges semantically equivalent web states into tree-structured trajectories for offline preference optimization. A process reward model scores subgoal progress, redundant behavior, and action validity, while dynamic weighting emphasizes consequential decisions.
UI-S1 simulates multi-turn online learning over offline trajectories, using a patch module to recover divergences while retaining the model’s own outputs in context. It combines discounted future returns with step- and episode-level advantages and introduces a semi-online performance proxy.
MobileRL adapts GRPO to mobile GUI training with difficulty-dependent replay of successes and filtering of failure curricula. A shortest-path reward adjustment accounts for task length, with evaluations on AndroidWorld and AndroidLab.
Introduces VeriOS-Agent, which learns when to execute GUI actions autonomously and when to query a user under untrustworthy conditions. Its staged training combines supervised learning and group-relative policy optimization.
UI-TARS-2 combines scalable trajectory generation, stabilized multi-turn reinforcement learning, and a sandbox that supports GUI, filesystem, and terminal actions. The report evaluates the resulting agent across web, desktop, Android, and game tasks and analyzes training stability.
Succeed or Learn Slowly applies direct off-policy updates to successful interactions and more conservative regularized updates to negative examples. Successful Transition Replay prioritizes successful experience during policy learning, with evaluation on AndroidWorld.
UItron combines GUI perception and planning data engineering with supervised training and curriculum reinforcement learning in connected mobile and PC environments. It also collects more than one million operation steps across 100 Chinese apps and builds offline and online evaluation environments.
CODA combines a generalist planner with a specialist GUI executor and trains application-specific planners through decoupled GRPO. Successful specialist trajectories then support supervised training of a consolidated planner, evaluated across four ScienceBoard applications.
InquireMobile trains mobile agents to seek human confirmation at critical decisions using staged reinforcement fine-tuning and pre-action reasoning. InquireBench evaluates safe interaction and proactive inquiry, distinguishing asking for help from completing the underlying task.
Combines API and GUI actions with distributed desktop reinforcement learning and alternating supervised training to address training instability.
CRAFT-GUI organizes GUI trajectories by difficulty for curriculum-based GRPO training. Its reward combines rule-based signals with model judgments, with evaluation on Android Control and an internal online benchmark.
Studies screenshot-based GUI grounding and navigation using curated interaction data, reward design, and reinforcement fine-tuning.
Uses multiple candidate answers and adaptive exploration rewards to improve spatial and semantic alignment in GUI grounding.
GUI-RC aggregates multiple grounding predictions into spatial voting grids to choose a consensus region. GUI-RCPO then turns agreement with that region into a reward for test-time policy updates on unlabeled GUI screenshots.
SEAgent explores unfamiliar software using a world-state assessor and an automatically generated task curriculum. It learns from failed and successful interactions through adversarial imitation and GRPO, then combines specialist experience into a generalist agent evaluated on five OSWorld software environments.
Introduces GuirlVG, an empirical study of reinforcement fine-tuning for GUI visual grounding. It analyzes training components and adds an adaptive KL-based mechanism to limit reward over-optimization.
The Self-Evolution Agent combines automatic generation of verifiable computer-use trajectories with step-wise reinforcement learning for long-horizon training. It also describes a model-enhancement procedure that integrates grounding and planning capabilities without an additional training stage.
NaviMaster trains a shared policy for GUI and embodied navigation using a common visual-trajectory formulation. It combines trajectory collection, reinforcement learning and distance-aware rewards, with evaluation across both task families.
Combines grounding-oriented reward and sampling improvements with inference-time image decomposition and target selection.
GUI-G2 replaces binary grounding rewards with Gaussian point rewards and spatial-coverage rewards. An adaptive variance tied to element dimensions supplies continuous training feedback for targets of different sizes, evaluated on the ScreenSpot benchmark family.
Combines a mobile GUI data pipeline, perception and planning capabilities, and reinforcement fine-tuning for interactive task execution.
GTA1 samples multiple action proposals at each GUI step and uses a judge to select among them. A separate grounding model is trained with reinforcement learning to map selected proposals to visual targets, linking planning-time search with click localization.
MobileGUI-RL constructs an online training curriculum through agent exploration and task filtering. It adapts GRPO with trajectory-aware advantages and rewards for task completion and execution efficiency, evaluated on three online mobile-agent benchmarks.
Studies compute allocation for web-agent post-training with teacher-generated supervised fine-tuning followed by on-policy reinforcement learning. It samples 1,370 training configurations and uses bootstrap analysis to examine hyperparameter sensitivity on WorkArena and MiniWoB++.
Uses a staged curriculum from response-format alignment through action feedback to interactive task-level training for mobile agents.
