Abstract: A method of presenting information to a party through a messaging application is described. Responsive to receipt of a communication from a party (e.g., the first user), a reply is sent. The communication and the reply is presented in an interface to the sender. The messaging system determines matching content that is relevant to one or both of the communication and the reply and determines a quality of the match. Determining the quality of the match may include determining a score for an advertisement based on the advertisement's responsiveness to content identified in the reply message that was sent. Based on a determination that the quality is above a threshold, the matching content is presented along with the communication and the reply.

Abstract: A method of combining data streams from fixed audio-visual sensors with data streams from personal mobile devices including, forming a communication link with at least one of one or more personal mobile devices; receiving at least one of an audio data stream and/or a video data stream from the at least one of the one or more personal mobile devices; determining the quality of the at least one of the audio data stream and/or the video data stream, wherein the audio data stream and/or the video data stream having a quality above a threshold quality is retained; and combining the retained audio data stream and/or the video data stream with the data streams from the fixed audio-visual sensors.

Abstract: A method of combining data streams from fixed audio-visual sensors with data streams from personal mobile devices including, forming a communication link with at least one of one or more personal mobile devices; receiving at least one of an audio data stream and/or a video data stream from the at least one of the one or more personal mobile devices; determining the quality of the at least one of the audio data stream and/or the video data stream, wherein the audio data stream and/or the video data stream having a quality above a threshold quality is retained; and combining the retained audio data stream and/or the video data stream with the data streams from the fixed audio-visual sensors.

Abstract: A plurality of corpora is received from one or more sources. A separate model is trained on each corpus of the plurality of corpora. The models for the plurality of corpora are merged into a joint model using parameter interpolation. The models for each corpus of the plurality of corpora are retrained separately using the joint model. A single model is created based on the retrained models.

Abstract: A method, executed by a computer, includes monitoring a conversation between a plurality of meeting participants, identifying a conversational focus within the conversation, generating at least one question corresponding to the conversational focus, and retrieving at least one answer corresponding to the at least one question. A computer system and computer program product corresponding to the method are also disclosed herein.

Abstract: A method of combining data streams from fixed audio-visual sensors with data streams from personal mobile devices including, forming a communication link with at least one of one or more personal mobile devices; receiving at least one of an audio data stream and/or a video data stream from the at least one of the one or more personal mobile devices; determining the quality of the at least one of the audio data stream and/or the video data stream, wherein the audio data stream and/or the video data stream having a quality above a threshold quality is retained; and combining the retained audio data stream and/or the video data stream with the data streams from the fixed audio-visual sensors.

Abstract: A method of presenting information to a party through a messaging application is described. Responsive to receipt of a communication from a party (e.g., the first user), a reply is sent. The communication and the reply is presented in an interface to the sender. The messaging system determines matching content that is relevant to one or both of the communication and the reply and determines a quality of the match. Determining the quality of the match may include determining a score for an advertisement based on the advertisement's responsiveness to content identified in the reply message that was sent. Based on a determination that the quality is above a threshold, the matching content is presented along with the communication and the reply.

Abstract: LED drivers and lighting systems having a dynamically reconfigurable switching network is disclosed. The lighting system comprises a switching network which provides interconnections between LEDs or LED lamps that can dynamically reconfigure between largely parallel circuits and largely serial circuits as the instantaneous voltage of a power supply varies. When the instantaneous voltage of the power source is low, the switching network interconnects the LED lamps in a largely parallel configuration. As the instantaneous voltage increases, the switching network reconfigures and interconnects the LED lamps with increasing serial configurations. As the instantaneous voltage begins to decrease, the switching network reconfigures and interconnects the LED lamps back to a largely parallel configuration.

Abstract: A method of combining data streams from fixed audio-visual sensors with data streams from personal mobile devices including, forming a communication link with at least one of one or more personal mobile devices; receiving at least one of an audio data stream and/or a video data stream from the at least one of the one or more personal mobile devices; determining the quality of the at least one of the audio data stream and/or the video data stream, wherein the audio data stream and/or the video data stream having a quality above a threshold quality is retained; and combining the retained audio data stream and/or the video data stream with the data streams from the fixed audio-visual sensors.

Abstract: Some embodiments relate to using an unnormalized neural network language model in connection with a speech processing application. The techniques include obtaining a language segment sequence comprising one or more language segments in a vocabulary; accessing an unnormalized neural network language model having a normalizer node and an output layer comprising a plurality of output nodes, each of the plurality of output nodes associated with a respective language segment in the vocabulary; and determining, using the unnormalized neural network language model, a first likelihood that a first language segment in the vocabulary follows the language segment sequence.

Abstract: Techniques for conversion of non-back-off language models for use in speech decoders. For example, an apparatus for conversion of non-back-off language models for use in speech decoders. For example, an apparatus is configured convert a non-back-off language model to a back-off language model. The converted back-off language model is pruned. The converted back-off language model is usable for decoding speech.

