Abstract: The technology disclosed provides a so-called “joint many-task neural network model” to solve a variety of increasingly complex natural language processing (NLP) tasks using growing depth of layers in a single end-to-end model. The model is successively trained by considering linguistic hierarchies, directly connecting word representations to all model layers, explicitly using predictions in lower tasks, and applying a so-called “successive regularization” technique to prevent catastrophic forgetting. Three examples of lower level model layers are part-of-speech (POS) tagging layer, chunking layer, and dependency parsing layer. Two examples of higher level model layers are semantic relatedness layer and textual entailment layer. The model achieves the state-of-the-art results on chunking, dependency parsing, semantic relatedness and textual entailment.

Abstract: Embodiments described herein provide a pipelined natural language question answering system that improves a BERT-based system. Specifically, the natural language question answering system uses a pipeline of neural networks each trained to perform a particular task. The context selection network identifies premium context from context for the question. The question type network identifies the natural language question as a yes, no, or span question and a yes or no answer to the natural language question when the question is a yes or no question. The span extraction model determines an answer span to the natural language question when the question is a span question.

Abstract: A multi-tenant system performs custom configuration of a tenant-specific chatbot to process and act upon natural language requests. The multi-tenant system configures the tenant-specific chatbots without requiring tenant-specific training. The multi-tenant system providing a user interface for configuring a tenant-specific set of permitted actions. The multi-tenant system determines a set of example phrases for each of the selected permitted actions. The multi-tenant system receives a natural language request from a user and identifies the action that the user wants to perform. The multi-tenant system uses a neural network to compare the natural language request with example phrases to identify an example phrase that matches the natural language request. The multi-tenant system performs the action corresponding to the matching example phrase.

Abstract: An online system that allows users to interact with it using expressions in natural language form includes an intent inference module allowing it to infer an intent represented by a user expression. The intent inference module has a set of possible intents, along with a small set of example natural language expressions known to represent that intent. When a user interacts with the system using a natural language expression for which the intent is not already known, the intent inference module applies a natural language inference model to compute scores indicating whether the user expression textually entails the various example natural language expressions. Based on the scores, the intent inference module determines an intent that is most applicable for the expression. If an intent cannot be determined with sufficient confidence, the intent inference module may further attempt to determine whether the various example natural language expressions textually entail the user expression.

Abstract: Embodiments described herein provide methods and systems for training task-oriented dialogue (TOD) language models. In some embodiments, a TOD language model may receive a TOD dataset including a plurality of dialogues and a model input sequence may be generated from the dialogues using a first token prefixed to each user utterance and a second token prefixed to each system response of the dialogues. In some embodiments, the first token or the second token may be randomly replaced with a mask token to generate a masked training sequence and a masked language modeling (MLM) loss may be computed using the masked training sequence. In some embodiments, the TOD language model may be updated based on the MLM loss.

Abstract: Approaches to zero-shot learning include partitioning training data into first and second sets according to classes assigned to the training data, training a prediction module based on the first set to predict a cluster center based on a class label, training a correction module based on the second set and each of the class labels in the first set to generate a correction to a cluster center predicted by the prediction module, presenting a new class label for a new class to the prediction module to predict a new cluster center, presenting the new class label, the predicted new cluster center, and each of the class labels in the first set to the correction module to generate a correction for the predicted new cluster center, augmenting a classifier based on the corrected cluster center for the new class, and classifying input data into the new class using the classifier.

Abstract: Embodiments described herein provide methods and systems for training a sequential recommendation model. A system receives a plurality of user behavior sequences, and encodes those sequences into a plurality of user interest representations. The system predicts a next item using a sequential recommendation model, producing a probability distribution over a set of items. The next interacted item in a sequence is selected as a positive sample, and a negative sample is selected based on the generated probability distribution. The positive and negative samples are used to compute a contrastive loss and update the sequential recommendation model.

Abstract: A text-to-database neural network architecture is provided. The architecture receives a natural language question and a database schema and generates a serialized question-schema representation that includes a question and at least one table and at least one field from the database schema. The serialized question-schema representation is appended with at least one value that matches a word in the natural language question and at least one field in a database picklist. An encoder in the architecture generates question and schema encodings from the appended question-schema representation. Schema encodings are associated with metadata that indicates a data type of the fields and whether fields are associated with primary or foreign keys. A decoder in the architecture generates an executable query from the question encodings and schema encodings.

Abstract: Computing systems may support image classification and image detection services, and these services may utilize object detection/image classification machine learning models. The described techniques provide for normalization of confidence scores corresponding to manipulated target images and for non-max suppression within the range of confidence scores for manipulated images. In one example, the techniques provide for generating different scales of a test image, and the system performs normalization of confidence scores corresponding to each scaled image and non-max suppression per scaled image These techniques may be used to provide more accurate image detection (e.g., object detection and/or image classification) and may be used with models that are not trained on modified image sets. The model may be trained on a standard (e.g. non-manipulated) image set but used with manipulated target images and the described techniques to provide accurate object detection.

Abstract: Embodiments described herein provide methods and systems for presenting a document and generating a human-AI summary. A system provides a user with a selection of an amount of time to spend reading the document, or a list of questions from which the user may select which questions they would like answered by reading the document. The system highlights sections of the document according to the user selection. Implicit and explicit user data such as dwell times, user highlights, and user notes, are collected while displaying the document. A human-AI summary is generated based on the document and the user data.

