Aspect pre-selection using machine learning
Aspect pre-selection techniques using machine learning are described. In one example, an artificial assistant system is configured to implement a chat bot. A user then engages in a first natural-language conversation. As part of this first natural-language conversation, a communication is generated by the chat bot to prompt the user to specify an aspect of a category that is a subject of a first natural-language conversation and user data is received in response. Data that describes this first natural-language conversation is used to train a model using machine learning. Data, is then be received by the chat bot as part of a second natural-language conversation. This data, from the second natural-language conversation, is processed using the model as part of machine learning to generate the second search query to include the aspect of the category automatically and without user intervention.
1. A method implemented by a computing device, the method comprising:
generating, of the computing device, a communication to prompt a user to specify an aspect of a category that is a subject of a first natural-language conversation;
receiving, by the computing device via a user interface, user data describing the specified aspect in response to the communication;
generating, by the computing device, a first search query including the specified aspect as text generated from the user data;
training, by the computing device, a model using machine learning based on the specified aspect using the text generated from the user data;
receiving, by the computing device, data describing a second natural- language conversation;
identifying, by the computing device, the specified aspect by processing the data describing the second natural-language conversation using the model as part of machine learning;
automatically adding, by the computing device, the text of the specified aspect to a second search query responsive to the identifying without user intervention and without indicating, in the user interface, the adding of the text to the second search query; and
outputting, by the computing device, a result of a search performed using the second search query and an indication in the user interface that the text of the specified aspect is added to the second search query.
2. The method as described in claim 1 , further comprising identifying, by the computing device, a product type and further comprising locating the model based on the product type.
3. The method as described in claim 2 , wherein the identifying includes extracting a dominant object as the product type.
4. The method as described in claim 3 , wherein the extracting is performed by walking a category tree based on the received data as part of the second natural-language conversation.
5. The method as described in claim 3 , wherein the extracting is based at least in part on a digital image included as part of the first natural- language conversation.
6. The method as described in claim 1 , further comprising identifying, by the computing device, a user associated with the second natural- language conversation and further comprising locating the model based on the identified user.
7. The method as described in claim 6 , wherein the identifying is performed to identify the user from a plurality of users associated with a single user account.
8. The method as described in claim 1 , wherein the outputting of the result of the search includes displaying an indication that the specified aspect is used as part of the search.
9. The method as described in claim 8 , wherein the indication is user selectable to initiate a subsequent search query that does not include the specified aspect.
10. The method as described in claim 1 , wherein the outputting is configured to cause output of the result of the search as part of a live camera feed used to generate the user data.
11. The method as described in claim 1 , wherein the second natural-language conversation does not include the prompt.
12. A method implemented by a computing device, the method comprising:
extracting, by the computing device, a user identifier from input data, the input data generated as part of a natural-language conversation between a user and an artificial assistant system;
locating, by the computing device, a model based on the user identifier, the model trained as part of machine learning;
generating, by the computing device, data describing an aspect of a category, the generating performed by processing the input data using the located model as part of machine learning;
generating, by the computing device, a search query based on the input data, the generating including automatically adding the data describing the aspect to the search query without user intervention;
automatically initiating, by the computing device responsive to the generating of the search query, a search using the search query that includes the data; and
outputting, by the computing device, a result of the search performed using the search query and the data describing the aspect, the result including an indication that the data is added to the search query and an option displayed proximal to the data that is user selectable to remove the data from the search query.
13. The method as described in claim 12 , wherein the model is trained based on a previous natural-language conversation between the user and the artificial assistant system.
14. The method as described in claim 13 , wherein the previous natural-language conversation includes generating a communication by the artificial assistant system to prompt the user to specify the aspect of the category.
15. The method as described in claim 12 , wherein the result is configured for display in a user interface and includes an indication that the aspect is used to generate the search result.
16. The method as described in claim 12 , further comprising extracting a product identifier as part of the natural-language conversation and the locating is also based at least in part on the extracted product identifier.
17. The method as described in claim 16 , wherein the input data includes a digital image and the extracting is based on the digital image.
18. A computing device comprising:
a processing system; and
a computer-readable storage medium having instructions stored thereon that, responsive to execution by the processing system, causes the processing system to perform operations including:
receiving data describing a natural-language conversation performed via a user interface;
identifying an aspect by processing the data describing the natural-language conversation using a model trained as part of machine learning;
automatically adding, without user intervention, text of the aspect to a search query responsive to the identifying and without indicating, in the user interface, the adding of the text to the search query; and
outputting a result of a search performed using the search query and an indication that the text of the specified aspect is added to the search query.
19. The computing device as described in claim 18 , further comprising automatically initiating, without user intervention, the search using the search query.
20. The computing device as described in claim 18 , wherein the outputting is configured to cause output of the result of the search as part of a live camera feed used to generate the user data.