Data Modeling - Predicting User Behavior (Microsoft) - Patent Application - PRIOR ART REQUEST

Question

This Patent Application has received a non-final rejection by the US Patent Office! An initial rejection is part of the typical course of a patent application.

---

Thanks to YOU, the Ask Patents community, overly-broad claims have at least been narrowed. Follow @askpatents to block more overly-broad patent applications. 10 minutes of your time can help narrow US patent applications before they become patents.

AN OVERBROAD PATENT ON PREDICTING USER BEHAVIOR - This application from Microsoft seeks to patent the idea of...automatically predicting actions a user is likely to take in response to receiving data! 10 minutes of your time can help narrow US patent applications before they become patents. Follow @askpatents on twitter to help.

QUESTION - Have you seen anything that was published before Dec 15, 2011 that discusses:

1. A method of MONITORING USER ACTION in response to INCOMING DATA FOR A USER; determining a PATTERN OF USER ACTIONS; and
2. Predicting the USER’S ACTION in response to his receiving NEW INCOMING DATA;

If so, please submit evidence of prior art as an answer to this question.. We welcome multiple answers from the same individual.

EXTRA CREDIT - A reference to anything that meets all of the criteria to the question above AND ALSO uses a SEED MODEL to make predictions for new users, to utilize OTHER USER BEHAVIOR for existing and new users.

TITLE: PREDICTION OF USER RESPONSE TO RECEIVED DATA

Summary: [Translated from Legalese into English] A method of detecting when a user receives new data, monitoring what the user does in response the the data, and building a model of the user’s future behavior, and predicting the user’s action in response to new incoming data.

• Publication Number: US 20130159220 A1
• Application Number: 13/326,877
• Assignee: Microsoft, Inc.
• Prior Art Date: Seeking prior Art predating Dec 15, 2011
• Open for Challenge at USPTO: Open through Dec 20, 2013

Claim 1 requires each and every step below:

A method executed for predicting user actions in response to receiving incoming data, the method comprising:

1. detecting incoming data from a data source for a user;

2. monitoring the user's actions in response to receiving the incoming data;

3. determining a pattern of user actions in response to the received incoming data;

4. detecting new incoming data from a data source; and

5. predicting the user's action in response to receiving the new incoming data based on the pattern of user actions, an incoming data context and an incoming data content.

Claim 19 gives insight into what Microsoft claims to have invented with more specificity

19. {A method, comprising}:

20. detecting incoming data from a data source for a user;

21. extracting one or more features of the incoming data utilizing a feature extractor component;

22. enabling a sensor component to continuously detect the user's actions in response to receiving the incoming data;

23. uploading detected user action information from the sensor component and extracted feature information from the feature extractor component to an actions processor;

24. enabling the actions processor to combine, store, and organize the information from the sensor component and the feature extractor component into a user action database, wherein the user action database includes a list of past actions the user has taken in response to receiving data having one or more features;

25. determining a pattern of user actions in response to the received incoming data;

26. detecting new incoming data from a data source; and

27. predicting the user's action in response to receiving the new incoming data based on the pattern of user actions, an incoming data context and an incoming data content.

In English this means:

A method predicting user actions comprising:

1. Detecting that user has received some new data; and

2. Monitoring what the user does in response to the data; and

3. Extracting patterns of user action in response to the new data; and

4. Predicting the user’s action based on the previous pattern of user actions and the incoming data

Good prior art would be evidence of a system that did each and every one of these steps prior to the Dec 15, 2011.

You're probably aware of ten pieces of art that meet this criteria already... separately, the applicant is claiming using a SEED MODEL to make predictions for new users, to utilize OTHER USER BEHAVIOR for existing and new users.

---

"Cloud-based system for observing and predicting user actions” from Microsoft

---

What is good prior art? Please see our FAQ.

Want to help? Please vote or comment on submissions below. We welcome you to post your own request for prior art on other questionable US Patent Applications.

---

Answer 1

Benyon, David. "Accommodating individual differences through an adaptive user interface." HUMAN FACTORS IN INFORMATION TECHNOLOGY 10 (1993).

Computer systems can monitor the interaction at a level of detail unavailable to other artefacts. If the system is supplied with a suitable theory of interaction and how interaction can be improved, the computer is in a position to change its functioning, its structure or the representations provided at the interface to better match the preferences, needs and desires of the users. The computer can adapt itself to individuals or groups of users

Also: Exploring mouse movements for inferring query intent (2008)

We explore the potential of a complementary, more sensitive signal -mouse movements- in providing insights into the intent behind a web search query Here also is a paper from 2008 with other seemingly good references regarding user mouse tracking on a web page. Probably more directly relevant:

Also: WWW '06 Proceedings of the 15th international conference on World Wide Web 203-212

In this paper, we investigate how detailed tracking of user interaction can be monitored using standard web technologies. Our motivation is to enable implicit interaction and to ease usability evaluation of web applications outside the lab. To obtain meaningful statements on how users interact with a web application, the collected information needs to be more detailed and fine-grained than that provided by classical log files

Answer 2

I think that any prior art on user predictive analytics can match the request. In particular, I think that Search completion (like Google search completion) is probably exactly that.

I found the following:

Systems and methods for joint analytics on user level and network level data of a communications network US 8447269 B2

Embodiments utilize joint analytics to generate patterns and rules concerning the content and services accessed by a user, when they are accessed, and how they are accessed.

Contextual prediction of user words and user actions WO 2005036413 A8

The invention concerns user entry of information into a system with an input device (12). A scheme is provided in which an entire word that a user wants to enter is predicted, and shown a display (10), after the user enters a specific symbol, such as a space character. If the user presses an ambiguous key thereafter, rather than accept the prediction, the selection list is reordered. The invention can also make predictions on context such as the person to whom the message is sent, the person writing the message, the day of the week, the time of the week, etc.

Action Prediction and Identification Temporal User Behavior US 20120123993 A1

From the mined temporal-based user actions, future actions can be predicted. Certain implementations include providing information and/or services based on the predicted future actions. Some implementations, include providing relevant information, services, and/or goods regarding the predicted future action

Answer 3

This all sounds awfully familiar to inverse reinforcement learning, where the (technical) goal is to construct a reward function from observations of how a user responds to incoming data. This reward function can then be used (with ordinary reinforcement learning or dynamic programming) to make predictions about what the actions of new user would be. See, e.g., Ng & Russell, (2000)

Answer 4

Dated from April 2011, a method that plays rock-paper-scissor :

The method:

• Detecting that user has received some new data (aka last response to the previous challenge)
• Monitoring what the user does in response to the data (aka observing the response to the new challenge)
• Extracting patterns of user action in response to the new data (aka observed specific sequences of moves)
• Predicting the user’s action based on the previous pattern of user actions and the incoming data (that's exactly the point of the algorithm)

Furthermore, concerning seed model that's exactly what has been done in the present case to have a good "default" behavior.

Overall, it looks to me as a quasi-identical setup, just declined for rock-paper-scissors.