Abstract: A memory includes a memory device, a reading device and a feedback device. The memory device stores a plurality of bits. The reading device includes first and second reading circuits coupled to the memory device. The second reading circuit is coupled to the first reading circuit at a first node. The first and second reading circuits cooperates with each other to generate a first voltage signal at the first node based on at least one first bit of the plurality of bits. The feedback device adjusts at least one of the first reading circuit or the second reading circuit based on the first voltage signal. The first and second reading circuits generate a second voltage signal, different from the first voltage signal, corresponding to the bits, after the at least one of the first reading circuit or the second reading circuit is adjusted by the feedback device.

Abstract: Methods, computer systems, and apparatus, including computer programs encoded on computer storage media, for training an audio-processing neural network that includes at least (1) a first encoder network having a first set of encoder network parameters and (2) a decoder network having a set of decoder network parameters. The system obtains a set of un-labeled audio data segments, and generates, from the set of un-labeled audio data segments, a set of encoder training examples. The system performs training of a second encoder neural network that includes at least the first encoder neural network on the set of generated encoder training examples. The system also obtains one or more labeled training examples, and performs training of the audio-processing neural network on the labeled training examples.

Abstract: A near-field sensing information transmission and pairing system for air-land unmanned vehicles and a method thereof are disclosed. The unmanned aviation vehicle is driven to move along a forward route of an unmanned ground vehicle in advance, and continuously senses whether an obstacle or a traffic control appears in the forward route; when sensing the presence of the obstacle or the traffic control feature, the unmanned aviation vehicle generates and transmits a re-routing signal to the unmanned ground vehicle, an alternative route is calculated to replace the original forward route based on map information, a current coordinate, an obstacle coordinate and a destination coordinate, so as to increase a reaction time of the unmanned ground vehicle, thereby achieving the technical effect of improving immediacy of the unmanned ground vehicle to avoid a roadblock.

Abstract: An avatar carrier generating system for metaverse and method thereof are disclosed. Before a living body enters a metaverse, a sensing element senses the living body to generate a physiological feature, and after a real-name identity message is received, the real-name identity message and the physiological feature are converted into metadata compliant with a non-fungible token standard; the non-fungible token, which is owned by the living body, is generated on the blockchain based on the metadata. When the living body enters the metaverse, an avatar carrier for entering the metaverse is generated based on the metadata of the non-fungible token, and when the avatar carrier receives an identification request, the avatar carrier is permitted to provide at least one of the physiological feature and the real-name identity message, so as to achieve the effect of improving identifiability and authentication of the avatar carrier of the metaverse.

Abstract: A system for companion robot with 3D display device and a method are disclosed. In the system, a companion robot continuously senses a user behavior status and an input speech, converts the input speech to a text, transmits the text and the user behavior status to finite state machines of a server-end host to perform parsing and transition, so that the server-end host generates a chat message having an emotion label, transmits the chat message to an AI device, receives and stores the companion conversation message to a companion conversation list; the companion robot selects companion conversation message from the companion conversation list as an output message, and convert the output message to speech which is broadcasted by a speaker; while broadcasting the output message, the companion robot continuously obtains image data from a multi-source image and outputs the image data to a 3D display device for displaying.

Abstract: A generative chatbot system for virtual community and a method thereof are disclosed. In the system, a client-end host is linked to a virtual community to receive a chat message; when the trigger signal is detected, the client-end host generates an operation interface for inputting the customized instruction, and transmits the chat message and the customized instruction to the server-end host, which can integrate the chat message as the context message having a timing logic and transmit the context message and the customized instruction to an Al device to generate a response message. The server-end host stores the response message received from the artificial intelligence device to a response list, so that the client-end host can select one of the response messages from the response list and output the selected response message to the virtual community.

