Dual Information Enhanced Multiview Attributed Graph Clustering.

Multiview attributed graph clustering is an important approach to partition multiview data based on the attribute characteristics and adjacent matrices from different views. Some attempts have been made in using graph neural network (GNN), which have achieved promising clustering performance. Despite this, few of them pay attention to the inherent specific information embedded in multiple views. Meanwhile, they are incapable of recovering the latent high-level representation from the low-level ones, greatly limiting the downstream clustering performance. To fill these gaps, a novel dual information enhanced multiview attributed graph clustering (DIAGC) method is proposed in this article. Specifically, the proposed method introduces the specific information reconstruction (SIR) module to disentangle the explorations of the consensus and specific information from multiple views, which enables graph convolutional network (GCN) to capture the more essential low-level representations. Besides, the contrastive learning (CL) module maximizes the agreement between the latent high-level representation and low-level ones and enables the high-level representation to satisfy the desired clustering structure with the help of the self-supervised clustering (SC) module. Extensive experiments on several real-world benchmarks demonstrate the effectiveness of the proposed DIAGC method compared with the state-of-the-art baselines.

Contrastive Multiview Attribute Graph Clustering With Adaptive Encoders.

Multiview attribute graph clustering aims to cluster nodes into disjoint categories by taking advantage of the multiview topological structures and the node attribute values. However, the existing works fail to explicitly discover the inherent relationships in multiview topological graph matrices while considering different properties between the graphs. Besides, they cannot well handle the sparse structure of some graphs in the learning procedure of graph embeddings. Therefore, in this article, we propose a novel contrastive multiview attribute graph clustering (CMAGC) with adaptive encoders method. Within this framework, the adaptive encoders concerning different properties of distinct topological graphs are chosen to integrate multiview attribute graph information by checking whether there exists high-order neighbor information or not. Meanwhile, the number of layers of the GCN encoders is selected according to the prior knowledge related to the characteristics of different topological graphs. In particular, the feature-level and cluster-level contrastive learning are conducted on the multiview soft assignment representations, where the union of the first-order neighbors from the corresponding graph pairs is regarded as the positive pairs for data augmentation and the sparse neighbor information problem in some graphs can be well dealt with. To the best of our knowledge, it is the first time to explicitly deal with the inherent relationships from the interview and intraview perspectives. Extensive experiments are conducted on several datasets to verify the superiority of the proposed CMAGC method compared with the state-of-the-art methods.

Spectrally flat mid-infrared supercontinuum pumped by a high power 2 µm noise-like pulse.

We demonstrate spectrally flat high-power mid-infrared supercontinuum (MIR SC) generation with record-breaking power of 33.1 W and power conversion efficiency of 75.06%. It is pumped by a 2 µm master oscillator power amplifier system consisting of a figure-8 mode-locked noise-like pulse seed laser and dual-stage Tm-doped fiber amplifiers with repetition rate of 4.08 MHz. Through cascading a piece of ZBLAN fiber with 13.5 µm large core diameter by direct-low-loss fusion splicing, SCs with spectral ranges of 1.9-3.68 µm, 1.9-3.84 µm, 1.9-4.02 µm and average powers of 33.1 W, 29.8 W, 25.9 W are generated. To the best of our knowledge, all of them have achieved the highest output power under the same condition of MIR spectrum range. This high-power all-fiber MIR SC laser system has relatively simple architecture, high efficiency and flat spectrum, demonstrating the advantages of 2 µm noise-like pulse pump in high-power MIR SC generation.

Cross-Subject Tinnitus Diagnosis Based on Multi-Band EEG Contrastive Representation Learning.

Electroencephalogram (EEG) is an important technology to explore the central nervous mechanism of tinnitus. However, it is hard to obtain consistent results in many previous studies for the high heterogeneity of tinnitus. In order to identify tinnitus and provide theoretical guidance for the diagnosis and treatment, we propose a robust, data-efficient multi-task learning framework called Multi-band EEG Contrastive Representation Learning (MECRL). In this study, we collect resting-state EEG data from 187 tinnitus patients and 80 healthy subjects to generate a high-quality large-scale EEG dataset on tinnitus diagnosis, and then apply the MECRL framework on the generated dataset to obtain a deep neural network model which can distinguish tinnitus patients from the healthy controls accurately. Subject-independent tinnitus diagnosis experiments are conducted and the result shows that the proposed MECRL method is significantly superior to other state-of-the-art baselines and can be well generalized to unseen topics. Meanwhile, visual experiments on key parameters of the model indicate that the high-classification weight electrodes of tinnitus' EEG signals are mainly distributed in the frontal, parietal and temporal regions. In conclusion, this study facilitates our understanding of the relationship between electrophysiology and pathophysiology changes of tinnitus and provides a new deep learning method (MECRL) to identify the neuronal biomarkers in tinnitus.

