Multi-label Learning Based on Kernel Extreme Learning Machine

Abstract

In recent years, with the increase of data scale, multi-label learning with large scale class labels has turned out to be the research hotspots. Due to the huge solution space, the problem becomes more complex. Therefore, we propose a multi-label algorithm based on kernel learning machine in this paper. Besides, the Cholesky matrix decomposition inverse method is adopted to calculate the network output weight of the kernel extreme learning machine. In particular, in terms of large matrix inverse problem, the large matrix is divided into small matrices for parallel computation through using matrix block method. Compared with several state-of -the-art algorithms on several benchmark data sets, results of the experiments show that the proposed algorithm makes a better performance with large scale class labels.

Text‐to‐3D Shape Generation

“…Most methods we described so far treat the output shape as an unstructured 3D representation. An emerging line of work attempts to produce structured 3D shapes based on parts and part connectivity [LLJ22; TYW23]. ShapeScaffolder [TYW23] first uses shape data to pretrain a structured shape decoder that can take a global latent vector encoding (consisting of both shape and color latents) and hierarchically decode it into parts.…”

“…ShapeGlot [AFH*19] and ShapeTalk [AHS*23] provided discriminative text that selected one object from multiple objects. However, as noted by Luo et al [LLJ22], this style of text data omits important information that is shared between the three objects (e.g. the object category) and is not suitable for aligning text and 3D spaces for 3D shape generation.…”

“…There is also growing interest in generating shapes based on programs [JGMR23; YXP*23]. A recent work that connects text to programs is Neural Shape Compiler [LLJ22]. Translating text to programs that can generate 3D shapes is an interesting direction for future investigation.…”

Background modeling from video sequences via online motion-aware RPCA

“…When the label y i is 1, the corresponding frame will be selected. Due to the removal of invalid data frames, this method can restore the background of the video more accurately [23][24] [26].…”