Abstract: A system of generative chatbot in a real multi-person response situation and a method thereof are disclosed. In the system, speech signals are sensed and converted into feature vectors and text messages, and a timing label and a classification label are embedded into the text messages, the text messages are stored as a context message, so that a server-end host can determine a timing logic of a multi-person conversation, and the context message and the timing logic are transmitted to an artificial intelligence device which determines a current conversation stage, a topic evolution, predicts a conversation development, and actively generates and stores a response message to the server-end host; the server-end host can filter out the response messages and transmit the filtered response message to a portable device for output. Therefore, the technical effect of improving conversational initiative and response efficiency can be achieved.

Abstract: A PCIe clock detection circuit includes a clock detector, a clock receiver, a counter coupled to the clock receiver, a multiplexer coupled to the counter, and an AND gate coupled to the clock detector and the multiplexer. The clock detector is used to detect amplitude of a clock signal and generate a clock detection signal accordingly. The clock receiver is used to generate a reference clock signal according to the clock signal. The counter is used to generate a counter signal according to the reference clock signal. The multiplexer is used to generate a MUX output signal according to the counter signal and a reference signal. The AND gate is used to generate a clock detection output signal according to the clock detection signal and the MUX output signal.

Abstract: A sense amplifier of a memory device that includes sense amplifier circuits and a reference sharing circuit is introduced. The sense amplifier circuits are configured to sense the plurality of bit lines according to an enable signal. The reference sharing circuit includes first switches and second switches that are coupled to the reference nodes and second reference nodes of the sense amplifier circuits, respectively. The first switches and second switches are controlled according to a control signal to control a first electrical connection among the first reference nodes, and to control a second electrical connection among the second reference nodes. An operation method of the sense amplifier and a memory device including the sense amplifier are also introduced.

Abstract: In a method of fabricating at least one IC, doped regions are formed on a semiconductor wafer using a first photolithography mask, including at least one doped region of a test structure. Active regions are formed on the semiconductor wafer using a second photolithography mask, including active regions of the test structure. Electrical contacts are formed on the active regions of the test structure. Electrical resistances are measured between pairs of active regions of the test structure using the electrical contacts. At least one metric is determined indicating whether the doped regions are spatially aligned with the active regions based on the measured electrical resistances. In response to the at least one metric indicating the doped regions are spatially aligned with the active regions, completing fabrication of the at least one integrated circuit.

Abstract: A memory includes a memory device, a reading device and a feedback device. The memory device stores a plurality of bits. The reading device includes first and second reading circuits coupled to the memory device. The second reading circuit is coupled to the first reading circuit at a first node. The first and second reading circuits cooperates with each other to generate a first voltage signal at the first node based on at least one first bit of the plurality of bits. The feedback device adjusts at least one of the first reading circuit or the second reading circuit based on the first voltage signal. The first and second reading circuits generate a second voltage signal, different from the first voltage signal, corresponding to the bits, after the at least one of the first reading circuit or the second reading circuit is adjusted by the feedback device.

Abstract: A method for automatic speech recognition using joint acoustic echo cancellation, speech enhancement, and voice separation includes receiving, at a contextual frontend processing model, input speech features corresponding to a target utterance. The method also includes receiving, at the contextual frontend processing model, at least one of a reference audio signal, a contextual noise signal including noise prior to the target utterance, or a speaker embedding including voice characteristics of a target speaker that spoke the target utterance. The method further includes processing, using the contextual frontend processing model, the input speech features and the at least one of the reference audio signal, the contextual noise signal, or the speaker embedding vector to generate enhanced speech features.

Abstract: There is provided a pixel circuit capable of outputting time difference data or image data, and including a first temporal circuit and a second temporal circuit. The first temporal circuit is used to store detected light energy of a first interval and a second interval as the time difference data. The second temporal circuit is used to store detected light energy of the second interval as the image data. The pixel circuit is used to output a pulse width signal corresponding to the time difference data or the image data in different operating modes.

