Abstract: Techniques for training robust machine learning models for adversarial input data. Training data for a machine learning (ML) model is received. The training data includes a plurality of labels for data elements. First modified training data is generated by modifying one or more of the plurality of labels in the training data using parameterized label smoothing with a first optimization parameter. The ML model is trained using the first modified training data.

Abstract: An encoder and a signal processing method are disclosed. The method includes: receiving an analog signal, and generating a filtered analog signal by an analog filter according to the input signal and a first frequency indication signal; generating a digital signal by an analog-to-digital converter according to the filtered analog signal; generating a filtered digital signal by a digital filter according to the digital signal and a second frequency indication signal; generating a seventh signal and an eighth signal by a dynamic offset calibration unit according to the filtered digital signal and a period indication signal; and generating a position information by a position detection unit according to the seventh signal and the eighth signal. The first frequency indication signal, the second frequency indication signal and the period indication signal are generated by a frequency generation module according to the filtered analog signal or the digital signal.

Abstract: Example methods and systems to obtain images associated with target cells distributed in a cytology specimen have been disclosed. One example method includes obtaining a first image associated with a first region of the cytology specimen through a first object lens, determining a first subregion of the first region, obtaining a second set of one or more images associated with the first subregion through a second object lens, identifying a third image from the first image, and in response to a ratio of a second number of the target cells in the second set of one or more images to the first number of the target cells in the third image being in a predetermined range, obtaining images associated with the target cells based on the second set of images.

Abstract: An imaging optical system includes, in order from an object side to an image side, a first lens element, a second lens element, a third lens element, a fourth lens element and a fifth lens element. The first lens element with refractive power has an object-side surface being convex in a paraxial region and an image-side surface being concave in a paraxial region. The second lens element with positive refractive power has an image-side surface being convex in a paraxial region. The third lens element with negative refractive power has an image-side surface being concave in a paraxial region. The fourth lens element with positive refractive power has an image-side surface being convex in a paraxial region. The fifth lens element with negative refractive power has an image-side surface being concave in a paraxial region and having a convex shape in an off-axial region thereof.

Abstract: A trained machine learning model and a training dataset used to train the trained machine learning model can be received. Based on the training dataset, unsupervised adversarial examples can be generated. Robustness of the trained machine learning model can be determined using the generated unsupervised adversarial examples. The training dataset can be augmented with the generated unsupervised adversarial examples. The trained machine learning model can be retrained using the augmented training dataset.

Abstract: Semiconductor devices and methods of forming the same are provided. A method includes providing a workpiece having a semiconductor structure; depositing a two-dimensional (2D) material layer over the semiconductor structure; forming a source feature and a drain feature electrically connected to the semiconductor structure and the 2D material layer, wherein the source feature and drain feature include a semiconductor material; and forming a gate structure over the two-dimensional material layer and interposed between the source feature and the drain feature. The gate structure, the source feature, the drain feature, the semiconductor structure and the 2D material layer are configured to form a field-effect transistor. The semiconductor structure and the 2D material layer function, respectively, as a first channel and a second channel between the source feature and the drain feature.

Abstract: A wireless power supply device and method applied to an electronic lock is implemented between a power supply unit and a power receiving unit. The power supply unit is connected to an external power source and has a first radio frequency coil and a first wireless communication module. The power receiving unit is connected to an electronic lock and has a second radio frequency coil and a second wireless communication module. When the power supply unit wirelessly transmits the power provided by the external power source to the power receiving unit, the power receiving unit and the power supply unit perform a handshaking process to confirm whether they are paired objects, if yes, the power is allowed to be transmitted to the electronic lock, so that the security of wireless power supply can be improved.

Abstract: A smart electric heating device comprises a storage unit, a first heating unit, a second heating unit, a control unit and a first temperature sensing unit. With the first temperature sensing unit to obtain an ambient temperature, the control unit compares the ambient temperature with a maximum increased temperature and a set temperature for controlling the first heating unit and the second heating unit to actuate. In this way, each user can use hot water of sufficient temperature better.

Abstract: Methods for tuning threshold voltages of fin-like field effect transistor (FinFET) devices are disclosed herein. An exemplary integrated circuit device includes a high voltage n-type FinFET, a high voltage p-type FinFET, a low voltage n-type FinFET, and a low voltage p-type FinFET. Threshold voltages of the high voltage n-type FinFET and the high voltage p-type FinFET are greater than threshold voltages of the low voltage n-type FinFET and the low voltage p-type FinFET, respectively. The high voltage n-type FinFET, the high voltage p-type FinFET, the low voltage n-type FinFET, and the low voltage p-type FinFET each include a threshold voltage tuning layer that includes tantalum and nitrogen. Thicknesses of the threshold voltage tuning layer of the low voltage n-type FinFET and the low voltage p-type FinFET are less than thicknesses of the threshold voltage tuning layer of the high voltage n-type FinFET and the high voltage p-type FinFET, respectively.

Abstract: A resisting structure of ratchet wrench includes a ratchet wheel having a ratchet toothed portion fitting an outer periphery thereof, an engagement unit having an engagement toothed portion meshing with the ratchet toothed portion, a changeover groove, and a curved surface portion disposed on two sides of the changeover groove, and a changeover unit having an assembling portion disposed on a front side of the changeover unit and situated correspondingly to the changeover groove and a support portion disposed on two sides of the assembling portion and situated correspondingly to the curved surface portion. When the ratchet wheel moves back against a direction of applied force, the support portion of the changeover unit can limit a downward distance caused when the curved surface portion is pressed downwards to thereby ensure each moving back angle of the ratchet wheel.

