Digital document update
Techniques and systems are described in which a document management system is configured to update content of document portions of digital documents. In one example, an update to the digital document is initially triggered by a document management system by detecting a triggering change applied to an initial portion of the digital document. The document management system, in response to the triggering change, then determines whether trailing changes are to be made to other document portions, such as to other document portions in the same digital document or another digital document. To do so, triggering and trailing change representations are generated and compared to determine similarity of candidate document portions with an initial document portion.
1. In a digital medium document update environment, a method implemented by a computing device, the method comprising:
detecting, by the computing device, an input specifying a triggering change to an initial document portion of a digital document;
generating, by the computing device, a triggering change representation of the initial document portion;
generating, by the computing device, a plurality of candidate trailing change representations from respective ones of a plurality of candidate document portions;
generating, by the computing device, a plurality of similarity scores by comparing the triggering change representation to the respective ones of the plurality of candidate trailing change representations;
ranking, by the computing device, the plurality of similarity scores; and
outputting, by the computing device, the ranked plurality of similarity scores to control application of trailing changes to respective ones of the plurality of candidate document portions that are based on the triggering change to the initial document portion.
2. The method as described in claim 1 , wherein the generating of the triggering change representation is based on the initial document portion before the triggering change is applied.
3. The method as described in claim 1 , wherein the generating of the triggering change representation includes:
generating word representations from the initial document portion; and
generating the triggering change representation of the initial document portion by aggregating the generated word representations.
4. The method as described in claim 3 , wherein the aggregating includes concatenating the generated word representations, averaging the generated word representations, or employing machine learning including a recurrent neural network or long short-term memory network.
5. The method as described in claim 1 , wherein the generating of the plurality of candidate trailing change representations from respective ones of a plurality of candidate document portions includes for a respective said candidate trailing change representation:
generating word representations from a respective said candidate document portion of the plurality of candidate document portions; and
generating the respective said triggering change representation of the respective said candidate document portion by aggregating the generated word representations.
6. The method as described in claim 5 , wherein the aggregating includes concatenating the generated word representations, averaging the generated word representations, or employing machine learning including a recurrent neural network or long short-term memory network.
7. The method as described in claim 1 , wherein the generating of the plurality of similarity scores is based on Euclidean distance or cosine similarity of the triggering change representation to the respective ones of the plurality of candidate trailing change representations.
8. The method as described in claim 1 , wherein the triggering change involves receipt of a user input to manually specify the triggering change to the initial document portion.
9. The method as described in claim 1 , wherein the triggering change includes detecting a change, automatically and without user intervention, to a document repository outside of the digital document, the document repository identified as having subject matter that corresponds to subject matter of the initial document portion of the digital document.
10. The method as described in claim 1 , wherein at least one of the plurality of candidate document portions is taken from another digital document that is different than the digital document.
11. The method as described in claim 1 , wherein the triggering change involves receipt of a user input to manually specify the triggering change to the initial document portion.
12. The method as described in claim 1 , further comprising applying at least one said trailing change to a respective said candidate document portion based on the ranking automatically and without user intervention using natural language processing.
13. The method as described in claim 1 , wherein the outputting includes at least one option that is user selectable via a user interface to cause application of least one said trailing change to a respective said candidate document portion.
14. In a digital medium document update environment, a system comprising:
a triggering change module implemented at least partially in hardware of a computing device to perform operations including:
detecting an input specifying a triggering change to an initial document portion of a digital document; and
applying the triggering change to the initial document portion; and
a trailing change module implemented at least partially in hardware of the computing device to perform operations comprising:
generating a triggering change representation of the initial document portion;
generating a plurality of candidate trailing change representations from respective ones of a plurality of candidate document portions;
generating a plurality of similarity scores by comparing the triggering change representation to the respective ones of the plurality of candidate trailing change representations; and
outputting an indication of a triggering change to be made to a respective one of the plurality of candidate trailing change representations based on the generated plurality of similarity scores.
15. The system as described in claim 14 , wherein the generating of the triggering change representation includes:
generating word representations from the initial document portion; and
generating the triggering change representation of the initial document portion by aggregating the generated word representations.
16. The system as described in claim 15 , wherein the aggregating includes concatenating the generated word representations, averaging the generated word representations, or employing machine learning including a recurrent neural network or long short-term memory network.
17. The system as described in claim 15 , wherein the aggregating includes concatenating the generated word representations, averaging the generated word representations, or employing machine learning including a recurrent neural network or long short-term memory network.
18. The system as described in claim 14 , wherein the generating of the plurality of candidate trailing change representations from respective ones of a plurality of candidate document portions includes for a respective said candidate trailing change representation:
generating word representations from a respective said candidate document portion of the plurality of candidate document portions; and
generating the respective said triggering change representation of the respective said candidate document portion by aggregating the generated word representations.
19. In a digital medium document update environment, a system comprising:
means for detecting an input specifying a triggering change to an initial document portion of a digital document;
means for generating a triggering change representation of the initial document portion;
means for generating a plurality of candidate trailing change representations from respective ones of a plurality of candidate document portions;
means for generating a plurality of similarity scores by comparing the triggering change representation to the respective ones of the plurality of candidate trailing change representations; and
means for outputting the plurality of similarity scores in a user interface to control application of trailing changes to respective ones of the plurality of candidate document portions that are based on the triggering change to the initial document portion.
20. The system as described in claim 19 , wherein the outputting means includes at least one option that is user selectable via a user interface to cause application of least one said trailing change to a respective said candidate document portion.