Abstract: An active chatbot system with behavioral awareness and on-demand conversation and a method thereof are disclosed. In the active chatbot system, a client-end host continuously senses a client behavior state, and transmits the sensed client behavior state and an on-demand conversation setting to a server-end host to generate a rough question message having a natural language structure, and input the rough question message to a plurality of finite state machines to generate a precise question message. The server-end host transmits the precise question message to an artificial intelligence platform to obtain a corresponding answer message, stores the answer message to an answer list, filters out an answer message matching the on-demand conversation setting as an on-demand conversation message, transmits the on-demand conversation message to the client-end host for output. Therefore, the technical effect of improving human-computer interaction and initiative of chatbot can be achieved.

Abstract: An active chatbot system with customized setting and updating download and a method thereof are disclosed. In the active chatbot system, a rough question message having a natural language structure is generated based on a client behavior state and an customized parameter, and the rough question message is inputted to a logic circuit to generate a precise question message, and the precise question message is transmitted to an artificial intelligence platform to obtain a corresponding answer message which is used as first training data, the customization is used as second training data, the first training data and the second training data are inputted to an AI model to perform a training, weighting values of the AI model are translated to correspond to different customized parameter, a translating result is provided for the client-end host to download for customization setting and update. Therefore, the technical effect of improving convenience and specificity in model customization can be achieved.

Abstract: An active chatbot system with composite finite state machine and a method thereof are disclosed. In the active chatbot system, a rough question message having a natural language structure is generated based on a client behavior state and an on-demand conversation setting, and the rough question message is inputted to a question optimization circuit to generate a precise question message, and the precise question message is transmitted to an artificial intelligence platform to obtain a corresponding answer message, the answer message is inputted to a trained emotional AI model to generate an emotional answer message and the emotional answer message is stored in an answer list, so that the emotional answer message matching the on-demand conversation setting can be filtered out as an on-demand conversation message, the on-demand conversation message is transmitted to the client-end host for output. Therefore, the technical effect of improving conversational flexibility and realism of chatbot can be achieved.

Abstract: A method includes receiving a sequence of audio features characterizing an utterance and processing, using an encoder neural network, the sequence of audio features to generate a sequence of encodings. At each of a plurality of output steps, the method also includes determining a corresponding hard monotonic attention output to select an encoding from the sequence of encodings, identifying a proper subset of the sequence of encodings based on a position of the selected encoding in the sequence of encodings, and performing soft attention over the proper subset of the sequence of encodings to generate a context vector at the corresponding output step. The method also includes processing, using a decoder neural network, the context vector generated at the corresponding output step to predict a probability distribution over possible output labels at the corresponding output step.

Abstract: A method for performing speech recognition using sequence-to-sequence models includes receiving audio data for an utterance and providing features indicative of acoustic characteristics of the utterance as input to an encoder. The method also includes processing an output of the encoder using an attender to generate a context vector, generating speech recognition scores using the context vector and a decoder trained using a training process, and generating a transcription for the utterance using word elements selected based on the speech recognition scores. The transcription is provided as an output of the ASR system.

Abstract: A sense amplifier of a memory device that includes sense amplifier circuits and a reference sharing circuit is introduced. The sense amplifier circuits are configured to sense the plurality of bit lines according to an enable signal. The reference sharing circuit includes first switches and second switches that are coupled to the reference nodes and second reference nodes of the sense amplifier circuits, respectively. The first switches and second switches are controlled according to a control signal to control a first electrical connection among the first reference nodes, and to control a second electrical connection among the second reference nodes. An operation method of the sense amplifier and a memory device including the sense amplifier are also introduced.

Abstract: A clamp may include two substantially planar plates opposite from and generally parallel to one another, each plate having a mounting hole formed therein for receiving a fastener, two teeth, each tooth extending from a respective one of the two plates, such that each tooth is generally perpendicular to the plates, and such that each tooth extends towards the other from its respective plate, and a tongue opposite from and mechanically interfaced between the two plates, such that the tongue has a concave curved profile between the plates, and such that when a force is applied in a first direction by a fastener engaging with the clamp through the mounting holes, compression of the plates towards one another causes deflection of the tongue in a second direction substantially perpendicular to the first direction.