Structural information preserving for graph-to-text generation

1. A method of training a graph-to-text generation network, comprising:

receiving encoded graph information corresponding to a target sentence;

decoding the encoded graph information based on a biaffine attention score; and

determining one or more loss values based on the decoded information, wherein a second loss of the one or more loss values is associated with a loss of graph structural information and is used to predict a graph as a linearized sequence, and wherein the loss of graph structural information associated with the second loss is minimized using a depth-first traversal, and wherein the graph-to-text generation network is trained by minimizing the one or more loss values.

2. The method of claim 1 , wherein a first loss value is generated by reconstructing one or more triple relations based on the biaffine attention score.

3. The method of claim 2 , wherein each of the one or more triple relations contains a pair of nodes and a labeled relation associated with the pair of nodes.

4. The method of claim 3 , wherein each of the one or more triple relations are mapped to the target sentence based on one or more pre-generated alignments between one or more nodes of the graph and target words.

5. The method of claim 1 , wherein the linearized graph comprises a sequence of one or more tokens, wherein each token includes one of more of: a graph node, an edge label, and an inserted bracket.

6. A computer system for training a graph-to-text generation network, the computer system comprising:

one or more computer-readable non-transitory storage media configured to store computer program code; and

one or more computer processors configured to access said computer program code and operate as instructed by said computer program code, said computer program code including:

receiving code configured to cause the one or more computer processors to receive encoded graph information corresponding to a target sentence;

decoding code configured to cause the one or more computer processors to decode the encoded graph information based on a biaffine attention score; and

determining code configured to cause the one or more computer processors to determine one or more loss values based on the decoded information, wherein a second loss of the one or more loss values is associated with a loss of graph structural information and is used to predict a graph as a linearized sequence, and wherein the loss of graph structural information associated with the second loss is minimized using a depth-first traversal, and wherein the text-to-graph generation network is trained by minimizing the one or more loss values.

7. The computer system of claim 6 , wherein a first loss value is generated by reconstructing one or more triple relations based on the biaffine attention score.

8. The computer system of claim 7 , wherein each of the one or more triple relations contains a pair of nodes and a labeled relation associated with the pair of nodes.

9. The computer system of claim 8 , wherein each of the one or more triple relations are mapped to the target sentence based on one or more pre-generated alignments between one or more nodes of the graph and target words.

10. The computer system of claim 6 , wherein the linearized graph comprises a sequence of one or more tokens, wherein each token includes one of more of: a graph node, an edge label, and an inserted bracket.

11. A non-transitory computer readable medium having stored thereon a computer program for training a graph-to-text generation network, the computer program configured to cause one or more computer processors to:

receive encoded graph information corresponding to a target sentence;

decode the encoded graph information based on a biaffine attention score; and

determine one or more loss values based on the decoded information, wherein a second loss of the one or more loss values is associated with a loss of graph structural information and is used to predict a graph as a linearized sequence, and wherein the loss of graph structural information associated with the second loss is minimized using a depth-first traversal, and wherein the text-to-graph generation network is trained by minimizing the one or more loss values.

12. The computer readable medium of claim 11 , wherein a first loss value is generated by reconstructing one or more triple relations based on the biaffine attention score.

13. The computer readable medium of claim 12 , wherein each of the one or more triple relations contains a pair of nodes and a labeled relation associated with the pair of nodes.

14. The computer readable medium of claim 13 , wherein each of the one or more triple relations are mapped to the target sentence based on one or more pre-generated alignments between one or more nodes of the graph and target words.

15. The computer readable medium of claim 11 , wherein the linearized graph comprises a sequence of one or more tokens, wherein each token includes one of more of: a graph node, an edge label, and an inserted bracket.