jacob-valdez/synthux-economy-r4-w026
收藏资源简介:
SynthUX计算机使用数据集(经济模拟版本r4)是一个基于视觉的计算机使用轨迹数据集,由SynthUX生成。每个记录代表一个工人在模拟公司内的设备会话:目标扩展为节点树,终端节点通过低级鼠标/键盘输入驱动真实桌面模拟器应用(包括终端、笔记、VS Code、浏览器、Slack/Teams、邮件、表格、幻灯片等),观察到的轨迹被记录为屏幕视频加上每个节点的帧。该数据集支持应用原生多应用执行(无工作台覆盖),每演员内容多样性,跨操作系统消息传递,以及大型多公司经济模拟。每个经济模拟作为独立的数据集仓库发布。数据集内容包括data.jsonl文件(每个渲染轨迹一条记录)、screenshots文件夹(每个终端节点的帧,在节点的最后一个输入事件时拍摄)和videos文件夹(完整屏幕录制)。记录模式包括轨迹ID、演员ID、公司、角色、标题、环境、目标、节点树、输入事件、观察、对齐和媒体信息。统计显示本构建包含36个轨迹、36个视频、1026帧、6个公司,环境分布为browser-os(16)、macos-web-next(6)和windows-web-next(14)。数据生成使用SynthUX工具,观察基于因果模拟(突变通道为低级输入),捕获过程中未使用高级应用突变。
The SynthUX Computer Usage Dataset (Economic Simulation Version r4) is a vision-based computer usage trajectory dataset generated by SynthUX. Each record represents a worker's device session within a simulated company: goals are expanded into node trees, and terminal nodes drive real desktop simulator applications (including terminals, note-taking tools, VS Code, browsers, Slack/Teams, email clients, spreadsheets, presentation software, etc.) via low-level mouse/keyboard inputs. The observed trajectories are recorded as screen videos plus frames for each terminal node. This dataset supports native multi-application execution (without workspace overlay), content diversity per actor, cross-OS messaging, and large-scale multi-company economic simulations. Each economic simulation is released as an independent dataset repository. The dataset contents include the data.jsonl file (one record per rendered trajectory), the screenshots folder (frames of each terminal node captured at the time of the node's final input event), and the videos folder (full screen recordings). Recorded fields include trajectory ID, actor ID, company, role, title, environment, goal, node tree, input events, observations, alignments, and media information. Statistics show that this build contains 36 trajectories, 36 videos, 1026 frames, and 6 companies, with an environment distribution of browser-os (16), macos-web-next (6), and windows-web-next (14). The data was generated using the SynthUX tool, with observations based on causal simulation (using low-level inputs as the mutation channel), and no high-level application mutations were utilized during the capture process.