Abstract: A method for fault location to multi-terminal traveling wave in a direct current distribution line, which belongs to the field of power line fault ranging and location technology. The method includes a main site and a plurality of acquisition points installed into the distribution line and includes steps as follows. Step 1001: collecting and uploading a traveling wave signal by each of the acquisition points after a fault occurs in the line; step 1002: generating a fault record set; step 1003: computing the shortest paths from a central site to other sites and their lengths; step 1004: using expanded two-terminal traveling wave ranging principle for pairing computation; step 1005: converting positions of possible disturbance points into possible disturbance occurrence time points; step 1006: extrapolating from equal path lengths of the possible disturbance points to obtain disturbance time data; and step 1007: determining a final disturbance point.

Abstract: A lock monitoring device includes a detection unit and a plurality of sensors, where the detection unit is configured with a thin circuit board configured with a plurality of detection elements, allowing the detection elements to be connected in parallel with a circuit and then connected to a control circuit; the sensors are configured on parts of the lock such as lock tongues, rotating shaft and correspond to the detection elements; whereby, when the lock is unlocked and locked, the detection elements will detect that the detection elements depart from detection positions and send detection signals to a remote monitoring host, and the monitoring host monitors and records the number and time of the unlocking of the lock, achieving smart access control and further allowing the lock to be safer.

Abstract: The present disclosure provides a motor structure including a winding set, a rotor and a motor base. The rotor is driven by the winding set. The motor base includes a central portion, a surrounding portion and at least one connecting element. The central portion is configured to dispose the winding set thereon. The surrounding portion is disposed around the central portion. The connecting element is deformable and connected between the central portion and the surrounding portion. The connecting element includes a first supporting part, a second supporting part and a deformation part. The first supporting part and the second supporting part are connected with the central portion and the surrounding portion, respectively, and the deformation part is connected therebetween. The hardnesses of the first supporting part and the second supporting part are greater than or equal to those of the central portion and the surrounding portion, respectively.

Abstract: A device includes a substrate and a gate structure over the substrate. The device further includes source/drain (S/D) features in the substrate. At least one of the S/D features is located in a trench. The at least one S/D feature includes a first semiconductor material covering an entirety of a bottom surface of the trench. The at least one S/D feature further includes a second semiconductor material over the first semiconductor material. The at least one S/D feature further includes a third semiconductor material over the second semiconductor material. The second semiconductor material has a composition different from the first semiconductor material and the third semiconductor material. The first semiconductor material includes physically discontinuous portions directly contacting the substrate. The second semiconductor material surrounds the third semiconductor material.

Abstract: A wavelength converting layer is partially diced to generate a first and second wavelength converting layer segment and to allow partial isolation between the first segment and the second segment such that the wavelength converting layer segments are connected by a connecting wavelength converting layer. The first and second wavelength converting layer segments are attached to a first and second light emitting device, respectively to create a first and second pixel. The connecting wavelength converting layer segment is removed to allow complete isolation between the first pixel and the second pixel. An optical isolation material is applied to exposed surfaces of the first and second pixel and a sacrificial portion of the wavelength converting layer segments and optical isolation material attached to the sacrificial portion is removed from a surface facing away from the first light emitting device, to expose a emitting surface of the first wavelength converting layer segment.

Abstract: A chip package structure is provided. The chip package structure includes a redistribution structure and a first chip structure over the redistribution structure. The chip package structure also includes a first solder bump between the redistribution structure and the first chip structure and a first molding layer surrounding the first chip structure. The chip package structure further includes a second chip structure over the first chip structure and a second molding layer surrounding the second chip structure. In addition, the chip package structure includes a third molding layer surrounding the first molding layer, the second molding layer, and the first solder bump. A portion of the third molding layer is between the first molding layer and the redistribution structure.

Abstract: Zwitterionic surfactant compounds and their use in connection with absorbent paper manufacture in debonder and softener formulations. An exemplary surfactant is an imidazolinium compound of formula: wherein X is SO3, R1 is a hydrocarbyl spacer group having a chain length of from 2-6 carbon atoms, R2 is a hydrocarbyl group having from 8 to 22 carbon atoms; and R3 is an alkenylamidoalkyl moiety having from 8 to 30 carbon atoms.

Abstract: An image capturing lens assembly includes seven lens elements, which are, in order from an object side to an image side, a first lens element, a second lens element, a third lens element, a fourth lens element, a fifth lens element, a sixth lens element and a seventh lens element. The fifth lens element with positive refractive power has an object-side surface being concave in a paraxial region thereof and an image-side surface being convex in a paraxial region thereof. The sixth lens element has an object-side surface being convex in a paraxial region thereof and an image-side surface being concave in a paraxial region thereof. The seventh lens element has an object-side surface being convex in a paraxial region thereof and an image-side surface being concave in a paraxial region thereof, and the image-side surface thereof includes at least one convex shape in an off-axis region thereof.

Abstract: One or more computing devices, systems, and/or methods for implementing an automated model update pipeline are provided. User behavior data associated with content provided to users may be collected. An automatic model training is invoked to train a new model to output a set of model parameters based upon a configuration specifying a target audience, features extracted from user behavior data, and training model parameters.

Abstract: A semiconductor structure and a manufacturing method thereof are provided. The semiconductor structure includes a first under-bump metallization (UBM) pattern, a first conductive via, and a first dielectric layer laterally covering the first UBM pattern and the first conductive via. Entireties of a top surface and a bottom surface of the first UBM pattern are substantially planar. The first conductive via landing on the top surface of the first UBM pattern includes a vertical sidewall and a top surface connected to the vertical sidewall, and a planarized mark is on the top surface of the first conductive via. A bottom surface of the first dielectric layer is substantially flush with the bottom surface of the first UBM, and a top surface of the first dielectric layer is substantially flush with the top surface of the first conductive via.