Abstract: Embodiments described herein provide a method of evaluating a natural language processing model. The method includes receiving an evaluation dataset that may include a plurality of unit tests, the unit tests having: an input context, and a first candidate and a second candidate that are generated in response to the input context, where the first test candidate is associated with a first quality notation, and the second candidate is associated with a second quality notation. The method includes determining, via a model, a first likelihood of generating the first candidate and a second likelihood of generating the second candidate in response to the input context. The method also includes determining whether the first likelihood being greater than the second likelihood. The method also includes determining whether the first model passed the unit test, where the first quality notation indicates a higher quality candidate and the second quality notation indicate a lower quality candidate.

Abstract: Embodiments described herein provide methods and systems for presenting a document and generating a human-AI summary. A system provides a user with a selection of an amount of time to spend reading the document, or a list of questions from which the user may select which questions they would like answered by reading the document. The system highlights sections of the document according to the user selection. Implicit and explicit user data such as dwell times, user highlights, and user notes, are collected while displaying the document. A human-AI summary is generated based on the document and the user data.

Abstract: An online system extracts information from non-fixed form documents. The online system receives an image of a form document and obtains a set of phrases and locations of the set of phrases on the form image. For at least one field, the online system determines key scores for the set of phrases. The online system identifies a set of candidate values for the field from the set of identified phrases and identifies a set of neighbors for each candidate value from the set of identified phrases. The online system determines neighbor scores, where a neighbor score for a candidate value and a respective neighbor is determined based on the key score for the neighbor and a spatial relationship of the neighbor to the candidate value. The online system selects a candidate value and a respective neighbor based on the neighbor score as the value and key for the field.

Abstract: Systems and methods are provided for weakly supervised natural language localization (WSNLL), for example, as implemented in a neural network or model. The WSNLL network is trained with long, untrimmed videos, i.e., videos that have not been temporally segmented or annotated. The WSNLL network or model defines or generates a video-sentence pair, which corresponds to a pairing of an untrimmed video with an input text sentence. According to some embodiments, the WSNLL network or model is implemented with a two-branch architecture, where one branch performs segment sentence alignment and the other one conducts segment selection. These methods and systems are specifically used to predict how a video proposal matches a text query using respective visual and text features.

Abstract: A conversation engine performs conversations with users using chatbots customized for performing a set of tasks that can be performed using an online system. The conversation engine loads a chatbot configuration that specifies the behavior of a chatbot including the tasks that can be performed by the chatbot, the types of entities relevant to each task, and so on. The conversation may be voice based and use natural language. The conversation engine may load different chatbot configurations to implement different chatbots. The conversation engine receives a conversation engine configuration that specifies the behavior of the conversation engine across chatbots. The system may be a multi-tenant system that allows customization of the chatbots for each tenant.

Abstract: A method for training parameters of a first domain adaptation model. The method includes evaluating a cycle consistency objective using a first task specific model associated with a first domain and a second task specific model associated with a second domain, and evaluating one or more first discriminator models to generate a first discriminator objective using the second task specific model. The one or more first discriminator models include a plurality of discriminators corresponding to a plurality of bands that corresponds domain variable ranges of the first and second domains respectively. The method further includes updating, based on the cycle consistency objective and the first discriminator objective, one or more parameters of the first domain adaptation model for adapting representations from the first domain to the second domain.

Abstract: A method for generating a neural network for detecting one or more objects in images includes generating one or more self-supervised proposal learning losses based on the one or more proposal features and corresponding proposal feature predictions. One or more consistency-based proposal learning losses are generated based on noisy proposal feature predictions and the corresponding proposal predictions without noise. A combined loss is generated using the one or more self-supervised proposal learning losses and one or more consistency-based proposal learning losses. The neural network is updated based on the combined loss.

Abstract: A method for evaluating robustness of one or more target neural network models using natural typos. The method includes receiving one or more natural typo generation rules associated with a first task associated with a first input document type, receiving a first target neural network model, and receiving a first document and corresponding its ground truth labels. The method further includes generating one or more natural typos for the first document based on the one or more natural typo generation rules, and providing, to the first target neural network model, a test document generated based on the first document and the one or more natural typos as an input document to generate a first output. A robustness evaluation result of the first target neural network model is generated based on a comparison between the output and the ground truth labels.

Abstract: Embodiments described herein provide a composed variational natural language generation (CLANG) model that is configured to generate training samples for few-shot intents. Specifically, the CLANG model may build connections between existing training samples of many-shot intents and new training samples of few-shot intents by modeling an intent as a combination of a domain and an action. In this way, the CLANG model transfers knowledge from existing many-shot intents to few-shot intents in natural language generation by learning how to compose utterances with many-shot intents and transferring such knowledge to few-shot intents.

Abstract: Systems and methods for unifying question answering and text classification via span extraction include a preprocessor for preparing a source text and an auxiliary text based on a task type of a natural language processing task, an encoder for receiving the source text and the auxiliary text from the preprocessor and generating an encoded representation of a combination of the source text and the auxiliary text, and a span-extractive decoder for receiving the encoded representation and identifying a span of text within the source text that is a result of the NLP task. The task type is one of entailment, classification, or regression. In some embodiments, the source text includes one or more of text received as input when the task type is entailment, a list of classifications when the task type is entailment or classification, or a list of similarity options when the task type is regression.