Deep learning approach to mitigate the cold-start problem in textual items recommendations
AssigneeIntuit Inc.
InventorsElik Sror, Oren Sar Shalom, Rami Cohen
Life of this patent
Estimating the term…
Abstract
A method for mitigating cold starts in recommendations includes receiving a request that identifies a requested page and identifying a content vector of the requested page. The content vector is generated based on providing text of the requested page to a neural network text encoder. The method further includes selecting, based on the content vector, a link to a cold start page that does not satisfy a threshold level of interaction data. The selected link is ranked above a second link to a warm page that does satisfy the threshold level of the interaction data. The method further includes presenting the requested page with the selected link.
Claims20
A method comprising:
- receiving a request that identifies a requested page;
- identifying a content vector of the requested page, the content vector generated based on providing text of the requested page to a neural network text encoder;
- selecting, based on the content vector, a link to a cold start page, the cold start page not satisfying a threshold level of interaction data, the selected link ranked above a second link to a warm page that does satisfy the threshold level of the interaction data;
- and
- presenting the requested page with the selected link.
A system comprising:
- a processor;
- a memory coupled to the processor;
- and
- the memory comprising a server application, wherein the server application executes on the processor and is configured for:
- receiving a request that identifies a requested page;
- identifying a content vector of the requested page, the content vector generated based on providing text of the requested page to a neural network text encoder;
- selecting, based on the content vector, a link to a cold start page, the cold start page not satisfying a threshold level of interaction data, the selected link ranked above a second link to a warm page that does satisfy the threshold level of the interaction data;
- and
- presenting the requested page with the selected link.
A method comprising:
- identifying a requested page;
- identifying a content vector of the requested page, the content vector generated based on providing text of the requested page to a neural network text encoder;
- selecting, based on the content vector, a link to a cold start page, the cold start page not satisfying a threshold level of interaction data, the selected link ranked above a second link to a warm page that does satisfy the threshold level of the interaction data;
- and
- displaying the requested page with the selected link.
Dependencies are inferred from the claim text.