Abstract: This document generally describes systems and methods for dynamically adapting speech recognition for individual voice queries of a user using class-based language models. The method may include receiving a voice query from a user that includes audio data corresponding to an utterance of the user, and context data associated with the user. One or more class models are then generated that collectively identify a first set of terms determined based on the context data, and a respective class to which the respective term is assigned for each respective term in the first set of terms. A language model that includes a residual unigram may then be accessed and processed for each respective class to insert a respective class symbol at each instance of the residual unigram that occurs within the language model. A transcription of the utterance of the user is then generated using the modified language model.

Abstract: This document generally describes systems and methods for dynamically adapting speech recognition for individual voice queries of a user using class-based language models. The method may include receiving a voice query from a user that includes audio data corresponding to an utterance of the user, and context data associated with the user. One or more class models are then generated that collectively identify a first set of terms determined based on the context data, and a respective class to which the respective term is assigned for each respective term in the first set of terms. A language model that includes a residual unigram may then be accessed and processed for each respective class to insert a respective class symbol at each instance of the residual unigram that occurs within the language model. A transcription of the utterance of the user is then generated using the modified language model.

Abstract: This document generally describes systems and methods for dynamically adapting speech recognition for individual voice queries of a user using class-based language models. The method may include receiving a voice query from a user that includes audio data corresponding to an utterance of the user, and context data associated with the user. One or more class models are then generated that collectively identify a first set of terms determined based on the context data, and a respective class to which the respective term is assigned for each respective term in the first set of terms. A language model that includes a residual unigram may then be accessed and processed for each respective class to insert a respective class symbol at each instance of the residual unigram that occurs within the language model. A transcription of the utterance of the user is then generated using the modified language model.

Abstract: This document generally describes systems and methods for dynamically adapting speech recognition for individual voice queries of a user using class-based language models. The method may include receiving a voice query from a user that includes audio data corresponding to an utterance of the user, and context data associated with the user. One or more class models are then generated that collectively identify a first set of terms determined based on the context data, and a respective class to which the respective term is assigned for each respective term in the first set of terms. A language model that includes a residual unigram may then be accessed and processed for each respective class to insert a respective class symbol at each instance of the residual unigram that occurs within the language model. A transcription of the utterance of the user is then generated using the modified language model.

Abstract: This document generally describes systems and methods for dynamically adapting speech recognition for individual voice queries of a user using class-based language models. The method may include receiving a voice query from a user that includes audio data corresponding to an utterance of the user, and context data associated with the user. One or more class models are then generated that collectively identify a first set of terms determined based on the context data, and a respective class to which the respective term is assigned for each respective term in the first set of terms. A language model that includes a residual unigram may then be accessed and processed for each respective class to insert a respective class symbol at each instance of the residual unigram that occurs within the language model. A transcription of the utterance of the user is then generated using the modified language model.

Abstract: This document generally describes systems and methods for dynamically adapting speech recognition for individual voice queries of a user using class-based language models. The method may include receiving a voice query from a user that includes audio data corresponding to an utterance of the user, and context data associated with the user. One or more class models are then generated that collectively identify a first set of terms determined based on the context data, and a respective class to which the respective term is assigned for each respective term in the first set of terms. A language model that includes a residual unigram may then be accessed and processed for each respective class to insert a respective class symbol at each instance of the residual unigram that occurs within the language model. A transcription of the utterance of the user is then generated using the modified language model.

Abstract: This document generally describes systems and methods for dynamically adapting speech recognition for individual voice queries of a user using class-based language models. The method may include receiving a voice query from a user that includes audio data corresponding to an utterance of the user, and context data associated with the user. One or more class models are then generated that collectively identify a first set of terms determined based on the context data, and a respective class to which the respective term is assigned for each respective term in the first set of terms. A language model that includes a residual unigram may then be accessed and processed for each respective class to insert a respective class symbol at each instance of the residual unigram that occurs within the language model. A transcription of the utterance of the user is then generated using the modified language model.

Abstract: This document generally describes systems and methods for dynamically adapting speech recognition for individual voice queries of a user using class-based language models. The method may include receiving a voice query from a user that includes audio data corresponding to an utterance of the user, and context data associated with the user. One or more class models are then generated that collectively identify a first set of terms determined based on the context data, and a respective class to which the respective term is assigned for each respective term in the first set of terms. A language model that includes a residual unigram may then be accessed and processed for each respective class to insert a respective class symbol at each instance of the residual unigram that occurs within the language model. A transcription of the utterance of the user is then generated using the modified language model.

