Search

Question answering requiring numerical reasoning, which generally involves symbolic operations such as sorting, counting, and addition, is a challenging task. To address such a problem, existing mixture-of-experts (MoE)-based methods design several specific answer predictors to handle different types of questions and achieve promising performance. However, they ignore the modeling and exploitation of fine-grained reasoning-related operations to support numerical reasoning, encountering the inadequacy in reasoning capability and interpretability. To alleviate this issue, we propose OPERA, an operation-augmented numerical reasoning framework. Concretely, we systematically define a scalable operation set to model numerical reasoning. We first identify reasoning-related operations based on context and then softly execute them to imitate the answer reasoning procedure via an operation-aware cross-attention mechanism. Finally, we utilize the operation-augmented semantic representation of execution results to support answer prediction. We verify the effectiveness and generalization of OPERA in two scenarios with different knowledge sources and reasoning capabilities. Specifically, we conduct extensive experiments on two textual datasets, DROP and RACENum, and a table-text hybrid dataset TAT-QA. Experiment results show that OPERA outperforms previous strong methods on the DROP, RACENum, and TAT-QA datasets. Further, we statistically and visually analyze its interpretability.

Congenital acorea is a rare disease with the absence of a pupil in the eye. To date, only one family and two isolated cases with congenital acorea have been reported. The gene associated with acorea has not been identified. In this study, we recruited a Chinese family acorea-microphthalmia-cataract syndrome. By analyzing the whole-exome sequencing (WES) data of this Chinese family, we revealed the association of a novel heterozygous variant, NM_005267.5:c.137G>A (p.G46E) in the gap junction protein alpha 8 (GJA8) gene encoding connexin 50 or CX50, with familial acorea-microphthalmia-cataract syndrome. Additionally, another variant, NM_005267.5:c.151G>A (p.D51N) in GJA8, was identified to co-segregate with this syndrome in an unrelated Japanese family. Ectopic expression of p.G46E and p.D51N mutant GJA8genes in cultured cells caused protein mislocalization, suggesting that the p.G46E and p.D51N mutations in GJA8impaired the function of the gap junction channels. These results established GJA8as the first gene associated with familial acorea-microphthalmia-cataract syndrome.

The paper regards the 42 undergraduate colleges (the independent colleges are not included) of Henan province by the end of 2013 as the object of research, builds the composite indicator rank model of undergraduate colleges' educational resource allocation. On this basis, the authors use the score of integrated factor as the rank of composite indicator. In addition, they cluster to the educational resource allocation of 42 undergraduate colleges by use of cluster analysis under the new factors, and the 42 undergraduate colleges are divided into 4 classes. Finally, according to result of cluster analysis and factor scores, the authors put forward the related suggestion of optimizing educational resource allocation for each cluster colleges.

A dual-channel supply chain system composed of one manufacturer and one retailer is considered in this paper, which existed uncertainty demands in both distribution channels and capital constraint on retailer. It set up the profit model of manufacturer dominated dual-channel supply chain system, studied to the optimal pricing and inventory strategies of decentralized and centralized supply chain, obtained the optimal pricing and inventory strategy of the two kinds of supply chain system. The analysis to the optimal solution indicated that the demand uncertainty of two distribution channels, deferred-payment rate etc, are all have certain influence relation on the pricing strategies of manufacturer and retailer. Numerical experiment has showed the effectiveness of the conclusions.

By using an ELISA, increased levels of the soluble form (sgp130) of gp130, the IL‐6 signal transducer, were detected in the sera of various HTLV‐1‐associated conditions (HC, ATL, HAM) as compared to normal healthy individuals. Sgp130 levels seemed to be correlated with disease severity. The 94 KD of sgp130 was specifically precipitated in the sera of HTLV‐1‐infected patients as revealed by Western blot analysis. A reverse transcriptase (RT)‐polymerase chain reaction (PCR) method was used to detect the message for transmembrane (TM) lacking gp130 in mRNA isolated from peripheral blood mononuclear cells (PBMCs) of patients infected with or without HTLV‐1 and those of various hematopoietic cell lines. Two PCR products, 648 and 507 bp were observed in the PBMCs from HTLV‐1‐infected patients. But the 507 bp PCR product was not detected in the PBMCs from normal healthy individuals and HTLV‐1‐positive cell lines although the 648 bp product was equally expressed. A nucleotide sequence analysis of the 507 bp fragment showed deletion of the 141 bp at the region spanning from nucleotide 1702 (G) to 1842 (T) of the 648 bp product that matched completely with a conventional gp130 molecule. This deleted region was located upstream of the transmembrane (TM) domain, but not within the TM region itself. However, no frame shift was observed. These results indicate that the generation of sgp130 may not be due to an alternative splicing mechanism.