Method of and system for translating speech to text
AssigneeYandex Europe AG
InventorPavel Aleksandrovich ZELENKO
Life of this patent
Estimating the term…
Abstract
There is disclosed a method and system for translating speech to text, the speech having been received by a client device. A user utterance corresponding to the speech is received. A first predicted text corresponding to the user utterance and a first confidence score corresponding to the first predicted text are determined using a local graph. The user utterance is transmitted to a server. A second predicted text corresponding to the user utterance and a second confidence score corresponding to the second predicted text is received from the server. If the first confidence score is higher than the second confidence score, the first predicted text is output.
Claims20
A method for translating speech to text, the speech having been received by a client device, the method executable on the client device, the method comprising:
- receiving a user utterance corresponding to the speech;
- determining, based on a local graph stored by the client device, a first predicted text corresponding to the user utterance and a first confidence score corresponding to the first predicted text, wherein the local graph comprises a tree, wherein each leaf of the tree corresponds to a word, and wherein the local graph is generated based on text stored by the client device;
- transmitting, to a server, the user utterance;
- receiving, from the server, a second predicted text corresponding to the user utterance and a second confidence score corresponding to the second predicted text;
- comparing the first confidence score to the second confidence score;
- and
- after determining that the first confidence score is greater than the second confidence score, outputting the first predicted text.
A method for translating speech to text, the speech having been received by a client device, the method executable on a server, the method comprising:
- receiving, from the client device, a user utterance corresponding to the speech;
- determining, based on a graph, a first predicted text corresponding to the user utterance and a first confidence score corresponding to the first predicted text, wherein the graph comprises a tree, and wherein each leaf of the tree corresponds to a word;
- receiving, from the client device, a second predicted text corresponding to the user utterance and a second confidence score corresponding to the second predicted text, wherein the second predicted text was determined by the client device, wherein the second predicted text was determined based on a local graph stored by the client device, and wherein the local graph was generated based on text stored by the client device;
- comparing the first confidence score to the second confidence score;
- and
- after determining that the first confidence score is greater than the second confidence score, transmitting, to the client device, the first predicted text.
A method for translating speech to text, the speech having been received by a client device, the method executable on a server, the method comprising:
- receiving, from the client device, a user utterance corresponding to the speech;
- determining, based on a graph, a first predicted text corresponding to the user utterance and a first confidence score corresponding to the first predicted text, wherein the graph comprises a tree, and wherein each leaf of the tree corresponds to a word;
- and
- transmitting, to the client device, the first predicted text and the first confidence score for comparing the first predicted text and the first confidence score to a second predicted text and a second confidence score, wherein the second predicted text and the second confidence score was determined by the client device based on a local graph stored by the client device, and wherein the local graph was generated based on text stored by the client device.
A method for translating speech to text, the speech having been received by a client device, the method executable on the client device, the method comprising:
- receiving a user utterance corresponding to the speech;
- determining, based on a local graph stored by the client device, a predicted text corresponding to the user utterance and a confidence score corresponding to the predicted text, wherein the local graph comprises a tree, wherein each leaf of the tree corresponds to a word, and wherein the local graph is generated based on text stored by the client device;
- comparing the confidence score to a threshold confidence score;
- determining, based on a threshold confidence score, whether to transmit the user utterance to a server;
- after determining that the confidence score satisfies the threshold confidence score, determining not to transmit the user utterance to the server;
- and
- outputting the predicted text.
Dependencies are inferred from the claim text.