The intelligent automated assistant system engages with the user in an integrated, conversational manner using natural language dialog, and invokes external services when appropriate to obtain information or perform various actions. The system can be implemented using any of a number of different platforms, such as the web, email, smartphone, and the like, or any combination thereof. In one embodiment, the system is based on sets of interrelated domains and tasks, and employs additional functionally powered by external services with which the system can interact.

A method of performing a voice wake-up in multiple speech zones is provided, which relates to a field of artificial intelligence, in particular to fields of speech technology, natural language processing, speech interaction, etc., and may be used in Internet of vehicles, autonomous driving, and other scenarios. A specific implementation scheme includes: acquiring N channels of audio signals, wherein each channel of audio signal corresponds to one of N speech zones; inputting, based on a corresponding relationship between the N channels of audio signals and N synchronous audio processing threads in a wake-up engine, each channel of audio signal into a corresponding audio processing thread; and determining, in response to a thread with a wake-up result occurring in the N synchronous audio processing threads, a speech zone corresponding to the thread with the wake-up result as an awakened speech zone in the N speech zones.

Systems and methods allow users to leverage multiple disparate cloud solutions, offered by disparate service providers, in a unified and cohesive manner. A system includes an engine configured to allocate a task among two or more disparate cloud services according to a running mode. The two or more disparate cloud services include a dedicated solution and a shared solution. The running modes include a dedicated mode configured to direct the tasks to the dedicated solution, a serverless mode configured to direct the tasks to the shared solution, and a hybrid mode configured to direct the tasks to a combination of the dedicated solution and the shared solution.

Systems and methods allow users to leverage multiple disparate cloud solutions, offered by disparate service providers, in a unified and cohesive manner. A system includes an engine configured to allocate a task among two or more disparate cloud services according to a running mode. The two or more disparate cloud services include a dedicated solution and a shared solution. The running modes include a dedicated mode configured to direct the tasks to the dedicated solution, a serverless mode configured to direct the tasks to the shared solution, and a hybrid mode configured to direct the tasks to a combination of the dedicated solution and the shared solution.

A technique for aligning spike timing of models is disclosed. A first model having a first architecture trained with a set of training samples is generated. Each training sample includes an input sequence of observations and an output sequence of symbols having different length from the input sequence. Then, one or more second models are trained with the trained first model by minimizing a guide loss jointly with a normal loss for each second model and a sequence recognition task is performed using the one or more second models. The guide loss evaluates dissimilarity in spike timing between the trained first model and each second model being trained.

A technique for aligning spike timing of models is disclosed. A first model having a first architecture trained with a set of training samples is generated. Each training sample includes an input sequence of observations and an output sequence of symbols having different length from the input sequence. Then, one or more second models are trained with the trained first model by minimizing a guide loss jointly with a normal loss for each second model and a sequence recognition task is performed using the one or more second models. The guide loss evaluates dissimilarity in spike timing between the trained first model and each second model being trained.

Systems and methods herein provide for understanding the context of multiple conversation events and accurately linking them together. Such may allow for fewer financial transaction opportunities to be missed and enable sell side institutions to book more trades. In one embodiment, a method of classifying financial transaction messages with a trained machine learning model includes identifying entities in a financial transaction message, identifying subsequent passages relating to the financial transaction message, and classifying intent as valid or invalid in the financial transaction message. The method also includes linking events within a specific thread by sequentially processing the passages of the financial transaction message.

Systems and methods herein provide for understanding the context of multiple conversation events and accurately linking them together. Such may allow for fewer financial transaction opportunities to be missed and enable sell side institutions to book more trades. In one embodiment, a method of classifying financial transaction messages with a trained machine learning model includes identifying entities in a financial transaction message, identifying subsequent passages relating to the financial transaction message, and classifying intent as valid or invalid in the financial transaction message. The method also includes linking events within a specific thread by sequentially processing the passages of the financial transaction message.

The present disclosure provides a method and a system for interacting with a third-party application. The method includes receiving voice data from a user for launching the third-party application; sending the voice data to a cloud server; receiving the instruction for launching the third-party application from the cloud server; executing the instruction to launch the third-party application; receiving voice data for operating the third-party application from the user after the third-party application is launched; sending the voice data to the cloud server, such that the cloud server performs voice recognition and semantic understanding on the voice data to obtain an instruction for operating the third-party application; receiving the instruction for operating the third-party application sent by the cloud server and forwarding the instruction to the third-party application, such that the third-party application executes the instruction.

The present disclosure provides a method and a system for interacting with a third-party application. The method includes receiving voice data from a user for launching the third-party application; sending the voice data to a cloud server; receiving the instruction for launching the third-party application from the cloud server; executing the instruction to launch the third-party application; receiving voice data for operating the third-party application from the user after the third-party application is launched; sending the voice data to the cloud server, such that the cloud server performs voice recognition and semantic understanding on the voice data to obtain an instruction for operating the third-party application; receiving the instruction for operating the third-party application sent by the cloud server and forwarding the instruction to the third-party application, such that the third-party application executes the instruction.