Abstract: The present disclosure relates to systems, methods, and non-transitory computer-readable media that utilize a text-image alignment loss to train a diffusion model to generate digital images from input text. In particular, in some embodiments, the disclosed systems generate a prompt noise representation form a text prompt with a first text concept and a second text concept using a denoising step of a diffusion neural network. Further, in some embodiments, the disclosed systems generate a first concept noise representation from the first text concept and a second concept noise representation from the second text concept. Moreover, the disclosed systems combine the first and second concept noise representation to generate a combined concept noise representation. Accordingly, in some embodiments, by comparing the combined concept noise representation and the prompt noise representation, the disclosed systems modify parameters of the diffusion neural network.

Abstract: A method for learning a representation of a sequence of frames includes encoding, via an encoder network, the sequence of frames to obtain a set of feature maps and extracting, a motion-guided slot learning mechanism, mid-level features from the set of feature maps. The method further includes quantizing the mid-level features via a vector quantization process to obtain a set of tokens, and decoding, via a decoder network, the tokens to obtain a reconstructed sequence of frames. The method still further includes optimizing a combination of a reconstruction loss and a motion loss to train the encoder and decoder networks.

Abstract: A method of compositional feature representation learning for video understanding is described. The method includes individually processing a sequence of video frames received as an input of a feature map network to generate a plurality of feature maps. The method also includes binding the plurality of feature maps to a fixed set of slot variables using an attention model according to a motion segmentation signal. The method further includes combining slot states corresponding to the fixed set of slot variables into a combined feature map. The method also includes decoding the combined feature map to form a reconstructed sequence of video frames, in which objects discovered in the reconstructed sequence of video frames are identified.

Abstract: An integrated electric booster braking system with a pedal force compensation function, comprising a booster motor, a first gear, a second gear, a lead screw, a brake master cylinder, a pedal push rod, a master cylinder push rod, a fluid storage tank, a hydraulic control unit and an electric control unit, wherein first gear and second gear are assembled in a housing, first gear is engaged with the second gear, the booster motor is connected with the first gear and drives the first gear to rotate, the first gear drives the second gear to rotate during rotation, the second gear is in threaded connection with the lead screw, the second gear drives the lead screw to move during rotation, the lead screw is of a hollow structure, and the rear end of the lead screw is abutted with the front end of a first piston in the brake master cylinder.

Abstract: An integrated electric booster braking system with a pedal force compensation function, comprising a booster motor, a first gear, a second gear, a lead screw, a brake master cylinder, a pedal push rod, a master cylinder push rod, a fluid storage tank, a hydraulic control unit and an electric control unit, wherein first gear and second gear are assembled in a housing, first gear is engaged with the second gear, the booster motor is connected with the first gear and drives the first gear to rotate, the first gear drives the second gear to rotate during rotation, the second gear is in threaded connection with the lead screw, the second gear drives the lead screw to move during rotation, the lead screw is of a hollow structure, and the rear end of the lead screw is abutted with the front end of a first piston in the brake master cylinder.