View source for 17984685. PROBABILISTIC PROCEDURE PLANNING FOR INSTRUCTIONAL VIDEOS simplified abstract (SAMSUNG ELECTRONICS CO., LTD.)

=PROBABILISTIC PROCEDURE PLANNING FOR INSTRUCTIONAL VIDEOS=


==Organization Name==

[[:Category:SAMSUNG ELECTRONICS CO., LTD.|SAMSUNG ELECTRONICS CO., LTD.]]


[[Category:SAMSUNG ELECTRONICS CO., LTD.]]

==Inventor(s)==

[[:Category:He Zhao of Richmond Hill CA (US)|He Zhao of Richmond Hill CA (US)]][[Category:He Zhao of Richmond Hill CA (US)]]

[[:Category:Mikita Andreevich Dvornik of Toronto (CA)|Mikita Andreevich Dvornik of Toronto (CA)]][[Category:Mikita Andreevich Dvornik of Toronto (CA)]]

[[:Category:Isma Hadji of London (GB)|Isma Hadji of London (GB)]][[Category:Isma Hadji of London (GB)]]

[[:Category:Richard Wildes of Toronto (CA)|Richard Wildes of Toronto (CA)]][[Category:Richard Wildes of Toronto (CA)]]

[[:Category:Konstantinos Derpanis of Toronto (CA)|Konstantinos Derpanis of Toronto (CA)]][[Category:Konstantinos Derpanis of Toronto (CA)]]

[[:Category:Allan Douglas Jepson of Oakville (CA)|Allan Douglas Jepson of Oakville (CA)]][[Category:Allan Douglas Jepson of Oakville (CA)]]

==PROBABILISTIC PROCEDURE PLANNING FOR INSTRUCTIONAL VIDEOS - A simplified explanation of the abstract==


This abstract first appeared for US patent application 17984685 titled 'PROBABILISTIC PROCEDURE PLANNING FOR INSTRUCTIONAL VIDEOS

==Simplified Explanation==

The present disclosure describes a method and apparatus for probabilistic procedure planning using machine learning. Here are the key points:

* The method involves generating an action plan based on a user's request to achieve a specific end state.
* The action plan includes a series of intermediate actions between a start state and the end state.
* An input query matrix is constructed using various parameters such as the number of intermediate actions, start state, end state, positional encodings, and pseudo-random noise information.
* A machine learning transformer decoder is used to generate the action plan based on the input query matrix and a set of learnable vectors.
* The generated action plan provides a probability distribution of distinct action sequences that the user can perform to transform the start state to the end state.
* The action plan is then provided to the user for execution.

Potential applications of this technology:

* Robotic automation: This technology can be used to plan and generate action sequences for robots to perform specific tasks or achieve desired outcomes.
* Process optimization: It can be applied in industries to optimize procedures and workflows by generating efficient action plans.
* Personal assistants: Virtual assistants can utilize this technology to generate action plans for users to accomplish tasks or reach specific goals.

Problems solved by this technology:

* Complex planning: The method simplifies the process of generating action plans by using machine learning techniques, reducing the complexity and time required for manual planning.
* Uncertainty handling: By providing a probability distribution of action sequences, the method accounts for uncertainties and allows users to choose the most suitable plan based on their preferences or constraints.

Benefits of this technology:

* Efficiency: The automated generation of action plans saves time and effort compared to manual planning.
* Flexibility: The probability distribution allows users to select the most appropriate action sequence based on their needs and constraints.
* Adaptability: The machine learning approach enables the system to learn and improve over time, adapting to different scenarios and user preferences.


==Original Abstract Submitted==

The present disclosure provides methods and apparatuses for probabilistic procedure planning for generating a plan based on a goal relating to an end state. In some embodiments, a method includes receiving a request from a user to generate an action plan comprising T intermediate actions between a start state and the end state. The method further includes constructing an input query matrix based on T, the start state, the end state, positional encodings, and pseudo-random noise information. The method further includes generating, using a machine learning transformer decoder, the action plan based on the input query matrix and a plurality of learnable vectors. The method further includes providing the action plan to the user. The action plan indicates a probability distribution of a plurality of distinct action sequences, to be performed by the user, that transform the start state to the end state.



[[Category:G06F16/532]]
[[Category:G06N7/01]]