Abstract: A semiconductor device includes a first gate structure extending along a first lateral direction. The semiconductor device includes a first interconnect structure, disposed above the first gate structure, that extends along a second lateral direction perpendicular to the first lateral direction. The first interconnect structure includes a first portion and a second portion electrically isolated from each other by a first dielectric structure. The semiconductor device includes a second interconnect structure, disposed between the first gate structure and the first interconnect structure, that electrically couples the first gate structure to the first portion of the first interconnect structure. The second interconnect structure includes a recessed portion that is substantially aligned with the first gate structure and the dielectric structure along a vertical direction.

Abstract: A method of making a semiconductor device includes manufacturing a first transistor over a first side of a substrate. The method further includes depositing a spacer material against a sidewall of the first transistor. The method further includes recessing the spacer material to expose a first portion of the sidewall of the first transistor. The method further includes manufacturing a first electrical connection to the transistor, a first portion of the electrical connection contacts a surface of the first transistor farthest from the substrate, and a second portion of the electrical connect contacts the first portion of the sidewall of the first transistor. The method further includes manufacturing a self-aligned interconnect structure (SIS) extending along the spacer material, wherein the spacer material separates a portion of the SIS from the first transistor, and the first electrical connection directly contacts the SIS.

Abstract: A device includes: a stack of semiconductor nanostructures; a gate structure wrapping around the semiconductor nanostructures, the gate structure extending in a first direction; a source/drain region abutting the gate structure and the stack in a second direction transverse the first direction; a contact structure on the source/drain region; a backside conductive trace under the stack, the backside conductive trace extending in the second direction; a first through via that extends vertically from the contact structure to a top surface of the backside dielectric layer; and a gate isolation structure that abuts the first through via in the second direction.

Abstract: A system for training transfer models that includes a memory and a processor coupling to each other. The memory stores instructions. The processor accesses the instructions to retrieve raw datasets from a first-type device and train a feature extraction model, a transfer feature and a preliminary classify model based on the raw datasets. The processor further transfers the feature extraction model, the transfer feature and the preliminary classify model to a host of a manufacturer so that the host directly applies the feature extraction model and the transfer feature, and retrains the preliminary classify model as an adapted classify model based on practical datasets.

Abstract: A computer-aided recognition system for treatment response of rectal cancer is provided to predict the probability of a pathological complete response (pCR) of a rectal cancer patient after preoperative neoadjuvant chemoradiotherapy. The system includes a random forest model having several decision trees. Each decision tree has at least a feature node. Each feature node generates two branches according to a feature threshold, wherein each branch connects to another feature node or corresponds to an elementary pCR predicting probability. The random forest model integrates the elementary pCR probability of each decision tree, so as to generate a final pCR probability.

Abstract: A system for training transfer models that includes a memory and a processor coupling to each other. The memory stores instructions. The processor accesses the instructions to retrieve raw datasets from a first-type device and train a feature extraction model, a transfer feature and a preliminary classify model based on the raw datasets. The processor further transfers the feature extraction model, the transfer feature and the preliminary classify model to a host of a manufacturer so that the host directly applies the feature extraction model and the transfer feature, and retrains the preliminary classify model as an adapted classify model based on practical datasets.

Abstract: A prediction system of tumor image-aided prediction of prognosis for patients with uterine cervical cancer is provided. The prediction system includes a data augmentation module and a deep convolution neural network model. The data augmentation module is used to apply a data expansion process of the image data, so as to generate a plurality of slice sets of the uterine cervical cancer tumor. The deep convolution neural network model is used to apply a feature analysis to the slices, so as to predict the prognosis of a patient after chemoradiotherapy.

Abstract: A computer-aided recognition system for treatment response of rectal cancer is provided to predict the probability of a pathological complete response (pCR) of a rectal cancer patient after preoperative neoadjuvant chemoradiotherapy. The system includes a random forest model having several decision trees. Each decision tree has at least a feature node. Each feature node generates two branches according to a feature threshold, wherein each branch connects to another feature node or corresponds to an elementary pCR predicting probability. The random forest model integrates the elementary pCR probability of each decision tree, so as to generate a final pCR probability.

Abstract: A prediction system of tumor image-aided prediction of prognosis for patients with uterine cervical cancer is provided. The prediction system includes a data augmentation module and a deep convolution neural network model. The data augmentation module is used to apply a data expansion process of the image data , so as to generate a plurality of slice sets of the uterine cervical cancer tumor. The deep convolution neural network model is used to apply a feature analysis to the slices, so as to predict the prognosis of a patient after chemoradiotherapy.

Abstract: This invention relates to an epoxy resin composition which comprises polyalkyl phenol resins or polyalkyl phenol epoxy resins or both, wherein the polyalkyl phenol resins are represented by formula (I): ##STR1## wherein Ra, Rb and Rc may be the same or different from each other and stand for C.sub.1-5 alkyl styryl or halogen respectively;n is an integer of 0 to 4;k is an integer of 0 to 3;l is an integer of 0 to 4; andm is an integer of 1 to 10;and the polyalkyl phenol epoxy resins, which are obtained by reacting polyalkyl phenol resins of formula (I) with epihalohydrins, are represented by formula (II): ##STR2## wherein G stands for ##STR3## while the other symbols are the same as those defined above. The epoxy resin composition of this invention is able to provide a reactivity at lower temperatures and provide lower viscosity when used as hardener resins, as compared with traditional phenolic novolacs.