Intelligent Web and Desktop Assistant

The project has two main aspects. The intelligent assistant consists of an intelligent Web agent and a Desktop agent. The Web agent is implemented in JAVA and the Desktop agent is implemented in Visual Basic.

Web Agent

The Web-agent currently is currently equipped with following functionalities :-

The Web-agent can suggest user a list of web-sites which are of similar field of which user visited web-site was.

The Web-agent blocks the web-sites tagged as black-listed. The black-listing is also automated by web-agent.

The back-ground web-crawler investigates the previous visited web-sites to match the extent of user field of interest.

Desktop Agent

The Desktop Agent Application can assist a user in the following ways.

The Desktop agent maintains a daily log of user accessed files (with links).

The Agent frequently greets the user with date and time dialog.

Animated agent interface is provided to user.

The user can open files and applications using the Application Tray. Application Tray is the window that contains the list of recently used applications.

The Agent creates a virtual space to open any application or file that have been accessed at least once. These applications can be invoked using assigned tags (a keyword resembling the application).

The Agent Application rebuilds the Desktop. This includes changing the wallpaper frequently, monitoring Recycle Bin, and prompting the user to move less frequently used desktop items to a temporary folder.

Alerts the user on “disk full” conditions and provides the possible means for recovering Disk space.

Schedules time-based applications.

Brief Description:

The aim of the project is to build an Autonomous Agent which keeps track of user-working habits and builds a “profile” in terms of application parameters, uses the profile in assisting the user. The project also aims at observing the learning process (converging to a stable situation which best reflects the user behavior), observing the user feedback to find out whether such an agent becomes boon or bane for the user and at the same time the project also concentrates at identifying the new means of communication between the user and the computer (Machine-Human Interaction).

The common mentality of a user is to keep the things arranged neatly; the same applies to a computer desktop too. User can put any number of shortcuts for the applications present in the computer but it causes irritation and the desktop seems ugly too. There are many ways to arrange the shortcuts for the applications like keeping the one used more frequently on the desktop and leaving the remaining in other menus and so on. But as the user needs changes he may use new applications more frequently which was not that much important for him previously. The agent comes into picture in such cases, the keeps tracks the applications the user is using and builds a profile for the user about the applications, creates a list of frequently used applications and keeps them a “click” away from the user. This is one application built atop the learnt “user profile”. The list reflects the user working nature and modifies accordingly.

The other applications are, providing a tool that takes (optionally) a hint or tag representing an application used by the user to invoke any application. Unlike many tools available to do the same this application never makes the user to take even a bit of burden to locate the path for a particular application, it has the ability to find them on its own as the user regular usage of an application is itself a “hint”. The agent also helps in automating some of the tasks like Recycle Bin and system tools like Disk Defragmenter (Scheduling tasks).