Abstract: Apparatuses, systems, and techniques to select, from an elastic neural network, a sub-network that satisfies deployment constraints. In at least one embodiment, a sub-network is selected from an elastic neural network by using routers trained to select candidate sets of attention heads for multi-head attention (MHA) sub-blocks and candidate sets of neurons for multi-layer perceptron (MLP) sub-blocks.

Abstract: A semiconductor device includes a semiconductor substrate, a channel region, a gate structure, two epitaxial structures, and two silicide structures. The channel region is disposed on the semiconductor substrate. The gate structure is disposed on the semiconductor substrate and over the channel region. The epitaxial structures are connected at opposite ends of the channel region and are disposed opposite to each other relative to the gate structure. The silicide structures respectively surround the epitaxial structures. A method of manufacturing a semiconductor device is also provided.

Abstract: The present disclosure describes a method of fabricating a semiconductor structure that includes forming a fin structure on a substrate, forming a polysilicon gate structure on a first portion of the fin structure, forming an opening in a second portion of the fin structure, wherein the first and second portions of the fin structure is adjacent to each other, forming a recess laterally on a sidewall of the first portion of the fin structure underlying the polysilicon gate structure, and forming an inner spacer structure within the recess. The inner spacer structure comprises an inner air spacer enclosed by a first dielectric spacer layer and a second dielectric spacer layer.

Abstract: The present disclosure describes an inner spacer structure for a semiconductor device and a method for forming the same. The method for forming the inner spacer structure in the semiconductor device can include forming a vertical structure over a substrate, forming a gate structure over a portion of the vertical structure, exposing sidewalls of the portion of the vertical structure, forming multiple spacers over the sidewalls of the portion of the vertical structure, and forming a void in each of the multiple spacers.

Abstract: A method for estimating at least one electrical property of a semiconductor device is provided. The method includes forming the semiconductor device and at least one testing unit on a substrate, irradiating the testing unit with at least one electron beam, estimating electrons from the testing unit induced by the electron beam, and estimating the electrical property of the semiconductor device according to intensity of the estimated electrons from the testing unit.

Abstract: The present invention provides a voltage regulator configured to receive a supply voltage to generate a regulated voltage. The voltage regulator comprises an operation amplifier and a power transistor. The operational amplifier is configured to receive a reference voltage and a feedback signal to generate an output signal. The power transistor is coupled to the operational amplifier, wherein a gate electrode receives the output signal of the operational amplifier, a first electrode is coupled to the supply voltage, and a second electrode is used to generate the regulated voltage. The operational amplifier comprises an input stage, a current source and a capacitor, wherein the current source with the capacitor are configured to provide an AC current to the gate electrode of the power transistor for ripple cancellation.

Abstract: The present disclosure relates to a semiconductor device including first and second terminals formed on a fin region and a seal layer formed between the first and second terminals. The seal layer includes a silicon carbide material doped with oxygen. The semiconductor device also includes an air gap surrounded by the seal layer, the fin region, and the first and second terminals.

Abstract: The present invention proposes a method for sensing underwater point cloud. The method includes: acquiring point cloud information through an underwater sonar array, where the point cloud information contains multiple points; rearranging these points into a matrix which includes the aforementioned points and multiple missing points; and for each missing point, interpolating its depth value based on the depth values of several neighboring points.

Abstract: The invention proposes a sensing system for point clouds and images. The system includes a point cloud sensor, an image sensor, and a computing module. The point cloud sensor captures point cloud information of a target object. The image sensor captures a two-dimensional image of the target object. The computing module extracts three-dimensional feature points from the point cloud information and two-dimensional feature points from the two-dimensional image, and calculates multiple coefficients in a transformation matrix based on coordinates of the three-dimensional feature points and coordinates of the two-dimensional feature points. The computing module also performs a coordinate transformation process on the point cloud information or the two-dimensional image according to the transformation matrix.

Abstract: A 3D-configured production structure of rubber products based on an intelligent manufacturing unit and a production method thereof. The structure includes a stereoscopic production warehouse used to store a mobile intelligent manufacturing unit and an ex-warehouse delivery system used to deliver the mobile intelligent manufacturing unit. The mobile intelligent manufacturing unit includes a unit functional assembly, a molding vulcanization apparatus, a blank feeder, a material-delivering apparatus, a product-collecting apparatus and a reclaimer. The molding vulcanization apparatus includes a upper heat plate, a upper mold, a lower mold, a lower heat plate and a support post. The upper mold and the lower mold are arranged on the inner sides of the upper heat plate and the lower heat plate, respectively. The upper heat plate is fixed on one end of the support post, and the lower heat plate is arranged through the support post.

Abstract: A data classification method, for classifying unlabeled images into an inlier data set or an outlier data set, include following steps. The unlabeled images are obtained. An assigned inlier image is selected among the unlabeled images. A similarity matrix is computed and the similarity matrix includes first similarity scores of the unlabeled images relative to the assigned inlier image. Each of the unlabeled images is classified into an inlier data set or an outlier data set according to the similarity matrix, so as to generate inlier-outlier predictions of the unlabeled images.

