Abstract: In various examples, a technique for slot filling includes receiving a natural language sentence from a user and identifying a first mention span included in the natural language sentence. The technique also includes determining, using a first machine learning model, that the first mention span is associated with a first slot class included in a set of slot classes based on a set of slot class descriptions corresponding to the set of slot classes.

Abstract: Various embodiments may be generally directed to the use of an adversarial learning framework for persona-based dialogue modeling. In some embodiments, automated multi-turn dialogue response generation may be performed using a persona-based hierarchical recurrent encoder-decoder-based generative adversarial network (phredGAN). Such a phredGAN may feature a persona-based hierarchical recurrent encoder-decoder (PHRED) generator and a conditional discriminator. In some embodiments, the conditional discriminator may include an adversarial discriminator that is provided with attribute representations as inputs. In some other embodiments, the conditional discriminator may include an attribute discriminator, and attribute representations may be handled as targets of the attribute discriminator. The embodiments are not limited in this context.

Abstract: Systems and methods provide for a machine learning system to tokenize, classify, and generate representations from a provided input to provide a combined output. An input may be received and processed into tokens for classification based on semiotic classes. For a given semiotic class, particular rule-based algorithms may be selected to generate a desired output for a selected output language. An input may include an auditory or textual input, which may be in a different language from the selected output language, where the particular rule-based algorithms may include morphological rules for particular semiotic classes. Different rule-based algorithms may be modularly generated for particular languages and semiotic classes to build a library of models for processing different inputs.

Abstract: In various examples, query response generation using entity linking for conversational AI systems and applications is described herein. Systems and methods are disclosed that generate embeddings associated with entities that a dialogue system is trained to interpret. The systems and methods may then use the embeddings to interpret requests. For instance, when receiving a request, the systems and methods may generate at least an embedding for an entity included in the request and compare the embedding to the stored embeddings in order to determine that the entity from the request is related to one of the stored entities. The systems and methods may then use this relationship to generate the response to the query. This way, even if the entity is not an exact match to a stored entity, the systems and methods are still able to interpret the query from the user.

Abstract: Aspects discussed herein may relate to methods and techniques for embedding constrained and unconstrained optimization programs as layers in a neural network architecture. Systems are provided that implement a method of solving a particular optimization problem by a neural network architecture. Prior systems required use of external software to pre-solve optimization programs so that previously determined parameters could be used as fixed input in the neural network architecture. Aspects described herein may transform the structure of common optimization problems/programs into forms suitable for use in a neural network. This transformation may be invertible, allowing the system to learn the solution to the optimization program using gradient descent techniques via backpropagation of errors through the neural network architecture. Thus these optimization layers may be solved via operation of the neural network itself.

Abstract: In a variety of embodiments, machine classifiers may model multi-turn dialogue as a one-to-many prediction task. The machine classifier may be trained using adversarial bootstrapping between a generator and a discriminator with multi-turn capabilities. The machine classifiers may be trained in both auto-regressive and traditional teacher-forcing modes, with the generator including a hierarchical recurrent encoder-decoder network and the discriminator including a bi-directional recurrent neural network. The discriminator input may include a mixture of ground truth labels, the teacher-forcing outputs of the generator, and/or noise data. This mixture of input data may allow for richer feedback on the autoregressive outputs of the generator. The outputs can be ranked based on the discriminator feedback and a response selected from the ranked outputs.

Abstract: Systems and methods for generating responses to user input such as dialogues, and images are discussed. The system may generate, by a response generation module of at least one server, an optimal generated response to the user communication by applying an generative adversarial network. In some embodiments, the generative adversarial network may include a hierarchical recurrent encoder decoder generative adversarial network including a generator and a discriminator component.

Abstract: Machine classifiers in accordance with embodiments of the invention capture long-term temporal dependencies in particular tasks, such as turn-based dialogues. Machine classifiers may be used to help users to perform tasks indicated by the user. When a user utterance is received, natural language processing techniques may be used to understand the user's intent. Templates may be determined based on the user's intent in the generation of responses to solicit information from the user. A variety of persona attributes may be determined for a user. The persona attributes may be determined based on the user's utterances and/or provided as metadata included with the user's utterances. A response persona may be used to generate responses to the user's utterances such that the generated responses match a tone appropriate to the task. A response persona may be used to generate templates to solicit additional information and/or generate responses appropriate to the task.

Abstract: Systems described herein may use transformer-based machine classifiers to perform a variety of natural language understanding tasks including, but not limited to sentence classification, named entity recognition, sentence similarity, and question answering. The exceptional performance of transformer-based language models is due to their ability to capture long-term temporal dependencies in input sequences. Machine classifiers may be trained using training data sets for multiple tasks, such as but not limited to sentence classification tasks and sequence labeling tasks. Loss masking may be employed in the machine classifier to jointly train the machine classifier on multiple tasks simultaneously. The user of transformer encoders in the machine classifiers, which treat each output sequence independently of other output sequences, in accordance with aspects of the invention do not require joint labeling to model tasks.

Abstract: Aspects described herein may allow for the application of stochastic gradient boosting techniques to the training of deep neural networks by disallowing gradient back propagation from examples that are correctly classified by the neural network model while still keeping correctly classified examples in the gradient averaging. Removing the gradient contribution from correctly classified examples may regularize the deep neural network and prevent the model from overfitting. Further aspects described herein may provide for scheduled boosting during the training of the deep neural network model conditioned on a mini-batch accuracy and/or a number of training iterations. The model training process may start un-boosted, using maximum likelihood objectives or another first loss function.

Abstract: In various examples, a technique for slot filling includes receiving a natural language sentence from a user and identifying a first mention span included in the natural language sentence. The technique also includes determining, using a first machine learning model, that the first mention span is associated with a first slot class included in a set of slot classes based on a set of slot class descriptions corresponding to the set of slot classes.

