Generation of digital content navigation data
AssigneeAdobe Inc.
InventorSuyash
Life of this patent
Estimating the term…
Abstract
Navigation data generation techniques and systems are described to address the complexities of digital content and that overcome the challenges of the conventional techniques. In one example, digital content is received by a flow graph system and used to generate a flow graph that models functionality available via the digital content as graph nodes and connections between the functions as edges between the graph nodes based on a markup language of the digital content. Each of the graph nodes includes a respective utterance that describes functionality available via that node and thus is usable to locate this functionality using semantic similarity to an input query. The flow graph is used as a basis to generate navigation data.
Claims20
A method implemented by a computing device, the method comprising:
- receiving, by the computing device, a flow graph generated from digital content, the flow graph including graph nodes describing functions available from the digital content and edges connecting the graph nodes that represent navigation between the functions that is supported within the digital content;
- generating, by the computing device using a language model as part of machine learning, a word embedding of an input query and a plurality of word embeddings of the plurality of graph nodes;
- comparing, by the computing device, the word embedding of the input query with the plurality of word embeddings of the plurality of graph nodes;
- determining, by the computing device, a graph node of the plurality of graph nodes corresponds to the input query based on the comparing;
- detecting, by the computing device, a graph node of the plurality of graph nodes corresponds to a state of execution of the digital content;
- determining, by the computing device, a path in the flow graph between the graph node corresponding the state of execution of the digital content and the graph node corresponding to the input query;
- and
- outputting, by the computing device, the path as indicating navigation in the digital content as a series of functions corresponding to respective said graph nodes included in the path between the graph node corresponding the state of execution of the digital content and the graph node corresponding to the input query and navigation between the series of functions in the path using respective said edges in the path.
A system comprising:
- a markup node detection module implemented at least partially in hardware of a computing device to detect a plurality of markup nodes in a markup language of a web application;
- a graph node generation module implemented at least partially in hardware of the computing device to generate a plurality of graph nodes that model functions of the web application based on the plurality of markup nodes, each of the plurality of graph nodes including a respective utterance;
- and
- a flow graph generation module implemented at least partially in hardware of the computing device to:
- generate a flow graph including the plurality of graph nodes and edges connecting the graph nodes that represent navigation between the functions that is supported within the web application, the flow graph supporting generation of navigation data based on a path within the flow graph defined using respective graph nodes and edges between:
- a graph node of the plurality of graph nodes that corresponds to a state of the web application;
- and
- a graph node of the plurality of graph nodes that corresponds to an input query;
- and
- output the path as indicating navigation in the digital content as a series of functions corresponding to the respective graph nodes included in the path between the graph node corresponding the state of execution of the digital content and the graph node corresponding to the input query and navigation between the series of functions in the path using respective said edges in the path.
A system comprising:
- means for receiving a flow graph generated from digital content, the flow graph including graph nodes describing functions available from the digital content and edges connecting the graph nodes that represent navigation between the functions that is supported within the digital content;
- means for generating a word embedding of an input query and a plurality of word embeddings of the plurality of graph nodes using a language model as part of machine learning;
- means for comparing the word embedding of the input query with the plurality of word embeddings of the plurality of graph nodes;
- means for determining a graph node of the plurality of graph nodes corresponds to the input query based on the comparing;
- means for detecting a graph node of the plurality of graph nodes corresponds to a current state of execution of the digital content;
- and
- means for determining a path in the flow graph indicating navigation in the digital content as a series of functions corresponding to respective said graph nodes included in the path between the graph node corresponding the state of execution of the digital content and the graph node corresponding to the input query and navigation between the series of functions in the path using respective said edges in the path.
Dependencies are inferred from the claim text.