Abstract: The present disclosure relates to a dual lens manual synchronous focusing structure, which can drive the movable frame to move through the manual driving assembly during focusing, while the first rack on the movable frame drives the first outer gear ring to rotate, the second rack drives the second outer gear ring to rotate, the first outer gear ring turns the first lens barrel to rotate, and the first lens barrel is screwing connected with the first base, thus, the first lens barrel is close to or far away from the first photosensitive chip, the second outer gear ring turns the second lens barrel to rotate, and the second lens barrel is screwing connected with the second base, thus the second lens barrel is close to or far from the second photosensitive chip, so that the focusing of the first lens assembly and the second lens assembly can be synchronized.

Abstract: Methods, systems, and apparatuses include training a graph neural network. Content items are received for a user of an online system, the content items including skills. An input graph is generated using the content items, the input graph including nodes and edges linking the nodes. The input graph is sampled using a source node and skills to generate a first computational graph. The input graph is sampled using a target node and skills to generate a second computational graph. A source node embedding is generated by encoding the first computational graph. A target node embedding is generated by encoding the second computational graph. A prediction score is calculated by decoding the source node embedding and the target node embedding. Weights of the graph neural network are updated using the prediction score.

Abstract: A naked-eye three-dimensional imaging device includes an objective lens module including a first and objective lenses arranged at intervals of a pupil distance, a display module including a first display screen for displaying a shooting image of the first objective lens, and a second display screen for displaying a shooting image of the second objective lens, an observation member including a first and second observation chambers respectively connected to the first and second display screens, an eyepiece module including a first eyepiece and a second eyepiece respectively arranged at eyepiece ends of the first and second observation chambers, and an eyepiece focusing module between the eyepiece module and the observation member, and including a first focusing member and a second focusing member for respectively adjusting a distance between the first eyepiece and the first display screen, and a distance between the second eyepiece and the second display screen.

Abstract: A naked-eye three-dimensional VR camera includes a first lens module and a second lens module arranged at intervals of a pupil distance to simulate human eyes and form a pupil distance difference between captured first and second images. When the first image through one eye and the second image through the other eye are observed from the display component through eyepiece member, a naked-eye three-dimensional effect can be directly formed. A montage module is connected to a control member or a memory card assembly to read files of the first and second images, and splice the first and second images according to a time axis. Because the first and second images are synchronously captured, so that the first and second images are combined to form a three-dimensional image file with the pupil distance difference, thereby playing the captured images on other naked-eye three-dimensional devices (such as VR glasses, VR helmets).