Abstract: A multi-channel lighting control system includes a control module and a plurality of dimming modules. The control module generates a multi-level control signal and the multi-level control signal includes a plurality of control signal data arranged in a preset sequence. The dimming modules are corresponding to the control signal data respectively and arranged in the preset sequence. Each of the dimming modules, upon receiving the multi-level control signal, performs the signal extraction to extract the first of the control signal data from the multi-level control signal. Then, the dimming modules converts the extracted control signal data by the dimming module into a dimming signal, and outputs the dimming signal to a lighting device corresponding thereto.

Abstract: Methods and systems for intelligently recommending selections for a selector control are disclosed. The method includes receiving a recommendation request from a selector control client, the recommendation request comprising a search string and a unique identifier of a user interacting with a selector control; identifying user identifiers of usernames matching the search string; retrieving machine learning features corresponding to the user identifiers of usernames matching the search string; applying a machine learning model to the retrieved machine learning features to assign weights to the retrieved machine learning features; computing recommendation scores for the user identifiers based on the assigned weights to the retrieved machine learning features; ranking the user identifiers based on the recommendation scores; and forwarding a ranked list of user identifiers to the selector control client for displaying in the selector control for selection by the user interacting with the selector control.

Abstract: A data processing device, for providing a pre-training data for a language model, includes the following elements. A collection unit, for receiving a first dataset having a first category. An evaluation unit, for analyzing the first dataset to generate a category analysis result, evaluating the first dataset based on several of indicators of an evaluation rule to generate a first evaluation result, and determining whether the first evaluation result meets an evaluation criteria. A feedback unit, for converting and aggregating the first evaluation result to generate a first evaluation summary which is sent to the collection unit. A storage unit, for selectively storing the first dataset based on the first evaluation result. When the first evaluation result meets the evaluation criteria, the storage unit stores the first dataset which serves as the pre-training data. Otherwise, the evaluation unit provides several suggestions of adjustment for the first dataset.

Abstract: Certain aspects of the present disclosure provide techniques and apparatuses for training and using machine learning models to estimate a representation of a channel between a transmitter and a receiver in a spatial environment. An example method generally includes estimating a representation of a channel using a machine learning model trained to generate the estimated representation of the channel based on a location of a transmitter in a spatial environment, a location of a receiver in the spatial environment, and information about the spatial environment. One or more actions are taken based on the estimated representation of the channel.

Abstract: A reconfigurable PUF with two PUF functions comprises 2m×n PUF cells, a sequential control module, a row selection module, n amplification modules, n first bit lines, and n second bit lines. Each of the 2m×n PUF cells comprises a first PMOS transistor, a second PMOS transistor, a first NMOS transistor, a second NMOS transistor, a third NMOS transistor, a fourth NMOS transistor, a fifth NMOS transistor, a sixth NMOS transistor and a seventh NMOS transistor. The PUF cells can provide two independent responses, and can operate in a SRAM mode and an inverter mode. Therefore, the reconfigurable PUF with two PUF functions can operate both in the SRAM mode and the inverter mode, and a PUF operating mode with higher reliability is selected for generating final responses.

Abstract: Methods and systems for intelligently recommending selections for a selector control are disclosed. The method includes receiving a recommendation request from a selector control client, the recommendation request comprising a search string and a unique identifier of a user interacting with a selector control; identifying user identifiers of usernames matching the search string; retrieving machine learning features corresponding to the user identifiers of usernames matching the search string; applying a machine learning model to the retrieved machine learning features to assign weights to the retrieved machine learning features; computing recommendation scores for the user identifiers based on the assigned weights to the retrieved machine learning features; ranking the user identifiers based on the recommendation scores; and forwarding a ranked list of user identifiers to the selector control client for displaying in the selector control for selection by the user interacting with the selector control.