Abstract: Techniques for conversion of non-back-off language models for use in speech decoders. For example, an apparatus for conversion of non-back-off language models for use in speech decoders. For example, an apparatus is configured convert a non-back-off language model to a back-off language model. The converted back-off language model is pruned. The converted back-off language model is usable for decoding speech.

Abstract: Techniques for conversion of non-back-off language models for use in speech decoders. For example, a method comprises the following step. A non-back-off language model is converted to a back-off language model. The converted back-off language model is pruned. The converted back-off language model is usable for decoding speech.

Abstract: Some embodiments relate to using an unnormalized neural network language model in connection with a speech processing application. The techniques include obtaining a language segment sequence comprising one or more language segments in a vocabulary; accessing an unnormalized neural network language model having a normalizer node and an output layer comprising a plurality of output nodes, each of the plurality of output nodes associated with a respective language segment in the vocabulary; and determining, using the unnormalized neural network language model, a first likelihood that a first language segment in the vocabulary follows the language segment sequence.

Abstract: Multi-stage LED drivers and lighting systems are disclosed. The lighting system comprises a bridge rectifier receiving power from an AC power line source, two or more current limiting devices, and a string of LEDs having a corresponding number of LED lamps connected in electrical series. Each of the current limiting devices is coupled to and provides current to the corresponding LED. When the instantaneous voltage of the alternating current source is near zero volts, no significant current flows through the string of LEDs. As the instantaneous voltage from the bridge rectifier increases beyond the forward voltages of the LEDs in the string, the current limiting devices energize and provide current to the corresponding LEDs as well as the LEDs connected to the cathode of the corresponding LED. The equivalent circuits of the lighting system vary throughout the power cycle of the AC line power source.

Abstract: A first portion of a midsole formed of a first material has a recess in a top surface thereof, and a first aperture positioned within the recess and extending therethrough. The first portion is placed in a recess in a bottom plate of a second mold assembly. A middle plate with a second aperture extending therethrough is placed in contact with the bottom plate. A second material is inserted through the second aperture such that it fills the recess and the first aperture in the first portion to form a second portion of the midsole. A top plate of the second mold assembly is positioned in contact with the middle plate so as to close the second mold assembly. The second mold assembly is subjected to heat and pressure such that the second portion cures and bonds to the first portion of the midsole.

Abstract: LED lighting systems employing plano optics, direct dissipation heat sinks, and linear AC direct current drivers are disclosed. The heat sink is configured to provide thermal management for the LED lights and AC linear current driver circuit elements, as well as to provide a means for holding lenses having diffractive optical elements. The LED lighting systems are compact, energy efficient, and may be used in many conventional incandescent or fluorescent lighting applications. Lenses having diffractive optical elements are designed to redirect light radiated from the LEDs into other directions.

Abstract: Multi-stage LED drivers and lighting systems are disclosed. The lighting system comprises a bridge rectifier receiving power from an AC power line source, two or more current limiting devices, and a string of LEDs having a corresponding number of LED lamps connected in electrical series. Each of the current limiting devices is coupled to and provides current to the corresponding LED. When the instantaneous voltage of the alternating current source is near zero volts, no significant current flows through the string of LEDs. As the instantaneous voltage from the bridge rectifier increases beyond the forward voltages of the LEDs in the string, the current limiting devices energize and provide current to the corresponding LEDs as well as the LEDs connected to the cathode of the corresponding LED. The equivalent circuits of the lighting system vary throughout the power cycle of the AC line power source.

Abstract: LED drivers and lighting systems having a dynamically reconfigurable switching network is disclosed. The lighting system comprises a switching network which provides interconnections between LEDs or LED lamps that can dynamically reconfigure between largely parallel circuits and largely serial circuits as the instantaneous voltage of a power supply varies. When the instantaneous voltage of the power source is low, the switching network interconnects the LED lamps in a largely parallel configuration. As the instantaneous voltage increases, the switching network reconfigures and interconnects the LED lamps with increasing serial configurations. As the instantaneous voltage begins to decrease, the switching network reconfigures and interconnects the LED lamps back to a largely parallel configuration.

Abstract: Methods, systems, and apparatus, including computer programs encoded on a non-transitory computer storage medium, for providing reverse auctions for goods and services in real-time. In one aspect, a method includes storing profile information associated with a plurality of information sources, each source associated with a category. A request for information from a user including information associated with a good or service sought is analyzed to identify information sources as responsive to the request based on a comparison of the good or service sought and each information source's associated category. Responses are solicited from the identified information sources. In response, a set of responses from at least a subset of the identified information sources is received, which are then ranked based on a comparison of the particular response to the request for information, and at least a subset of which are presented to the user based on the ranking.