Microsoft Technology Licensing, LLC

Daniel Boies

Ruhi Sarikaya

Larry Paul Heck

Tasos Anastasakos

Fethiye Asli Celikyilmaz

Anoop Deoras

Zhaleh Feizollahi

Dilek Zeynep Hakkani-Tur

Gokhan Tur

Fileno A Alleva

Jean-Philippe Robichaud

Alexandre Rochette

Paul Anthony Crook

Puyang Xu

Nikhil Holenarsipur Ramesh

Dustin Hillard

Ravikiran Janardhana

Tools are provided to allow developers to enable applications for Conversational Understanding (CU) using assets from a CU service. The tools may be used to select functionality from existing domains, extend the coverage of one or more domains, as well as to create new domains in the CU service. A developer may provide example Natural Language (NL) sentences that are analyzed by the tools to assist the developer in labeling data that is used to update the models in the CU service. For example, the tools may assist a developer in identifying domains, determining intent actions, determining intent objects and determining slots from example NL sentences. After the developer tags all or a portion of the example NL sentences, the models in the CU service are automatically updated and validated. For example, validation tools may be used to determine an accuracy of the model against test data.

Building conversational understanding systems using a toolset

Environmental conditions, along with other information, are used to adjust a response of a conversational dialog system. The environmental conditions may be used at different times within the conversational dialog system. For example, the environmental conditions can be used to adjust the dialog manager's output (e.g., the machine action). The dialog state information that is used by the dialog manager includes environmental conditions for the current turn in the dialog as well as environmental conditions for one or more past turns in the dialog. The environmental conditions can also be used after receiving the machine action to adjust the response that is provided to the user. For example, the environmental conditions may affect the machine action that is determined as well as how the action is provided to the user. The dialog manager and the response generation components in the conversational dialog system each use the available environmental conditions.

Environmentally aware dialog policies and response generation

In language evaluation systems, user expressions are often evaluated by speech recognizers and language parsers, and among several possible translations, a highest-probability translation is selected and added to a dialog sequence. However, such systems may exhibit inadequacies by discarding alternative translations that may initially exhibit a lower probability, but that may have a higher probability when evaluated in the full context of the dialog, including subsequent expressions. Presented herein are techniques for communicating with a user by formulating a dialog hypothesis set identifying hypothesis probabilities for a set of dialog hypotheses, using generative and/or discriminative models, and repeatedly re-ranks the dialog hypotheses based on subsequent expressions. Additionally, knowledge sources may inform a model-based with a pre-knowledge fetch that facilitates pruning of the hypothesis search space at an early stage, thereby enhancing the accuracy of language parsing while also reducing the latency of the expression evaluation and economizing computing resources.

Dialogue evaluation via multiple hypothesis ranking

Systems and methods for determining a user intent or goal for contextual language understanding by utilizing information from one or more previous user natural language inputs and one or more previous system generated responses to the user natural language inputs are provided. More specifically, the systems and methods utilize a common schema for determining features from the responses and natural language inputs and provide carryover tracking between responses and the natural language inputs. Accordingly, the systems and methods for contextual language understanding provide for a more accurate, a more reliable, and a more efficient context carryover and goal tracking system when compared to systems and methods that do not utilized the responses in determining the user goal/intent.

Context carryover in language understanding systems or methods

Embodiments provide for tracking a partial dialog state as part of managing a dialog state space, but the embodiments are not so limited. A method of an embodiment jointly models partial state update and named entity recognition using a sequence-based classification or other model, wherein recognition of named entities and a partial state update can be performed in a single processing stage at runtime to generate a distribution over partial dialog states. A system of an embodiment is configured to generate a distribution over partial dialog states at runtime in part using a sequence classification decoding or other algorithm to generate one or more partial dialog state hypothesis and/or a confidence score or measure associated with each hypothesis. Other embodiments are included.

Techniques for updating a partial dialog state

A scalable statistical language understanding (SLU) system uses a fixed number of understanding models that scale across domains and intents (i.e. single vs. multiple intents per utterance). For each domain added to the SLU system, the fixed number of existing models is updated to reflect the newly added domain. Information that is already included in the existing models and the corresponding training data may be re-used. The fixed models may include a domain detector model, an intent action detector model, an intent object detector model and a slot/entity tagging model. A domain detector identifies different domains identified within an utterance. All/portion of the detected domains are used to determine associated intent actions. For each determined intent action, one or more intent objects are identified. Slot/entity tagging is performed using the determined domains, intent actions, and intent object detector.

Scaling statistical language understanding systems across domains and intents

Processes capable of accepting linguistic input in one or more languages are generated by re-using existing linguistic components associated with a different anchor language, together with machine translation components that translate between the anchor language and the one or more languages. Linguistic input is directed to machine translation components that translate such input from its language into the anchor language. Those existing linguistic components are then utilized to initiate responsive processing and generate output. Optionally, the output is directed through the machine translation components. A language identifier can initially receive linguistic input and identify the language within which such linguistic input is provided to select an appropriate machine translation component. A hybrid process, comprising machine translation components and linguistic components associated with the anchor language, can also serve as an initiating construct from which a single language process is created over time.

Growing community of inventors

St. Lambert, Canada

Daniel Boies