[2505.10831] Creating General User Models from Computer Use

View the PDF file for the paper entitled Create General user models from the use of the computer, by Omar Shaykh and 6 other authors

PDF HTML (experimental) view

a summary:The interaction between man and computer has long imagined a technique that the United States understands our preferences and customs, to the timing and purposes of our daily actions. However, the current user models remain fragmented, tightly designed for specific applications, and are unable to think flexible to achieve these visions. This paper offers a structure for the general user model (gum) that learns about you by monitoring any interaction with your computer. The gums are taken as inputs any unorganized note to the user (for example, screenshots of the device) and adopting likely confidence proposals that pick up the user’s knowledge and preferences. The gums can conclude that the user is preparing for a wedding that brings it from messages with a friend. Or we realize that the user struggles through the collaborators’ observations on a draft by monitoring multiple related adjustments and switching to reading works. The gums are offered a brown raising new proposals about a user of multimedia notes, recalls related to the context, and the current proposals are continuously revised. To clarify the expansion of the gum applications, we explain how to increase the chat -based assistants with the context, manage the operating system notices to selectively important information, and enable interactive factors that adapt to preferences through applications. We also establish pre -emptive aides who discover and implement useful suggestions on behalf of the user using the gums. In our assessments, we find that the gums are conducting calibration and accurate conclusions about users, and that the integrated gum assistants determine proactively and perform the procedures that users do not think explicitly request. In all, the gums offer ways to benefit from multimedia models to understand the irregular context, allowing long -term visions of HCI and completely new interactive systems that expect the user’s needs.