Search

Most existing classical artificial neural networks (ANN) are designed as a tree structure to imitate neural networks. In this paper, we argue that the connectivity of a tree is not sufficient to characterize a neural network. The nodes of the same level of a tree cannot be connected with each other, i.e., these neural unit cannot share information with each other, which is a major drawback of ANN. Although ANN has been significantly improved in recent years to more complex structures, such as the directed acyclic graph (DAG), these methods also have unidirectional and acyclic bias for ANN. In this paper, we propose a method to build a bidirectional complete graph for the nodes in the same level of an ANN, which yokes the nodes of the same level to formulate a neural module. We call our model as YNN in short. YNN promotes the information transfer significantly which obviously helps in improving the performance of the method. Our YNN can imitate neural networks much better compared with the traditional ANN. In this paper, we analyze the existing structural bias of ANN and propose a model YNN to efficiently eliminate such structural bias. In our model, nodes also carry out aggregation and transformation of features, and edges determine the flow of information. We further impose auxiliary sparsity constraint to the distribution of connectedness, which promotes the learned structure to focus on critical connections. Finally, based on the optimized structure, we also design small neural module structure based on the minimum cut technique to reduce the computational burden of the YNN model. This learning process is compatible with the existing networks and different tasks. The obtained quantitative experimental results reflect that the learned connectivity is superior to the traditional NN structure., Comment: arXiv admin note: text overlap with arXiv:2008.08261 by other authors

This work proposes a new perspective on the structure of Neural Networks (NNs). Traditional Neural Networks are typically tree-like structures for convenience, which can be predefined or learned by NAS methods. However, such a structure can not facilitate communications between nodes at the same level or signal transmissions to previous levels. These defects prevent effective collaboration, restricting the capabilities of neural networks. It is well-acknowledged that the biological neural system contains billions of neural units. Their connections are far more complicated than the current NN structure. To enhance the representational ability of neural networks, existing works try to increase the depth of the neural network and introduce more parameters. However, they all have limitations with constrained parameters. In this work, we introduce a synchronous graph-based structure to establish a novel way of organizing the neural units: the Neural Modules. This framework allows any nodes to communicate with each other and encourages neural units to work collectively, demonstrating a departure from the conventional constrained paradigm. Such a structure also provides more candidates for the NAS methods. Furthermore, we also propose an elegant regularization method to organize neural units into multiple independent, balanced neural modules systematically. This would be convenient for handling these neural modules in parallel. Compared to traditional NNs, our method unlocks the potential of NNs from tree-like structures to general graphs and makes NNs be optimized in an almost complete set. Our approach proves adaptable to diverse tasks, offering compatibility across various scenarios. Quantitative experimental results substantiate the potential of our structure, indicating the improvement of NNs.

Ba(Ni$_{1-x}$Co$_x$)$_2$As$_2$ is a structural homologue of the pnictide high temperature superconductor, Ba(Fe$_{1-x}$Co$_x$)$_2$As$_2$, in which the Fe atoms are replaced by Ni. Superconductivity is highly suppressed in this system, reaching a maximum $T_c$ = 2.3 K, compared to 24 K in its iron-based cousin, and the origin of this $T_c$ suppression is not known. Using x-ray scattering, we show that Ba(Ni$_{1-x}$Co$_x$)$_2$As$_2$ exhibits a unidirectional charge density wave (CDW) at its triclinic phase transition. The CDW is incommensurate, exhibits a sizable lattice distortion, and is accompanied by the appearance of $\alpha$ Fermi surface pockets in photoemission [B. Zhou et al., Phys. Rev. B 83, 035110 (2011)], suggesting it forms by an unconventional mechanism. Co doping suppresses the CDW, paralleling the behavior of antiferromagnetism in iron-based superconductors. Our study demonstrates that pnictide superconductors can exhibit competing CDW order, which may be the origin of $T_c$ suppression in this system.

X-ray diffraction was employed to study the evolution of the charge density wave (CDW) in CuxTiSe2 as a function of copper intercalation in order to clarify the relationship between the CDW and superconductivity. The results show a CDW incommensuration arising at an intercalation value coincident with the onset of superconductivity at around x=0.055(5). Additionally, it was found that the charge density wave persists to higher intercalant concentrations than previously assumed, demonstrating that the CDW does not terminate inside the superconducting dome. A charge density wave peak was observed in samples up to x=0.091(6), the highest copper concentration examined in this study. The phase diagram established in this work suggests that charge density wave incommensuration may play a role in the formation of the superconducting state., Comment: 5 pages, 4 figures