Abstract: In various examples, hybrid models for determining intents in conversational AI systems and applications are disclosed. Systems and methods are disclosed that use a machine learning model(s) and a data file(s) that associates requests (e.g., questions) with responses (e.g., answers) in order to generate final responses to requests. For instance, the machine learning model(s) may determine confidence scores that indicate similarities between the requests from the data file(s) and an input request represented by text data. The data file(s) is then used to determine, based on the confidence scores, one of the responses that is associated with one of the requests that is related to the input request. Additionally, the response may then used to generate a final response to the input request.

Abstract: In various examples, systems and methods are disclosed relating to generating dialogue responses from structured data for conversational artificial intelligence (AI) systems and applications. Systems and methods are disclosed for training or updating a machine learning model—such as a deep neural network—for deployment using structured data from dialogues of multiple domains. The systems and methods can generate responses to users to provide a more natural user experience, such as by generating alternative outputs that vary in syntax with respect to how the outputs incorporate data used to respond to user utterances, while still accurately providing information to satisfy requests from users.

Abstract: Aspects discussed herein may relate to methods and techniques for embedding constrained and unconstrained optimization programs as layers in a neural network architecture. Systems are provided that implement a method of solving a particular optimization problem by a neural network architecture. Prior systems required use of external software to pre-solve optimization programs so that previously determined parameters could be used as fixed input in the neural network architecture. Aspects described herein may transform the structure of common optimization problems/programs into forms suitable for use in a neural network. This transformation may be invertible, allowing the system to learn the solution to the optimization program using gradient descent techniques via backpropagation of errors through the neural network architecture. Thus these optimization layers may be solved via operation of the neural network itself.

Abstract: In various examples, contextual data may be generated using structured and unstructured data for conversational AI systems and applications. Systems and methods are disclosed that use structured data (converted to unstructured form) and unstructured data, such as from a knowledge database(s), to generate contextual data. For instance, the contextual data may represent text (e.g., narratives), where a first portion of the text is generated using the structured data and a second portion of the text is generated using the unstructured data. The systems and methods may then use a neural network(s), such as a neural network(s) associated with a dialogue manager, to process input data representing a request (e.g., a query) and the contextual data in order to generate a response to the request. For instance, if the request includes a query for information associated with a topic, the neural network(s) may generate a response that includes the requested information.

Abstract: Machine classifiers in accordance with embodiments of the invention capture long-term temporal dependencies in the dialogue data better than the existing RNN-based architectures. Additionally, machine classifiers may model the joint distribution of the context and response as opposed to the conditional distribution of the response given the context as employed in sequence-to-sequence frameworks. Machine classifiers in accordance with embodiments further append random paddings before and/or after the input data to reduce the syntactic redundancy in the input data, thereby improving the performance of the machine classifiers for a variety of dialogue-related tasks. The random padding of the input data may further provide regularization during the training of the machine classifier and/or reduce exposure bias. In a variety of embodiments, the input data may be encoded based on subword tokenization.

Abstract: Aspects described herein may allow for the application of stochastic gradient boosting techniques to the training of deep neural networks by disallowing gradient back propagation from examples that are correctly classified by the neural network model while still keeping correctly classified examples in the gradient averaging. Removing the gradient contribution from correctly classified examples may regularize the deep neural network and prevent the model from overfitting. Further aspects described herein may provide for scheduled boosting during the training of the deep neural network model conditioned on a mini-batch accuracy and/or a number of training iterations. The model training process may start un-boosted, using maximum likelihood objectives or another first loss function.

Abstract: In a variety of embodiments, machine classifiers may model multi-turn dialogue as a one-to-many prediction task. The machine classifier may be trained using adversarial bootstrapping between a generator and a discriminator with multi-turn capabilities. The machine classifiers may be trained in both auto-regressive and traditional teacher-forcing modes, with the generator including a hierarchical recurrent encoder-decoder network and the discriminator including a bi-directional recurrent neural network. The discriminator input may include a mixture of ground truth labels, the teacher-forcing outputs of the generator, and/or noise data. This mixture of input data may allow for richer feedback on the autoregressive outputs of the generator. The outputs can be ranked based on the discriminator feedback and a response selected from the ranked outputs.

Abstract: Machine classifiers in accordance with embodiments of the invention capture long-term temporal dependencies in particular tasks, such as turn-based dialogues. Machine classifiers may be used to help users to perform tasks indicated by the user. When a user utterance is received, natural language processing techniques may be used to understand the user's intent. Templates may be determined based on the user's intent in the generation of responses to solicit information from the user. A variety of persona attributes may be determined for a user. The persona attributes may be determined based on the user's utterances and/or provided as metadata included with the user's utterances. A response persona may be used to generate responses to the user's utterances such that the generated responses match a tone appropriate to the task. A response persona may be used to generate templates to solicit additional information and/or generate responses appropriate to the task.

Abstract: Systems described herein may use machine classifiers to perform a variety of natural language understanding tasks including, but not limited to multi-turn dialogue generation. Machine classifiers in accordance with aspects of the disclosure may model multi-turn dialogue as a one-to-many prediction task. The machine classifier may be trained using adversarial bootstrapping between a generator and a discriminator with multi-turn capabilities. The machine classifiers may be trained in both auto-regressive and traditional teacher-forcing modes, with the maximum likelihood loss of the auto-regressive outputs being weighted by the score from a metric-based discriminator model. The discriminators input may include a mixture of ground truth labels, the teacher-forcing outputs of the generator, and/or negative examples from the dataset. This mixture of input may allow for richer feedback on the autoregressive outputs of the generator.