Abstract: The present invention discloses a highly reliable positive-to-negative voltage conversion circuit, including a first inverter, a first buffer, a second buffer, a first to seventh NMOS transistors, a first to third PMOS transistors, a first resistor, a second resistor, a third resistor, and a fourth resistor, wherein the third PMOS transistor, the third NMOS transistor, the fourth NMOS transistor, the fifth NMOS transistor, the sixth NMOS transistor and the seventh NMOS transistor form an on/off control circuit, with that a power supply path between a positive-voltage power supply terminal and a negative-voltage power supply terminal is physically isolated by the on/off control circuit to eliminate voltage reversing occurring when the positive-voltage power supply terminal and the negative-voltage power supply terminal cannot be synchronized during power-on.

Abstract: A machine learning attack resistant strong PUF with a dual-edge sampling function comprises switch units, a first arbiter and a second arbiter. The first arbiter is for determining a sequential order of delays at a rising edge of signals input to a first input terminal and a second input terminal of the first arbiter. The second arbiter is for determining a sequential order of delays at a falling edge of signals input to a first input terminal and a second input terminal of the second arbiter. Each switch unit is composed of eight MOS transistors. The strong PUF has a high capacity to resist machine learning attacks and small hardware expenditure through simple structural design of the switch units, realizing machine learning attack resistance and small hardware expenditure at the same time, and generating a large number of challenge response pairs through dual-edge sampling realized by the two arbiters.

Abstract: A PUF-based magnetometer with a safety protection circuit is provided. The PUF-based magnetometer includes a TMR magnetometer main structure, a control circuit, a column encoder circuit, a row encoder circuit, a multiplexer and a dynamic comparator. The TMR magnetometer main structure includes n*n TMR sensors, and the TMR magnetometer main structure can also form a PUF circuit together with a control circuit, a column encoder circuit, a row encoder circuit, a multiplexer and a dynamic comparator to realize a function of a TMR magnetometer. The PUF circuit generates PUF response signals based on random deviations of a TMR sensor fabrication process. The PUF response signals generated by the PUF circuit make it more difficult for attackers to locate specific positions of the TPR sensors, and even make it hard to know the existence of the deviations of the TMR sensors.

Abstract: Provided are a modified immunoglobulin single variable domain, and a protein or polypeptide comprising same. The modified immunoglobulin single variable domain has reduced binding to a pre-anti-drug antibody. Specifically, the C terminal of the immunoglobulin single variable domain comprises amino acid modification. The present invention further relates to an immunoglobulin single variable domain, the pharmaceutical use of a protein or polypeptide comprising same, a pharmaceutical composition, an encoding nucleic acid, and a preparation method.

Abstract: A microscope is provided to measure HbA1c fraction. The microscope measures the HbA1c fraction of a single red blood cell (RBC) in a trace blood sample. The HbA1c fraction can be measured through a non-invasive way while the RBC flows in a human epidermal microvessel, too. The microscope comprises a laser device, an upright microscope, a light splitter, a light detector, and a mainframe. Unlike traditional methods, the HbA1c fraction can be measured in vitro or in vivo at the level of a single RBC. Accurate measurement is achieved. Misdiagnosis rate is reduced. The microscope provides HbA1c fractions from hundreds of RBCs, instead of averaging HbA1c fractions obtained from a large number of blood samples. Hence, the present invention is a method of detecting a HbA1c fraction of a single RBC, and is a microscope supporting blood-drawing measurement and non-invasive measurement simultaneously.

Abstract: A lighting device power supplying system having switching mechanism includes a plurality of power controllers and each of the power controllers includes a driving module, a sampling module and a control module. The driving module generates a driving voltage to drive a lighting module. The sampling module samples the driving voltage to generate a sampled voltage. The control module compares the sampled voltage with a default oscillation frequency range and a default oscillation amplitude threshold. The control module generates an abnormal detecting result when the oscillation frequency of the sampled voltage is not within the default oscillation frequency range or when the oscillation amplitude of the sampled voltage exceeds the default oscillation amplitude threshold.

Abstract: A device encrypted authentication method based on a strong physical unclonable function is disclosed, including: generating ? matrixes formed by random transformation of a ??{square root over (b)}?-order unit matrix, such that unnecessary device expenditure is reduced; when a sensing device is identified, information to be synchronized between the sensing device and a back-end server is transmitted in the registration phase, such that an attacker cannot obtain confidential information in the encryption phase, and the attacker cannot traverse or completely crack all these possible combinations in a short time; even if the attacker intercepts some CRPs, it is scarcely possible that data sent by the sensing device matches the CRPs collected by the attacker, and with the increase of the number b of bits of challenges and the quick convergence to 0, the accuracy of random guessing is only 50%.

