Due to the limited length and freely constructed sentence structures, it is a difficult classification task for short text classification, especially in multi-class classification. An efficient meta learning framework is proposed for twitter classification. The tweets are clustered into many sentence styles corresponding to new class labels. Thus, the original text classification task becomes few-shot learning task. When applying few-shot learning on benchmark datasets, the proposed method Meta-CNN achieves improvement in accuracy and F1 scores on multi-class twitter classification, and outweigh some traditional machine learning methods and a few deep learning approaches.

By virtue of kinetic proofreading theory, the puzzles of selectivity and specificity common in cellar activities are successfully explained. The chemical master equation is adopted to describe stochastic processes of kinetic proofreading models with two and three middle steps. Stochastic properties around the steady state are analyzed with the linear noise approximation under various parameters. Relations between noise density and parameters are revealed with numerical simulation. The result suggest that noises propagate and escalate as reactions proceed, furthermore multiple reaction rates determine noise intensity.

The authors propose a translation similarity model based on bilingual compositional semantics to integrate the bilingual semantic similarity feature into decoding process to improve translation quality. In the proposed model, monolingual compositional vectors for phrases are obtained at the source and target side respectively using distributional approach. These monolingual vectors are then projected onto the same semantic space and therefore transformed into bilingual compositional vectors. Base on this semantic space, translation similarity between source phrases and their corresponding target phrases is calculated. The similarities are integrated into the decoder as a new feature. Experiments on Chinese-to-English NIST06 and NIST08 test sets show that the proposed model significantly outperforms the baseline by 0.56 and 0.42 BLEU points respectively.

A random substitution model for the study of the evolution of virus genomes, which is able to determine self-consistently the ancestral genome, as well as the substitution matrix have been developed in this paper. The model is applied to analyze the evolution of four coronavirus genomes including the newly discovered SARS-CoV genome. The nucleotide sequences of the most recent common ancestor are determined at a set of synonymous substitution sites of several conserved protein-coding genes, the substitution matrix is found, and the total variation rate of each sequence since the divergent evolution of the common ancestor is determined. The results suggest that the SARS-CoV has undergone an independent evolution path from the beginning of the divergent evolution from the common ancestor (e.g. the SARS-CoV is as old as any other known coronaviruses). It can be speculated that the SARS-CoV may have stayed silent in its natural host long before it spreads the current large-scale epidemic in human.