Abstract: Techniques for evaluating a natural language understanding (NLU) component and determining an action to resolve an issue processing a user input are described. The system determines which component is invoked by a baseline NLU component is processing the user input, and which component is invoked by an updated NLU component. Based on that information, the system selects the action to resolve the updated NLU component generating an undesired response to the user input.

Abstract: Techniques for using a federated learning framework to update machine learning models for spoken language understanding (SLU) system are described. The system determines which labeled data is needed to update the models based on the models generating an undesired response to an input. The system identifies users to solicit labeled data from, and sends a request to a user device to speak an input. The device generates labeled data using the spoken input, and updates the on-device models using the spoken input and the labeled data. The updated model data is provided to the system to enable the system to update the system-level (global) models.

Abstract: Techniques for evaluating a natural language understanding (NLU) component and determining an action to resolve an issue processing a user input are described. The system determines which component is invoked by a baseline NLU component is processing the user input, and which component is invoked by an updated NLU component. Based on that information, the system selects the action to resolve the updated NLU component generating an undesired response to the user input.

Abstract: A system and method are disclosed which enable more effective email response authoring by contact center agents, for example, by automatically suggesting prototypical (entire) email responses to the human agent and interactive suggestion of next sentence candidates during the writing process. In one method, a customer inquiry is received and a latent topic prediction is generated, based on a word-based representation of the customer inquiry. A latent topic prediction is generated for an entire agent's reply to the customer inquiry as a function of the latent topic prediction generated for the customer inquiry. A further latent topic prediction is generated for a next sentence of the agent's reply as a function of a topic prediction for the next sentence which is generated with a prediction model that has been trained on annotated sentences of agent replies. Information is output to assist the agent, based on the topic predictions.

Abstract: A system and method for estimating parameters for features of a translation scoring function for scoring candidate translations in a target domain are provided. Given a source language corpus for a target domain, a similarity measure is computed between the source corpus and a target domain multi-model, which may be a phrase table derived from phrase tables of comparative domains, weighted as a function of similarity with the source corpus. The parameters of the log-linear function for these comparative domains are known. A mapping function is learned between similarity measure and parameters of the scoring function for the comparative domains. Given the mapping function and the target corpus similarity measure, the parameters of the translation scoring function for the target domain are estimated. For parameters where a mapping function with a threshold correlation is not found, another method for obtaining the target domain parameter can be used.

Abstract: A method for generating a phrase table for a target domain includes receiving a source corpus for a target domain and, for each of a set of comparative domain phrase tables, computing a measure of similarity between the source corpus and the comparative domain phrase table. Based on the computed similarity measures, a subset of the comparative domain phrase tables may be identified from the set of comparative domain phrase tables, and/or weights for combining them, and a phrase table is generated for the target domain based on the at least a subset of phrase tables.

Abstract: A system and method permit analysis and generation to be performed with the same reversible probabilistic model. The model includes a set of factors, including a canonical factor, which is a function of a logical form and a realization thereof, a similarity factor, which is a function of a canonical text string and a surface string, a language model factor, which is a static function of a surface string, a language context factor, which is a dynamic function of a surface string, and a semantic context factor, which is a dynamic function of a logical form. When performing generation, the canonical factor, similarity factor, language model factor, and language context factor are composed to receive as input a logical form and output a surface string, and when performing analysis, the similarity factor, canonical factor, and semantic context factor are composed to take as input a surface string and output a logical form.

Abstract: A system and method are disclosed which enable more effective email response authoring by contact center agents, for example, by automatically suggesting prototypical (entire) email responses to the human agent and interactive suggestion of next sentence candidates during the writing process. In one method, a customer inquiry is received and a latent topic prediction is generated, based on a word-based representation of the customer inquiry. A latent topic prediction is generated for an entire agent's reply to the customer inquiry as a function of the latent topic prediction generated for the customer inquiry. A further latent topic prediction is generated for a next sentence of the agent's reply as a function of a topic prediction for the next sentence which is generated with a prediction model that has been trained on annotated sentences of agent replies. Information is output to assist the agent, based on the topic predictions.

Abstract: Agent utterances are generated for implementing dialog acts recommended by a dialog manager of a call center. To this end, a set of word lattices, each represented as a weighted finite state automaton (WFSA), is constructed from training dialogs between call center agents and second parties (e.g. customers). The word lattices are assigned conditional probabilities over dialog act type. For each dialog act received from the dialog manager, the word lattices are ranked by the conditional probabilities for the dialog act type. At least one word lattice is chosen from the ranking, and is instantiated to generate a recommended agent utterance for implementing the recommended dialog act. The word lattices may be constructed by clustering agent utterances of training dialogs using context features from preceding second party utterances and grammatical dependency link features between words within agent utterances. Path variations of the word lattices may define slots or paraphrases.

Abstract: Methods and systems for enriching translation models. The first strength metric associated with a phrase in a first translation model is determined. The second strength metric associated with the phrase is received from at least one second translation model. The first translation model is enriched based on one or more translations of the phrase received from the at least one second translation model. The one or more translations are received based on a comparison between the first strength metric and the second strength metric.

Abstract: The disclosed embodiments relate to a system and method for predicting the learning curve of an SMT system. A set of anchor points are selected. The set of anchor points correspond to a size of a corpus. Thereafter, a gold curve or a benchmark curve is fitted based on the set of anchor points to determine the BLEU score. Based on the BLEU score and a set of parameters associated with the first set of anchor points, a confidence score is computed.

