Source-reviewed research and assets tagged with Long-horizon tasks.
Membership follows reviewed tags; it does not establish a dependency between neighboring records. Dates preserve source precision. Annotations retain their original language.
ColorBrowserAgent converts human feedback into reusable website knowledge and uses that knowledge to guide progressive context summaries. It evaluates the resulting memory and adaptation mechanisms on long browser workflows.
AndroidLens provides 571 extended mobile tasks in Chinese and English, with multiple valid paths and preserved environmental anomalies. Its dynamic evaluation uses nested milestones and Average Task Progress to measure partial completion of long workflows.
Introduces long mobile workflows that combine agent-user interaction, cross-application tasks, and MCP-based tool access.
Introduces AgentProg, organizing mobile-agent interaction history as a program with variables and control flow to decide what context to retain. A global belief state handles partial observability, with evaluation on AndroidWorld and extended long-horizon tasks.
This work trains high-level schedulers with staged reinforcement learning driven by execution feedback. Its CES framework separates a Coordinator for planning and decomposition, a StateTracker for compressed context, and a low-level Executor.
Uses separate retrieval systems for high-level mobile-task planning and precise interface operations, with a benchmark for extended cross-app tasks.
Introduces MGA, which separates task-agnostic screen-state reading from decision making and connects steps through structured state memory. Validated state deltas replace full trajectory replay, evaluated on OSWorld and application tasks.
Surfer2 combines hierarchical context management with separate planning and execution, self-verification, and recovery for visual computer use. It evaluates the framework across web, desktop, and Android tasks without task-specific fine-tuning.
ColorBench represents mobile tasks as graphs of states collected from real devices, providing a quasi-dynamic environment with multiple valid paths and error paths. It evaluates single-app and cross-app tasks using completion, subtask progress, and atomic diagnostics.
Guides UI agents with symbolic planning over a UI transition graph initialized from app-code analysis and updated through online exploration, with evaluation on AndroidWorld.
Introduces Behavior Judge, which converts execution trajectories into behavior narratives and compares multiple candidate rollouts. Experiments examine this selection approach on OSWorld, WindowsAgentArena and AndroidWorld.
GUI-ReWalk begins with stochastic interface exploration and then infers goals to generate coherent action trajectories. It supports long-horizon workflows across applications and studies training on the resulting data using grounding and navigation benchmarks.
MobileRAG combines retrieval and memory components to support mobile tasks that need external knowledge or long action sequences. MobileRAG-Eval is introduced to study these requirements on complex real-world app tasks.
Describes a mobile-agent stack combining region-based coordinate correction, hierarchical task planning, multi-agent voting, and reuse of successful interaction traces.
FlashAdventure evaluates GUI agents on completing story arcs in 34 adventure games. It pairs an automated gameplay judge with COAST, a framework that keeps long-term clue memory to plan and complete sequential objectives.
NaturalGAIA builds verifiable long-horizon GUI tasks from human interaction intents while separating causal task structure from its verbal description. Its LightManus-Jarvis framework divides dynamic planning and context management from execution based on visual and structural interface information.
MobileUse reflects on errors at action and task timescales, invoking reflection when needed and using proactive exploration to learn unfamiliar apps. It evaluates this mobile execution framework on AndroidWorld and AndroidLab and describes a toolkit for physical-device operation.
AgentSynth composes easily generated subtasks into longer computer-use tasks and trajectories with adjustable difficulty. Its evaluation uses the number of subtasks to probe how agent success changes as task complexity grows.
Mirage-1 organizes interaction trajectories into execution, core, and meta-skills, then uses them to guide Monte Carlo tree search during online GUI tasks. It introduces AndroidLH for long-horizon evaluation and tests the approach on Android and web-navigation benchmarks.
UI-NEXUS evaluates compositional mobile tasks in local utility apps and online Chinese and English service apps. AGENT-NEXUS decomposes those long tasks into self-contained atomic subtasks to study generalization beyond isolated screens or short action chains.
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.
Introduces synthetic CAD interaction videos, a model for learning interface actions from video, and evaluation tasks for spatial and long-horizon reasoning.
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.
Introduces OSUniverse, a desktop GUI benchmark spanning precise clicks through multistep and multiapplication tasks. It provides both manual scoring and an automated validation mechanism.
RealWebAssist evaluates web assistance through sequences of instructions collected from real users across multiple websites. It tests intent interpretation, user-specific routines, long-term context, and visual grounding when instructions are ambiguous or evolve during interaction.
Separates planning from execution and creates synthetic plan annotations for training agents on extended web-navigation tasks.
Uses hierarchical mobile agents and persistent tips and reusable action shortcuts to improve complex tasks from accumulated experience.
Simulates a software-company workplace to evaluate agents on professional tasks involving websites, programs, files, and coworker communication.
WebPilot separates global subtask planning from local Monte Carlo tree search during web execution. It continually refines the high-level plan and adapts local search to uncertain transitions and incomplete observations, evaluated on WebArena and MiniWoB++.
Evaluates language agents on office workflows requiring information integration, long-term planning, and switching between applications.
Evaluates multimodal agents on compositional tasks across changing websites and studies replaying prior trajectories as memory support.
CompWoB adds 50 composed web-automation tasks to study generalization from simpler interactions. The paper compares prompted and fine-tuned agents, balances task training data, and probes sensitivity to changes in instruction order.
Introduces reproducible web environments with task-completion checks for evaluating autonomous web agents.
A graph-based mobile GUI benchmark that evaluates long tasks through screenshot states and multiple valid or erroneous transitions, including reflection, replanning and backtracking.
Long-horizon agentic web search benchmark whose Agent-as-a-Judge framework scores time-varying, citation-backed answers, with evaluation scripts for its dev and test sets.
Multimodal web tasks in a WebArena-based environment that chain from 1 to 10 hops across sites such as a shopping store and Wikipedia, testing long-chain reasoning.
Task definitions for Mobile-Agent-E’s long-horizon, multi-app Android evaluation, packaged as a 25-record test split with rubrics and human-reference operation strings.
An OSWorld-derived desktop benchmark with an Ubuntu virtual-machine environment and expanded task configurations for measuring breadth and depth of computer interaction.
Published AgentSynth task definitions and recorded GUI trajectories, with OSWorld and InSTA training configurations and packaged screenshots, thoughts and commands.
The planner component of Plan-and-Act, generating structured high-level plans that an executor turns into web actions.
Android GUI-agent framework that keeps long-running task context in structured programs, variables and control flow, with a CLI and AndroidWorld integration.
Generates synthetic computer-use tasks and records trajectories in OSWorld desktop environments and InSTA browser environments.
A public Android GUI evaluation framework with action-history, summary and anchored-memory prompt modes, plus step-action and dependency-aware task scoring.
Replay-based policy-optimization framework for GUI agents, with distributed OSWorld rollouts and screenshot/action histories.
MemGUI-Agent training and offline evaluation tools provide trajectory conversion, LoRA SFT templates, and step-level checks of GUI actions, memory updates and context folding.
Showing 61–105 of 105 resources.
Get full records and evidence through the topic export manifest. Source records can be passed to the standard RAG and citation exporters after cloning the catalog.