Abstract: Disclosed is an optical system using image restoration, including a light source, a pinhole, a testing platform, an image sensor and an image processing device. The pinhole is disposed on a light transmission path of the light source. The testing platform is disposed on the light transmission path of the light source and the pinhole is located between the light source and the testing platform. The testing platform is used to place a testing sample. The image sensor is disposed below the testing platform, and used to sense the testing sample so as to output an optical diffraction signal. The image processing device is electrically connected to the image sensor and used to perform signal processing and optical signal recognition on the optical diffraction signal of the testing sample so as to obtain a clear image of the testing sample.

Abstract: A memory cell structure includes a transistor structure and a capacitor structure, where the capacitor structure includes a hydrogen absorption layer. The hydrogen absorption layer absorbs hydrogen, which prevents or reduces the likelihood of the hydrogen diffusing into an underlying metal-oxide channel of the transistor structure. In this way, the hydrogen absorption layer minimizes and/or reduces the likelihood of hydrogen contamination in the metal-oxide channel, which may enable a low current leakage to be achieved for the memory cell structure and reduces the likelihood of data corruption and/or failure of the memory cell structure, among other examples.

Abstract: A data classification method includes following steps. Text samples are obtained from a dataset. The text samples are converted into text embeddings in a semantic space. An outlier-inlier ranking of the text samples is generated based on an outlier detection algorithm according to distances between the text embeddings in the semantic space. Partial samples are selected from the text samples according to the outlier-inlier ranking. A manual input command is received to assign manual-input labels on the partial samples. A prompt message is generated according to the partial samples with the manual-input labels and unlabeled samples of the text samples. The prompt message is provided to a generative pre-trained transformer model for generating inlier-outlier prediction labels about the unlabeled samples.

Abstract: A method includes following steps. First transistors are formed over a substrate. An interconnect structure is formed over the plurality of first transistors. A dielectric layer is formed over the interconnect structure. 2D semiconductor seeds are formed over the dielectric layer. The 2D semiconductor seeds are annealed. An epitaxy process is performed to laterally grow a plurality of 2D semiconductor films respectively from the plurality of 2D semiconductor seeds. Second transistors are formed on the plurality of 2D semiconductor films.

Abstract: An integrated chip having an integrated function capable of providing cell image and biochemical detection is provided and includes a sequentially stacked and sealed laminate set. The laminate set includes an upper laminate, a middle laminate and a lower laminate. The upper laminate has one or more holes for sample injection and sample or air discharging of. The middle laminate includes at least two hollow structures that define an imaging chamber and a biochemical detection area. The lower laminate includes at least one filtering element and at least one electrode sensing element disposed in the biochemical detection. The filtering element is for blocking suspended particles, and the electrode sensing element has electrode terminals for connecting to instruments or equipment to perform the measurements and analysis of electrochemistry and impedance.

Abstract: A method includes following steps. A dielectric layer is formed over a substrate. A transition metal-containing layer is deposited on the dielectric layer. The transition metal-containing layer is patterned into a plurality of transition metal-containing pieces. The transition metal-containing pieces are sulfurized or selenized to form a plurality of semiconductor seeds. Semiconductor films are grown from semiconductor seeds. Transistors are formed on the semiconductor films.

Abstract: A semiconductor device includes a substrate, a 2-D material layer, source/drain contacts, and a gate electrode. The 2-D material layer is over the substrate, the 2-D material layer includes source/drain regions and a channel region between the source/drain regions, in which the 2-D material layer is made of a transition metal dichalcogenide (TMD). The source/drain contacts are in contact with source/drain regions of the 2-D material layer, in which a binding energy of transition metal atoms at the channel region of the 2-D material layer is different from a binding energy of the transition metal atoms at the source/drain regions of the 2-D material layer. The gate electrode is over the substrate.

Abstract: A method for forming an interconnect structure is provided. The method includes the following operations. A contact is formed over a substrate. An interlayer dielectric (ILD) layer is formed over the contact and the substrate. An opening is formed in the ILD layer thereby exposing a portion of the contact. A densified dielectric layer is formed at an exposed surface of the ILD layer by the opening and an oxide layer over the portion of the contact by irradiating a microwave on the exposed surface of the ILD layer.

Abstract: An Axial Dispersion Bioreactor (ADBR) is designed for the photoautotrophic, mixotrophic or heterotrophic growth and production of microalgae and other microorganisms (bacteria, fungi, etc.) as well as cell cultures of plants, animals, insects and others. The ADBR is equipped with a plate having a plurality of holes that moves longitudinally or axially within the bioreactor to effect superior hydrodynamic and mixing patterns. The ADBR has the advantages of providing a low-shear culture environment, superior liquid mixing, and efficient gas mass transfer.