Experimental demonstration of adversarial examples in learning topological phases.

Classification and identification of different phases and the transitions between them is a central task in condensed matter physics. Machine learning, which has achieved dramatic success in a wide range of applications, holds the promise to bring unprecedented perspectives for this challenging task. However, despite the exciting progress made along this direction, the reliability of machine-learning approaches in experimental settings demands further investigation. Here, with the nitrogen-vacancy center platform, we report a proof-of-principle experimental demonstration of adversarial examples in learning topological phases. We show that the experimental noises are more likely to act as adversarial perturbations when a larger percentage of the input data are dropped or unavailable for the neural network-based classifiers. We experimentally implement adversarial examples which can deceive the phase classifier with a high confidence, while keeping the topological properties of the simulated Hopf insulators unchanged. Our results explicitly showcase the crucial vulnerability aspect of applying machine learning techniques in experiments to classify phases of matter, which can benefit future studies in this interdisciplinary field.

Observation of Non-Hermitian Topology with Nonunitary Dynamics of Solid-State Spins.

Non-Hermitian topological phases exhibit a number of exotic features that have no Hermitian counterparts, including the skin effect and breakdown of the conventional bulk-boundary correspondence. Here, we implement the non-Hermitian Su-Schrieffer-Heeger Hamiltonian, which is a prototypical model for studying non-Hermitian topological phases, with a solid-state quantum simulator consisting of an electron spin and a ^{13}C nuclear spin in a nitrogen-vacancy center in a diamond. By employing a dilation method, we realize the desired nonunitary dynamics for the electron spin and map out its spin texture in the momentum space, from which the corresponding topological invariant can be obtained directly. From the measured spin textures with varying parameters, we observe both integer and fractional winding numbers. The non-Hermitian topological phase with fractional winding number cannot be continuously deformed to any Hermitian topological phase and is intrinsic to non-Hermitian systems. Our result paves the way for further exploiting and understanding the intriguing properties of non-Hermitian topological phases with solid-state spins or other quantum simulation platforms.

ROS-responsive organosilica nanocarrier for the targeted delivery of metformin against cancer with the synergistic effect of hypoglycemia.

The controllable degradation of silica nanoparticles in anticancer therapy remains challenging. Here, we offer the first report that a thioketal (TK)-bond-containing bridged organoalkoxysilane has been synthesized. This allows for the fabrication of reactive oxygen species (ROS)-sensitive, degradable, bridged silsesquioxane nanoparticles (BS-NPs). These TK-bridged BS-NPs have a uniform size of 50 nm and are able to encapsulate a small molecule drug - metformin - using a reverse micro-emulsion method. After surface modification with a targeting peptide (RGD), these metformin-loaded BS-NPs exhibited a homologous tumor aggregation ability, leading to the efficient transport of metformin into the tumor cells. When combined with a clinically feasible fasting therapy, the RGD-decorated, metformin-loaded, ROS-responsive degradable BS-NPs remarkably increased the tumor sensitivity to metformin by 10 times compared with free metformin. The synergistic effects of metformin-loaded BS-NPs and fasting-induced hypoglycemia were verified through in vitro and in vivo experiments. This effect occurred by down-regulating the expression of pro-survival proteins pGSK3ß and MCL-1. Collectively, these results demonstrate that the ROS-sensitive organosilica nanocarrier is a promising nanoplatform for drug delivery and provides an alternative approach for the combinatorial therapy of metformin and fasting therapy.

[Effects of planting broadleaf trees and Moso bamboo on soil carbon mineralization and microbial community structure]. / ç§æ¤é˜”å¶æ ‘ç§å’Œæ¯›ç«¹å¯¹åœŸå£¤æœ‰æœºç¢³çŸ¿åŒ–ä¸Žå¾®ç”Ÿç‰©ç¾¤è½ç»“æž„çš„å½±å“.