CogniGUI combines hierarchical visual parsing of interface elements with a GRPO-based grounding agent that compares alternative interaction paths. Its ScreenSeek benchmark evaluates navigation across applications, changing interface states, and consistency between interfaces.
Location Preference Optimization uses an information-entropy signal to select interaction regions and a distance-based reward to assess click locations. It trains GUI localization through preference optimization and evaluates both offline grounding and online interaction.
Test-Time Interaction scales the interaction horizon so web agents can explore, backtrack and revise plans within a rollout. Its online reinforcement-learning curriculum adjusts rollout lengths and is evaluated on WebVoyager and WebArena.
Combines grounding pretraining, bilingual interaction trajectories, and reinforcement fine-tuning for mobile GUI actions.
Combines generated tasks, estimated success rewards, and online reinforcement learning to automate GUI-agent interaction training.
Uses rule-based rewards for output structure and action correctness to train reasoning for enterprise web navigation.
UI-Genie jointly improves a mobile GUI agent and a reward model that processes screenshots with interaction history and supplies action- and task-level feedback. Its data pipeline uses rule-based verification, corrupted trajectories, and hard negatives to support reward-guided exploration and iterative training.
Trains ScreenExplorer with reinforcement learning in dynamic GUI environments, using a world-model curiosity reward and experience-stream distillation to support sustained exploration.
ProgRM predicts per-step task completion progress to provide intermediate rewards during online GUI-agent training. It derives progress labels by using longest common subsequences to identify key trajectory steps, allowing useful partial progress to contribute even when a trajectory ultimately fails.
Trains web agents through asynchronous multi-turn environment interaction, using task-success rewards to optimize complete trajectories.
ARPO extends group-relative policy optimization with a replay buffer that reuses successful GUI trajectories across training iterations. It filters tasks using baseline performance and compares end-to-end reinforcement learning with offline preference optimization on OSWorld.
Analyzes reasoning templates, reward design, and policy updates in GUI grounding, then proposes targeted adjustments to the training process.
Web-Shepherd is a process reward model for assessing individual steps of web-navigation trajectories. It introduces a collection of 40,000 step-level preference pairs with checklists and a separate benchmark for comparing reward models, then evaluates verification during WebArena-lite navigation.
Combines self-generated localization reasoning, spatial consistency checks, and inference-time search to improve data-efficient web grounding.
GUI-Shift trains an inverse-dynamics task that predicts the initial action connecting two GUI states, using trajectories without natural-language instructions. It combines rule-based reinforcement learning with data filtering and evaluates transfer to Android automation and screenshot grounding.
Proposes GUI grounding training that combines curated seed examples, continuous location-dependent policy feedback, and iterative refinement guided by attention maps. The study evaluates the resulting reinforcement-learning procedure on three grounding benchmarks, including ScreenSpot-Pro.
MobileIPL samples trees of action-planning thoughts, propagates rule-based leaf rewards backward, and forms preference pairs for reasoning-level optimization. It also expands screenshot-based instruction data through staged evolution before evaluation on mobile GUI-agent benchmarks.
Formalizes GUI tasks as Markov decision processes and traces agent training from prompting through supervised fine-tuning to reinforcement learning.
Uses a staged reasoning-focused training framework to move GUI agents from reactive actions toward deliberative planning and recovery.
Applies reinforcement fine-tuning with unified action rules and curated cross-platform examples to vision-language GUI agents.
Uses rule-based action rewards and policy optimization to train GUI action prediction from a compact set of mobile tasks.
Labels individual computer-use actions by comparing surrounding screenshots, then trains agents from both successful and unsuccessful steps.
Trains GUI agents with a frozen value model learned from offline interactions, guiding action exploration without predicting complete next observations.
Learns action-value functions from offline device-control experience and uses them to rank candidate actions for policy improvement.
Uses variational subgoal-conditioned reinforcement learning to improve learning efficiency on sparse-reward mobile and web-control tasks.
Combines autonomous task proposal, web interaction, and visual success evaluation to discover skills and improve browser-agent policies.
Introduces StepAgent, which compares agent and expert actions to supply intermediate rewards for policy learning, with implicit-reward and inverse-RL variants. Evaluation includes WebShop and Mind2Web alongside non-web tasks.
Trains web agents with an evolving task curriculum, outcome supervision, and reinforcement-learning updates intended to manage policy drift.
Presents AutoGLM agents for browser and phone control, separating planning from grounding and training through an evolving interaction curriculum.
第 121–180 条，共 242 条资源。
通过主题导出清单获取完整记录和证据。克隆目录后，可将主题记录直接交给标准 RAG 与引用导出工具。
