No abstract

Natural language generation (NLG) is an essential component of task-oriented dialogue systems. Despite the recent success of neural approaches for NLG, they are typically developed for particular domains with rich annotated training examples. In this paper, we study NLG in a low-resource setting to generate sentences in new scenarios with handful training examples. We formulate the problem from a meta-learning perspective, and propose a generalized optimization-based approach (Meta-NLG) based on the well-recognized model-agnostic metalearning (MAML) algorithm. Meta-NLG defines a set of meta tasks, and directly incorporates the objective of adapting to new low-resource NLG tasks into the meta-learning optimization process. Extensive experiments are conducted on a large multidomain dataset (MultiWoz) with diverse linguistic variations. We show that Meta-NLG significantly outperforms other training procedures in various low-resource configurations. We analyze the results, and demonstrate that Meta-NLG adapts extremely fast and well to low-resource situations.

Natural language generation (NLG) is an essential component of task-oriented dialog systems. Despite the recent success of neural approaches for NLG, they are typically developed in an offline manner for particular domains. To better fit real-life applications where new data come in a stream, we study NLG in a "continual learning" setting to expand its knowledge to new domains or functionalities incrementally. The major challenge towards this goal is catastrophic forgetting, meaning that a continually trained model tends to forget the knowledge it has learned before. To this end, we propose a method called ARPER (Adaptively Regularized Prioritized Exemplar Replay) by replaying prioritized historical exemplars, together with an adaptive regularization technique based on Elastic Weight Consolidation. Extensive experiments to continually learn new domains and intents are conducted on MultiWoZ-2.0 to benchmark ARPER with a wide range of techniques. Empirical results demonstrate that ARPER significantly outperforms other methods by effectively mitigating the detrimental catastrophic forgetting issue.

Candida tropicalis is a globally distributed human pathogenic yeast, especially prevalent in tropical and sub-tropical regions. Over the last several decades, a large number of studies have been published on the genetic diversity and molecular epidemiology of C. tropicalis from different parts of the world. However, the global pattern of genetic variation remains largely unknown. Here we analyzed the published multilocus sequence data at six loci for 876 isolates from 16 countries representing five continents. Our results showed that 280 of the 2677 (10.5%) analyzed nucleotides were polymorphic, resulting in a mean of 82 (a range of 38–150) genotypes per locus and a total of 633 combined diploid sequence types (DSTs). Among these, 93 combined DSTs were shared by 336 strains, including 10 by strains from different continents. Analysis of Molecular Variance (AMOVA) showed that 89% of the observed genetic variations were found within regional and national populations while < 10% was due to among-country separations. Pairwise geographic population analyses showed overall low but statistically significant genetic differentiation between most geographic populations, with the Singaporean and Indian populations being the most distinct from other populations. However, the Mantel test showed no significant correlation between genetic distance and geographic distance among the geographic populations. Consistent with high genetic variation within and limited variations among geographic populations, results from STRUCTURE analyses showed that the 876 isolates could be grouped into 15 genetic clusters, with each cluster having a broad geographic distribution. Together, our results suggest frequent gene flows among certain regional, national, and continental populations of C. tropicalis , resulting in abundant regional and national genetic diversities of this important human fungal pathogen.

Background: Myostatin (MSTN) encodes a negative regulator of skeletal muscle mass that might have applications for promoting muscle growth in livestock. In this study, we aimed to test whether targeted MSTN editing, mediated by transcription activator-like effector nucleases (TALENs), is a viable approach to create myostatin-modified goats (Capra hircus). Results: We obtained a pair of TALENs (MTAL-2) that could recognize and cut the targeted MSTN site in the goat genome. Fibroblasts from pedigreed goats were co-transfected with MTAL-2, and 272 monoclonal cell strains were confirmed to have mono-or bi-allelic mutations in MSTN. Ten cell strains with different genotypes were used as donor cells for somatic cell nuclear transfer, which produced three cloned kids (K179/MSTN −/− , K52-2/MSTN +/− , and K52-1/MSTN +/+ ).