jacob-valdez/synthux-economy-r5-w065
收藏资源简介:
SynthUX计算机使用数据集—经济模拟r5是一个基于SynthUX生成的视觉计算机使用轨迹数据集。每个记录代表一个模拟公司内工人的设备会话:目标扩展为节点树,终端节点通过低级别鼠标/键盘输入驱动真实桌面模拟器应用程序(如终端、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和每个节点的帧。数据集支持应用原生多应用程序执行(无工作台覆盖)、每个参与者内容多样性、跨操作系统消息传递和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。内容包括data.jsonl(每个渲染轨迹的记录)、screenshots(每个终端节点的帧,在节点最后输入事件时拍摄)和videos(完整屏幕录制)。记录模式包括轨迹ID、参与者ID、公司、角色ID、标题、环境(模拟器类型)、目标、节点树、输入事件、观察结果、对齐和媒体文件。该版本统计显示轨迹36个,视频36个,帧985个,公司6个,按环境分布为browser-os(12)、macos-web-next(10)、windows-web-next(14)。数据集使用SynthUX生成,观察结果是因果的(突变通道为低级别输入),捕获过程中未使用高级别应用程序突变。
The SynthUX Computer Usage Dataset: Economic Simulation r5 is a visual computer interaction trajectory dataset generated via SynthUX. Each record corresponds to a device session of a worker within a simulated company: Targets are expanded into node trees, and terminal nodes drive real desktop emulator applications (e.g., "Terminal", "Notes", "VS Code", "Browser", "Slack/Teams", "Mail", "Spreadsheets", "Slides", etc.) through low-level mouse and keyboard inputs. The observed interaction trajectories are recorded as full screen videos and per-node frame captures. This dataset supports native multi-application execution (no workspace overlay), diverse content per participant, cross-operating system messaging, and large-scale multi-company economic simulations. Each individual economic simulation is released as a standalone dataset repository. The dataset contents include three core components: data.jsonl (records for each rendered trajectory), screenshots (frames of each terminal node captured at the node's final input event), and videos (full screen recordings of entire sessions). The record schema encompasses the following fields: trajectory ID, participant ID, company, role ID, title, environment (emulator type), target, node tree, input event, observation, alignment, and media files. Statistics for this release indicate 36 trajectories, 36 videos, 985 frames, and 6 participating companies, with distribution across environments as follows: browser-os (12), macos-web-next (10), and windows-web-next (14). This dataset is generated using SynthUX, with causal observations (the mutation channel relies on low-level inputs), and no high-level application mutations were utilized during the data capture process.