High-gain single-frequency Tm3+-doped ZBLAN fiber amplifier at 2.33 µm.

We report here, to the best of our knowledge, the first high-gain, single-frequency Tm3+-doped fiber amplifier operating at the 2.3-µm band with conventional ground-state pumping transition (3H6→3H4) at 793 nm. The gain fiber is an 8.5-m-long ZBLAN fiber with a Tm3+ doping concentration of 3 mol.%, and the seed is a single-frequency distributed feedback diode laser operated at 2331.9 nm. A gain up to 24.1 dB is generated for ∼14 W of launched pump power with the maximum output power of 246 mW. The competitive 3F4 →3H6 laser transition at ∼2 µm is also investigated, and the prospects for further power scaling are discussed.

2.1 µm, 1.52 µJ square-wave noise-like pulse from an all-fiber figure-of-9 Ho-doped oscillator and single-stage amplifier.

A 2.1 µm, high energy square-wave noise-like pulse (NLP) in an all-fiber Ho-doped fiber laser is proposed, which consists of an oscillator and a single-stage amplifier. In the figure-of-9 oscillator, mode-locking is achieved based on the nonlinear amplifying loop mirror, employing a long gain fiber to provide sufficient gain in 2.1 µm band and optimizing the cavity length to obtain maximum pulse energy output. With appropriate pump power and polarization state, the oscillator emits a 175.1 nJ square-wave NLP with center wavelength of 2102.2 nm and spike width of 540 fs. The 3-dB spectral width and pulse envelope width are 11.2 nm and 6.95 ns, respectively. The single-stage amplifier employs a bi-directional pump scheme. After amplification, 5.8 W NLP with a slope efficiency of 56.8% is obtained. The pulse energy of NLP is scaled to 1.52 µJ, which is the highest pulse energy of NLP at 2.1 µm to the best of our knowledge. The obtained high-energy square-wave NLP-fiber laser has great potential in mid-infrared laser generation.

Pulse evolution and multi-pulse state of coherently coupled polarization domain walls in a fiber ring laser.

Pulse evolution and multi-pulse state of coherently coupled polarization domain walls (PDW) is experimentally demonstrated in a novel fiber ring laser. Versatile pulse shapes benefit by wide range moving of PDW in the weakly birefringent fiber. The 8.6 m short-cavity structure is more compact and accessible based on a 976 nm pump with nearly zero negative dispersion (-0.0002 ps2). Besides, multi-pulse patterns such as PDW splitting, harmonic mode-locking, and periodic soliton collision are also observed under larger net negative dispersion (-3.09 ps2) and 151m-longer cavity. This is the first demonstration of coherently coupled PDW in a fiber laser using a bandpass filter and the formation of coherently coupled PDW is ascribed to the BPF's force filtering.

Direct Observation of Folding Energy Landscape of RNA Hairpin at Mechanical Loading Rates.

By applying a controlled mechanical load using optical tweezers, we measured the diffusive barrier crossing in a 49 nt long P5ab RNA hairpin. We find that in the free-energy landscape the barrier height (G‡) and transition distance (x‡) are dependent on the loading rate (r) along the pulling direction, x, as predicted by Bell. The barrier shifted toward the initial state, whereas ΔG‡ reduced significantly from 50 to 5 kT, as r increased from 0 to 32 pN/s. However, the equilibrium work (ΔG) during strand separation, as estimated by Crook's fluctuation theorem, remained unchanged at different rates. Previously, helix formation and denaturation have been described as two-state (F ↔ U) transitions for P5ab. Herein, we report three intermediate states I1, I, and I2 located at 4, 11, and 16 nm respectively, from the folded conformation. The intermediates were observed only when the hairpin was subjected to an optimal r, 7.6 pN/s. The results indicate that the complementary strands in P5ab can zip and unzip through complex routes, whereby mismatches act as checkpoints and often impose barriers. The study highlights the significance of loading rates in force-spectroscopy experiments that are increasingly being used to measure the folding properties of biomolecules.