Kai‐Wei Chang

LIBLINEAR is an open source library for large-scale linear classification. It supports logistic regres-sion and linear support vector machines. We provide easy-to-use command-line tools and library calls for users and developers. Comprehensive documents are available for both beginners and advanced users. Experiments demonstrate that LIBLINEAR is very efficient on large sparse data sets.

In many applications, data appear with a huge number of instances as well as features. Linear Support Vector Machines (SVM) is one of the most popular tools to deal with such large-scale sparse data. This paper presents a novel dual coordinate descent method for linear SVM with L1-and L2-loss functions. The proposed method is simple and reaches an ε-accurate solution in O(log(1/ε)) iterations. Experiments indicate that our method is much faster than state of the art solvers such as Pegasos, TRON, SVMperf, and a recent primal coordinate descent implementation.

Kernel techniques have long been used in SVM to handle linearly inseparable problems by transforming data to a high dimensional space, but training and testing large data sets is often time consuming. In contrast, we can efficiently train and test much larger data sets using linear SVM without kernels. In this work, we apply fast linear-SVM methods to the explicit form of polynomially mapped data and investigate implementation issues. The approach enjoys fast training and testing, but may sometimes achieve accuracy close to that of using highly nonlinear kernels. Empirical experiments show that the proposed method is useful for certain large-scale data sets. We successfully apply the proposed method to a natural language processing (NLP) application by improving the testing accuracy under some training/testing speed requirements.

INTRODUCTION: Pluripotent stem cells are increasingly used to build therapeutic models, including the transplantation of neural progenitors derived from human embryonic stem cells (hESCs). Recently, long non-coding RNAs (lncRNAs), including delta-like homolog 1 gene and the type III iodothyronine deiodinase gene (DLK1-DIO3) imprinted locus-derived maternally expressed gene 3 (MEG3), were found to be expressed during neural development. The deregulation of these lncRNAs is associated with various neurological diseases. The imprinted locus DLK1-DIO3 encodes abundant non-coding RNAs (ncRNAs) that are regulated by differential methylation of the locus. We aim to study the correlation between the DLK1-DIO3-derived ncRNAs and the capacity of hESCs to differentiate into neural lineages. METHODS: We classified hESC sublines into MEG3-ON and MEG3-OFF based on the expression levels of MEG3 and its downstream microRNAs as detected by quantitative reverse transcription-polymerase chain reaction (qRT-PCR). A cDNA microarray was used to analyze the gene expression profiles of hESCs. To investigate the capacity of neural differentiation in MEG3-ON and MEG3-OFF hESCs, we performed neural lineage differentiation followed by neural lineage marker expression and neurite formation analyses via qRT-PCR and immunocytochemistry, respectively. MEG3-knockdown via small interfering RNA (siRNA) and small hairpin RNA (shRNA) was used to investigate the potential causative effect of MEG3 in regulating neural lineage-related gene expression. RESULTS: DLK1-DIO3-derived ncRNAs were repressed in MEG3-OFF hESCs compared with those in the MEG3-ON hESCs. The transcriptome profile indicated that many genes related to nervous system development and neural-type tumors were differentially expressed in MEG3-OFF hESCs. Three independent MEG3-knockdown assays using different siRNA and shRNA constructs consistently resulted in downregulation of some neural lineage genes. Lower expression levels of stage-specific neural lineage markers and reduced neurite formation were observed in neural lineage-like cells derived from MEG3-OFF-associated hESCs compared with those in the MEG3-ON groups at the same time points after differentiation. CONCLUSIONS: Repression of ncRNAs derived from the DLK1-DIO3 imprinted locus is associated with reduced neural lineage differentiation potential in hESCs.

Linear support vector machines (SVM) are useful for classifying large-scale sparse data. Problems with sparse features are common in applications such as document classification and natural language processing. In this paper, we propose a novel coordinate descent algorithm for training linear SVM with the L2-loss function. At each step, the proposed method minimizes a one-variable sub-problem while fixing other variables. The sub-problem is solved by Newton steps with the line search technique. The procedure globally converges at the linear rate. As each sub-problem involves only values of a corresponding feature, the proposed approach is suitable when accessing a feature is more convenient than accessing an instance. Experiments show that our method is more efficient and stable than state of the art methods such as Pegasos and TRON.