Abstract: The invention discloses a challenge screening method for improving the stability of a strong Physical Unclonable Function (PUF). If the stability of the group of challenges (GOCs) is quantized by 1 or 0, the GOCs are defined as stable challenges or unstable challenges respectively; the GOCs and the stability of corresponding responses are defined as a challenge-stability pair (CSP); a machining learning training dataset is constructed through CSPs for obtaining a screening model, to associate the challenge stability with the screening model; during actual application of the strong PUF, the challenge stability of a strong PUF is determined through the screening model, and stable challenges of the strong PUF are screened out to form a stable challenge set, which is input into the strong PUF to extract identity information. The challenge screening method can effectively improve the stability of various strong PUFs and has better effect with external disturbance increasing.

Abstract: Disclosed is a TFET with an OR-AND logic function. By arranging a horizontal channel and a vertical channel in different directions, three gates are not connected and will not be affected by each other, and can control the current of a whole channel jointly; when a first gate and a second gate are both at a high level, the TFET will be turned on, and when the first gate and a third gate are both at a high level, the TFET will also be turned on; the horizontal channel not only isolates the second gate from the third gate, but also reduces the strength of coupling between the second gate and the third gate, so only a small current passes through the horizontal channel when the second gate and the third gate are both at a high level, and the TFET will not be turned on at this moment.

Abstract: A signal processing circuit for a tunneling magnetoresistance sensor comprises thirty-eight MOS transistors, two capacitors and ten switches. A first switch, a second switch, a third switch, a fourth switch, a first MOS transistor, a second MOS transistor, a third MOS transistor and a fourth MOS transistor form a high-frequency modulation circuit. A fifth switch, a sixth switch, a seventh switch, an eighth switch, a seventeenth transistor, a twenty-second MOS transistor and a twenty-third MOS transistor form a high-frequency demodulation circuit. A signal output by a tunneling magnetoresistance sensor is modulated to a chopping frequency by the high-frequency modulation circuit, and is then demodulated by the high-frequency demodulation circuit.

Abstract: A bistable physical unclonable function circuit based on subthreshold leakage current deviation comprises a time sequence control circuit, a decoder, 16 bias voltage converters and a PUF array, wherein the time sequence control circuit is used for generating a precharge signal and an enable signal, the decoder is used for converting an external stimulus signal into 16 decoded signals under the control of the precharge signal and the enable signal, the kth bias voltage converter is used for converting the kth decoded signal into a kth word line signal which is input to the PUF array, the PUF array is used for generating 16 response signals under the control of the precharge signal and the 16 word line signals, and comprises four PMOS transistors, four NMOS transistors, two two-input NAND gates and sixteen PUF units, and each PUF unit comprises two NMOS transistors.

Abstract: A refrigerant pipe constituting a refrigerant circuit of a refrigeration apparatus, includes: a pipe body made of stainless steel; and a first connecting tube made of copper or a copper alloy and that is configured to connect a first different refrigerant pipe to the refrigerant pipe. The first connecting tube is connected to an outer circumferential surface of a first end of the pipe body in a pipe axis direction of the pipe body. The first connecting tube overlaps with the pipe body in a pipe diameter direction in an entirety of the first connecting tube in the pipe axis direction.

Abstract: The invention provides a multi-function intelligent lighting device control system and the method thereof. The system includes several lighting devices and an intelligent lighting device controller. The controller includes a detecting module having a detecting circuit, main power source module, intelligent control module and intelligent dimming module. The first electrode connectors of the main power source module are connected to the detecting circuit and connected to the first electrodes of the lighting devices via the detecting circuit. The common second electrode connector of the main power source module is connected to the second electrodes of the lighting devices and outputs an electricity signal to drive the lighting devices, such that the detecting circuit generates several detecting signals. The intelligent control module generates a feedback signal to perform the adaptive adjustment function accordingly.