End-to-End Feature-Aware Label Space Encoding for Multilabel Classification With Many Classes

Abstract

To make the problem of multilabel classification with many classes more tractable, in recent years, academia has seen efforts devoted to performing label space dimension reduction (LSDR). Specifically, LSDR encodes high-dimensional label vectors into low-dimensional code vectors lying in a latent space, so as to train predictive models at much lower costs. With respect to the prediction, it performs classification for any unseen instance by recovering a label vector from its predicted code vector via a decoding process. In this paper, we propose a novel method, namely End-to-End Feature-aware label space Encoding (EFE), to perform LSDR. Instead of requiring an encoding function like most previous works, EFE directly learns a code matrix formed by code vectors of the training instances in an end-to-end manner. Another distinct property of EFE is its feature awareness attributable to the fact that the code matrix is learned by jointly maximizing the recoverability of the label space and the predictability of the latent space. Based on the learned code matrix, EFE further trains predictive models to map instance features into code vectors, and also learns a linear decoding matrix for efficiently recovering the label vector of any unseen instance from its predicted code vector. Theoretical analyses show that both the code matrix and the linear decoding matrix in EFE can be efficiently learned. Moreover, similar to previous works, EFE can be specified to learn an encoding function. And it can also be extended with kernel tricks to handle nonlinear correlations between the feature space and the latent space. Comprehensive experiments conducted on diverse benchmark data sets with many classes show consistent performance gains of EFE over the state-of-the-art methods.

Hybrid Multi-Label Classification Model for Medical Applications Based on Adaptive Synthetic Data and Ensemble Learning

“…Lin et al [12] proposed that Label Space Dimension Reduction (LSDR) is performed using a unique technique called End-to-End Feature-Aware Label Space Encoding (E 2FE).…”

Wallpaper Texture Generation and Style Transfer Based on Multi-Label Semantics

“…Giunchiglia et al [5] recently proposed C-HMCNN(h) which exploits the hierarchy information in order to produce predictions coherent with the constraint. Lin et al [6] proposed E 2 F E which directly learns a feature-aware code matrix via jointly maximizing the recoverability of the label space and the predictability of the latent space, and gains performance improvements over other state-of-the-art LSDR methods.…”

“…We suspect this is because our style transfer model is trained on art works with artistic styles such as blurring, dim and sharp contrasts, hence cannot use the same aesthetic scoring rules as ordinary images. The average value of the score distributions for all wallpaper images is mainly concentrated in [3,6]. After investigation, we suppose this is because NIMA is trained on the AVA dataset [43], and the images in this dataset are high-definition photos.…”

Research on Data Science Ensembles for Covid-19 Detection and Length of Stay Prediction

“…It is a supervised statistical technique which is used to check the inter-relation between a set of input data or variables. These two techniques help to pre-process data to transform the data in a model-acceptable format to get a desired result [22]. IV.…”