Abstract: This document generally describes systems and methods for dynamically adapting speech recognition for individual voice queries of a user using class-based language models. The method may include receiving a voice query from a user that includes audio data corresponding to an utterance of the user, and context data associated with the user. One or more class models are then generated that collectively identify a first set of terms determined based on the context data, and a respective class to which the respective term is assigned for each respective term in the first set of terms. A language model that includes a residual unigram may then be accessed and processed for each respective class to insert a respective class symbol at each instance of the residual unigram that occurs within the language model. A transcription of the utterance of the user is then generated using the modified language model.

Abstract: A method is provided for producing fine denier multicomponent thermoplastic polymer filaments incorporating high melt-flow rate polymers. Multicomponent filaments are extruded such that the high melt-flow rate polymer component is substantially surrounded by one or more low melt-flow rate polymer components. The extruded multicomponent filament is then melt-attenuated with a significant drawing force to reduce the filament diameter and form continuous, fine denier filaments.

Abstract: The present invention provides multicomponent fine fiber webs and multilayer laminates thereof having an average fiber diameter less than about 7 micrometers and comprising a first olefin polymer component and a second distinct polymer component such as an amorphous polyolefin or polyamide. Multilayer laminates incorporating the fine multicomponent fiber webs are also provided such as, for example, spunbond/meltblown/spunbond laminates or spunbond/meltblown/meltblown/spunbond laminates. The fine multicomponent fiber webs and laminates thereof provide laminates having excellent softness, peel strength and/or controlled permeability.

Abstract: The present invention provides multicomponent fine fiber webs and multilayer laminates thereof having an average fiber diameter less than about 7 micrometers and comprising a first olefin polymer component and a second distinct polymer component such as an amorphous polyolefin or polyamide. Multilayer laminates incorporating the fine multicomponent fiber webs are also provided such as, for example, spunbond/meltblown/spunbond laminates or spunbond/meltblown/meltblown/spunbond laminates. The fine multicomponent fiber webs and laminates thereof provide laminates having excellent softness, peel strength and/or controlled permeability.

Abstract: An improved nonwoven web composite is formed by combining splittable bicomponent thermoplastic filaments with a component selected from other fibers and particles. The bicomponent filaments include distinct regions of first and second incompatible polymers extending the length of the filaments. After the bicomponent filaments are combined with the other fibers and/or particles, the bicomponent filaments are caused to split lengthwise along boundaries between the regions of different polymers, resulting in a web or matrix of finer filaments which entrap, ensnare and contain the other fibers and/or particles within the web or matrix. The nonwoven web composite is particularly useful for making absorbent articles, which require durability and optimum levels of absorbent fibers and/or particles.

Abstract: A method for producing super fine meltblown fibers increases the length of the meltblown jet thermal core to increase the dwell time of the extruded thermoplastic polymer within the jet thermal core. Through use of the method it is practical to use low viscosity resins and further to provide meltblown nonwovens with superior barrier properties to the passage of fluids and particularly gases. The method further provides a useful means for blooming internal additives to the surface of the fibers.

Abstract: An improved nonwoven web composite is formed by combining bicomponent thermoplastic filaments having adhesive properties with a component selected from other fibers and particles. The bicomponent filaments include distinct regions of first and second incompatible polymers across a cross-section of individual filaments. After the bicomponent filaments are combined with the other fibers and/or particles, the adhesive properties of the bicomponent filaments result in a web or matrix of filaments having improved ability to entrap, ensnare and contain the other fibers and/or particles within the web or matrix. The nonwoven web composite is particularly useful for making absorbent articles, which require stability and optimum levels of absorbent fibers and/or particles.

Abstract: A method for producing super fine meltblown fibers increases the length of the meltblown jet thermal core to increase the dwell time of the extruded thermoplastic polymer within the jet thermal core. Through use of the method it is practical to use low viscosity resins and further to provide meltblown nonwovens with superior barrier properties to the passage of fluids and particularly gases. The method further provides a useful means for blooming internal additives to the surface of the fibers.

Abstract: A method is provided for producing fine denier multicomponent thermoplastic polymer filaments incorporating high melt-flow rate polymers. Multicomponent filaments are extruded such that the high melt-flow rate polymer component is substantially surrounded by one or more low melt-flow rate polymer components. The extruded multicomponent filament is then melt-attenuated with a significant drawing force to reduce the filament diameter and form continuous, fine denier filaments.