We conducted a pot experiment to investigate the effects of planting broadleaf tree species (i.e., Cinnamomum camphora, Schima superba, and Quercus glauca) and Moso bamboo (Phyllostachys edulis) on soil carbon mineralization and microbial community structure. The rates of soil carbon mineralization were measured via alkali trapping method. The structural and functional diversity of soil bacterial and fungal communities were analyzed by terminal restriction fragment length polymorphism (T-RFLP) and real-time quantitative PCR techniques. The soil planted with Moso bamboo exhibited a significantly higher carbon mineralization rate and labile carbon content than those in the soils planted with broadleaf tree species. The underground biomass of Moso bamboo was higher than that of broadleaf tree species. The soil bacterial communities were more sensitive than fungal communities to the planting of different plant species . Moreover, soil fungal diversity of Moso bamboo was distinctly different from that of broadleaf tree species. Compared to the diversity of soil bacterial communities, the diversity of soil fungal communities was more closely related with soil pH, organic carbon content, and carbon mineralization. In comparison to the broadleaf tree species, the Moso bamboo planting could substantially increase soil organic carbon minera-lization, which was affected mainly by the soil fungal community structure.

The complete mitochondrial genomes sequences of Asio flammeus and Asio otus and comparative analysis.

The complete mitochondrial genomes of Asio flammeus and Asio otus were sequenced and found to span 18858 bp and 18493 bp, respectively. It is surprising to find the former to be the largest among all avian mitochondrial genomes sequenced so far. The two genomes have very similar gene order with that of Gallus gallus, neither contains the pseudo control region, but both have a single extra base, namely Cytidine, at position 174 in ND3 gene. The control regions of Asio flammeus and Asio otus' mitochondrial genomes span 3288 bp and 2926 bp respectively, which are the longest among vertebrates except for Myxine glutinosa and contribute to the large size of two genomes. The 3' end of the control region of Asio flammeus and Asio otus contains many tandemly repeated sequences, which are highly similar to a putative control element, i.e. Mt5, and may form stable stem-loop secondary structures. Such repeated sequences probably play an important role in regulating transcription and replication of mitochondrial genome. Our results may provide important clues for uncovering the origin and evolution mechanisms of mitochondrion genome.

[Compariative study of mitochondrial tRNA gene sequence and secondary structure among fifteen Predatory birds].

Three major clusters of mitochondrial tRNA genes (tRNA(Ile) -tRNA(Gln) -tRNA(Met), tRNA(Trp)- tRNA(Ala) -tRNA(Asn)- tRNA(CYs) -tRNA(Tyr) and tRNA(His) tRNA(Ser)(AGY) -tRNA(Leu)(CUN) from 13 species of Predatory birds were amplified and sequenced. The length of these tRNA clusters was similar among species (212 approximately 214 bp, 353 approximately 362 bp, 205 approximately 208 bp, respectively), and 47% of the sequences were variable, 67% of which were involved in the loop regions. The stem regions were relatively conserved, and the variable base pairs were under the restriction of compensatory changes or G-U wobble pairing which could be regarded as mechanisms for maintaining a stable secondary structure. Maximum-parsimony (MP) and Neighbor joining (NJ) phylogenetic trees were constructed using all the tRNA gene sequences or stein-forming nucleotides with Caprimulgus indicus as outgroup. We found that the bootstrap values for branches of trees using the tRNA sequences were commonly higher than the others, therefore the phylogenetic relationship of Predatory birds reflected by these data may be closer to the truth. Phylogenetic analyses indicated that Accipitridae was closer to Strigidae instead of Falconidae, and the classification of Tytonidae was different from the conclusion from the previously morphological and DNA-DNA hybridization studies. By comparing the secondary structure among taxa we found that the characters of nucleotide insertions and deletions in some tRNA genes have synapomorphies, suggesting that these characters may be useful for resolving the phylogenetic relationship of different families in Predatory birds with higher phylogenetic performance.