Knowledge Graph Representation Learning Based on Automatic Network Search for Link Prediction

Link prediction,also known as Knowledge Graph Completion(KGC),is the common task in Knowledge Graphs(KGs)to predict missing connections between entities.Most existing methods focus on designing shallow,scalable models,which have less expressive than deep,multi-layer models.Furthermore,most operations like addition,matrix multiplications or factorization are handcrafted based on a few known relation patterns in several wellknown datasets,such as FB15k,WN18,etc.However,due to the diversity and complex nature of real-world data distribution,it is inherently difficult to preset all latent patterns.To address this issue,we proposeKGE-ANS,a novel knowledge graph embedding framework for general link prediction tasks using automatic network search.KGEANS can learn a deep,multi-layer effective architecture to adapt to different datasets through neural architecture search.In addition,the general search spacewe designed is tailored forKGtasks.We performextensive experiments on benchmark datasets and the dataset constructed in this paper.The results show that our KGE-ANS outperforms several state-of-the-art methods,especially on these datasets with complex relation patterns.

Facile and diverse logic circuits based on dumbbell DNA-templated fluorescent copper nanoclusters and S1 nuclease detection

DNA-based logic gates promote the development of molecular computing and show enormous potential in the fields of nanotechnology and biotechnology. Dumbbell oligonucleotides(DNA) with poly-thymine(poly-T) loops and a nicked random double strand have been demonstrated to be an efficient template for the formation of fluorescent copper nanoclusters(Cu NCs) in our previous work. Herein, a new platform technology is presented with which to construct molecular logic gates by employing Cu NCs probe as a basic output generator, coupling of functional nucleases as the inputs. Two dumbbell DNAs are used with the difference in stem length(8 bp and 16 bp, respectively). The degradation of DNA templates can be tuned by various nucleic acid enzymes, single-stranded nuclease(S1), double-stranded specific nuclease(DSN), E. coli DNA ligase, exonucleases Ⅰ and Ⅲ. Briefly, S1 can digest both DNA templates, while the cleavage ability of DSN will be resistant by the short stem of SS-DNA(short-stem DNA). Exonuclease Ⅰ and Ⅲ can degrade these two nicked DNA templates, which are inhibited due to the ligation of E. coli DNA ligase. With this novel strategy, a set of logic gates is successfully constructed at the molecular level,including “YES”, “PASS 0”, “OR”, “INHIBIT”, which take the advantages of no label, easy operation, fast speed, high efficiency and low cost. Furthermore, S1 nuclease, as the biomarker of numerous carcinogens,is selectively detected in the range of 0.05–50 U/m L with the detection limit of 0.005 U/m L(1×10^(−6)U)based on this platform.