Abstract: A method for generating a phrase table for a target domain includes receiving a source corpus for a target domain and, for each of a set of comparative domain phrase tables, computing a measure of similarity between the source corpus and the comparative domain phrase table. Based on the computed similarity measures, a subset of the comparative domain phrase tables may be identified from the set of comparative domain phrase tables, and/or weights for combining them, and a phrase table is generated for the target domain based on the at least a subset of phrase tables.

Abstract: A system and method for estimating parameters for features of a translation scoring function for scoring candidate translations in a target domain are provided. Given a source language corpus for a target domain, a similarity measure is computed between the source corpus and a target domain multi-model, which may be a phrase table derived from phrase tables of comparative domains, weighted as a function of similarity with the source corpus. The parameters of the log-linear function for these comparative domains are known. A mapping function is learned between similarity measure and parameters of the scoring function for the comparative domains. Given the mapping function and the target corpus similarity measure, the parameters of the translation scoring function for the target domain are estimated. For parameters where a mapping function with a threshold correlation is not found, another method for obtaining the target domain parameter can be used.

Abstract: An authoring method includes generating an authoring interface configured for assisting a user to author a text string in a source language for translation to a target string in a target language. Initial source text entered by the user is received through the authoring interface. Source phrases are selected that each include at least one token of the initial source text as a prefix and at least one other token as a suffix. The source phrase selection is based on a translatability score and optionally on fluency and semantic relatedness scores. A set of candidate phrases is proposed for display on the authoring interface, each of the candidate phases being the suffix of a respective one of the selected source phrases. The user may select one of the candidate phrases, which is appended to the source text following its corresponding prefix, or may enter alternative text. The process may be repeated until the user is satisfied with the source text and the SMT model can then be used for its translation.

Abstract: A method, a system, and a computer program product for processing the output of an OCR are disclosed. The system receives a first character sequence from the OCR. A first set of characters from the first character sequence are converted to a corresponding second set of characters to generate a second character sequence based on a look-up table and language scores.

Abstract: Rejection sampling is performed to acquire at least one target language translation for a source language string s in accordance with a phrase-based statistical translation model p(x)=p(t, a|s) where t is a candidate translation, a is a candidate alignment comprising a biphrase sequence generating the candidate translation t, and x is a sequence representing the candidate alignment a. The rejection sampling uses a proposal distribution comprising a weighted finite state automaton (WFSA) q(n) that is refined responsive to rejection of a sample x* obtained in a current iteration of the rejection sampling to generate a refined WFSA q(n+1) for use in a next iteration of the rejection sampling. The refined WFSA q(n+1) is selected to satisfy the criteria p(x)?q(n+1)(x)?q(n)(x) for all x?X and q(n+1)(x*)<q(n)(x*) where the space X is the set of sequences x corresponding to candidate alignments a that generate candidate translations t for the source language string s.

Abstract: A method for rewriting source text includes receiving source text including a source text string in a first natural language. The source text string is translated with a machine translation system to generate a first target text string in a second natural language. A translation confidence for the source text string is computed, based on the first target text string. At least one alternative text string is generated, where possible, in the first natural language by automatically rewriting the source string. Each alternative string is translated to generate a second target text string in the second natural language. A translation confidence is computed for the alternative text string based on the second target string. Based on the computed translation confidences, one of the alternative text strings may be selected as a candidate replacement for the source text string and may be proposed to a user on a graphical user interface.

Abstract: Rejection sampling is performed to acquire at least one target language translation for a source language string s in accordance with a phrase-based statistical translation model p(x)=p(t, a|s) where t is a candidate translation, a is a candidate alignment comprising a biphrase sequence generating the candidate translation t, and x is a sequence representing the candidate alignment a. The rejection sampling uses a proposal distribution comprising a weighted finite state automaton (WFSA) q(n) that is refined responsive to rejection of a sample x* obtained in a current iteration of the rejection sampling to generate a refined WFSA q(n+1) for use in a next iteration of the rejection sampling. The refined WFSA q(n+1) is selected to satisfy the criteria p(x)?q(n+1)(x)?q(n)(x) for all x?X and q(n+1)(x*)<q(n)(x*) where the space X is the set of sequences x corresponding to candidate alignments a that generate candidate translations t for the source language string s.

Abstract: An authoring method includes generating an authoring interface configured for assisting a user to author a text string in a source language for translation to a target string in a target language. Initial source text entered by the user is received through the authoring interface. Source phrases are selected that each include at least one token of the initial source text as a prefix and at least one other token as a suffix. The source phrase selection is based on a translatability score and optionally on fluency and semantic relatedness scores. A set of candidate phrases is proposed for display on the authoring interface, each of the candidate phases being the suffix of a respective one of the selected source phrases. The user may select one of the candidate phrases, which is appended to the source text following its corresponding prefix, or may enter alternative text. The process may be repeated until the user is satisfied with the source text and the SMT model can then be used for its translation.

Abstract: The disclosed embodiments relate to a system and method for predicting the learning curve of an SMT system. A set of anchor points are selected. The set of anchor points correspond to a size of a corpus. Thereafter, a gold curve or a benchmark curve is fitted based on the set of anchor points to determine the BLEU score. Based on the BLEU score and a set of parameters associated with the first set of anchor points, a confidence score is computed.