Building robust multimodal models are crucial for achieving reliable deployment in the wild. Despite its importance, less attention has been paid to identifying and improving the robustness of Multimodal Sentiment Analysis (MSA) models. In this work, we hope to address that by (i) Proposing simple diagnostic checks for modality robustness in a trained multimodal model. Using these checks, we find MSA models to be highly sensitive to a single modality, which creates issues in their robustness; (ii) We analyze well-known robust training strategies to alleviate the issues. Critically, we observe that robustness can be achieved without compromising on the original performance. We hope our extensive study-performed across five models and two benchmark datasets-and proposed procedures would make robustness an integral component in MSA research. Our diagnostic checks and robust training solutions are simple to implement and available at https://github.com/ declare-lab/MSA-Robustness

Background To assess the indications and complications of late amniocentesis and the advanced genetic test results in a tertiary university fetal medical medicine unit. Methods In this retrospective study, women that underwent amniocentesis at 24+ 0 to 39+ 4 weeks, between January 2014 and December 2019, were recruited. Indications, complications, genetic test results, and pregnancy outcomes were reported for each pregnancy and compared with those who underwent the traditional amniocentesis at 16+ 0 to 23+ 6 weeks (control group). Information was retrieved from patient medical records, checked by research staff, and analyzed. Results Of the 1287 women (1321 fetuses) included in the late amniocentesis group, late detected sonographic abnormalities (85.5%) were the most common indication. The overall incidence of preterm birth and intrauterine demise after amniocentesis were 2.5 and 1.3%, respectively. Sixty-nine fetuses with aneuploidy (5.3%) and seventy-two fetuses with pathogenic copy number variations (5.5%) were identified by chromosomal microarray analysis. The maximal diagnostic yield (70%) was in the subgroup of fetuses with the abnormal diagnostic test results, followed by abnormal NIPT results (35.7%) and multiple abnormalities (23.8%). And 35.4% of the pregnancies were finally terminated. Conclusions Due to the high detection rates of advanced genetic technologies and the safety of the invasive procedure (3.9% vs 4.0%), it is reasonable to recommend late amniocentesis as an effective and reliable method to detect late-onset fetal abnormalities. However, chromosomal microarray and whole-exome sequencing may result in uncertain results like variants of uncertain significance. Comprehensive genetic counseling is necessary.

Multi-hop question answering models based on knowledge graph have been extensively studied. Most existing models predict a single answer with the highest probability by ranking candidate answers. However, they are stuck in predicting all the right answers caused by the ranking method. In this paper, we propose a novel model that converts the ranking of candidate answers into individual predictions for each candidate, named heterogeneous knowledge graph based multi-hop and multi-answer model (HGMAN). HGMAN is capable of capturing more informative representations for relations assisted by our heterogeneous graph, which consists of multiple entity nodes and relation nodes. We rely on graph convolutional network for multi-hop reasoning and then binary classification for each node to get multiple answers. Experimental results on MetaQA dataset show the performance of our proposed model over all baselines.

No abstract

Objective We aimed to investigate the validity of noninvasive prenatal diagnosis (NIPD) based on direct haplotype phasing without the proband and its feasibility for clinical application in the case of Duchenne Muscular Dystrophy (DMD). Methods Thirteen singleton-pregnancy families affected by DMD were recruited.Firstly, we resolved maternal haplotypes for each family by performing targeted linked-read sequencing of their high molecular weight DNA, respectively. Then, we identified SNPs of the DMD gene in all carrier mothers and inferred the DMD gene mutation status of all fetuses. Finally, the fetal genotypes were further validated by using chorionic villus sampling. ResultsThe method of directly resolving maternal haplotype through targeted linked-read sequencing was smoothly performed in all participated families. The predicted mutational status of 13 fetuses was correct, which had been confirmed by invasive prenatal diagnosis. ConclusionDirect haplotyping of NIPD based on linked-read sequencing for DMD is accurate.