Intent-oriented internet browsing
Inventors
Assignees
Publication Number
US-11625449-B2
Publication Date
2023-04-11
Expiration Date
2039-11-18
Interested in licensing this patent?
MTEC can help explore whether this patent might be available for licensing for your application.
Abstract
A system that provides intent-oriented browsing powered by machine learning and crowdsourcing. The system allows users to enter their intents, which are then assigned to target pages via supervised learning models based on hyperlinks and contributions made by other users. The system has a prediction server that is programmed to receive hyperlinks from a website and return target hyperlinks based on known intent, a user interface for inputting user intent, and a browser programmed to connect to the intent repository and to the prediction server via a user script. The list of supported intents can grow over time based on correct page marks for intent-page mappings as well as via continuous training of machine learning models.
Core Innovation
The invention provides a system for intent-oriented internet browsing that utilizes machine learning and crowdsourcing to enable users to reach specific webpages across websites with varying architectures. Users can input their browsing intents into a user interface, and the system identifies and directs them to the corresponding target pages using supervised learning models trained on hyperlink features, as well as user-contributed mappings. The system supports a growing set of intents as more users interact and mark correct intent-page mappings.
A key component of the system is an intent repository that stores known intents, associated target pages, and hyperlink hashes, allowing the browser to quickly access and match user-entered intents. The prediction server receives hyperlinks from a website and, by leveraging machine learning models, classifies and returns target hyperlinks that correspond to the known intents. This process is augmented by crowdsourcing, where users can confirm target pages, manually mark correct or incorrect associations, or request real-time assistance from other users.
The problem addressed by the invention is the challenge users face in achieving high-level goals on websites with inconsistent navigation structures and lack of explicit intent support, particularly for users with visual impairments who struggle with inter-page navigation. Existing solutions focus mostly on making individual webpages accessible rather than supporting these user goals directly. The invention aims to reduce unnecessary navigation and improve accessibility by allowing direct mapping of user intents to target pages, thus catering to tasks that are frequently encountered or difficult to locate.
Claims Coverage
The patent claims two main independent inventive features: a system and a method for intent-based web page selection, each with specific inventive aspects.
System for providing intent-based web page selection
The system comprises: - An intent repository containing known intents mapped to specific target pages and hyperlink hashes. - A website server programmed to transmit multiple hyperlinks representing pages. - A prediction server designed to receive hyperlinks from the website server and return target hyperlinks corresponding to known intents, using a machine learning model trained with hyperlink features. - A browser running a user script that connects to the intent repository and prediction server, allowing users to enter a selected intent and displaying any matching target hyperlinks received from the prediction server.
Method for providing intent-based web page selection
The method includes: 1. Provision of an intent repository with known intents, target pages, and hyperlink hashes. 2. Provision of a website server that transmits multiple page-representing hyperlinks. 3. Operation of a prediction server to receive these hyperlinks and return target hyperlinks tied to known intents. 4. Use of a browser with a user script to connect to the intent repository and prediction server. 5. Allowing users to enter selected intents. 6. Displaying any target hyperlinks from the prediction server that match the user-selected intent.
In summary, the claims cover both the system and method for enabling users to input their intents and receive direct links to target pages by integrating an intent repository, prediction server with machine learning capabilities, and browser interface, facilitating efficient and intent-driven web navigation.
Stated Advantages
Allows users to be directly taken to the desired page, avoiding unnecessary navigation and inconsistent website designs.
Improves accessibility for users with visual impairments by reducing the need for complex inter-page navigation and reliance on assistive technology.
Enables efficient completion of common or infrequently performed tasks across different websites, even if those sites have different structures.
Provides a straightforward and easy-to-use approach, allowing users to enter intents directly and be guided to target pages with minimal effort.
Supports crowdsourcing and human assistance features, permitting users to get real-time help or benefit from contributions made by others.
Reduces time spent on task completion, particularly for blind and low-vision users, as shown by user study results.
Enables the system to grow and adapt over time by learning from user-contributed correct intent-page assignments.
Documented Applications
Facilitating users’ completion of navigation and transactional tasks (such as login, sign up, reset password, contact, browse items, track order) across various websites regardless of their architecture.
Enhancing web accessibility for blind and low-vision users by supporting direct access to target pages.
Serving as a supplemental or fallback tool for users navigating unfamiliar or complex websites.
Utilization as a feature within personal assistants or assistive technology aimed at supporting user goals on the web.
Applicable to browser extensions or user scripts (such as those managed by TamperMonkey) to enable intent-based browsing in browsers on desktops and mobile devices.
Interested in licensing this patent?