Yan Li

Bio: Yan Li is an academic researcher from Shenzhen Polytechnic. The author has contributed to research in topics: Feature extraction & Statistical classification. The author has an hindex of 9, co-authored 34 publications receiving 302 citations. Previous affiliations of Yan Li include South China Normal University.

Complementary resistive switching behaviors evolved from bipolar TiN/HfO2/Pt device

Abstract: In this letter, the dynamic evolution of TiN/HfO2/Pt device from bipolar resistive switching (BRS) to complementary resistive switching (CRS) was reported. The device exhibits the uniform BRS with long retention, good endurance, and self-compliance characteristics after the asymmetric two-step electroforming. However, BRS of the device eventually transforms to CRS after the transitional processes through controlling the compliance current. Meanwhile, the effective barrier height rises up accordingly as the device evolves from BRS to CRS. These superior resistive switching performances of TiN/HfO2/Pt device here can be elucidated in views of evolution of asymmetric filament. This work confirms the intimate correlation and discrepancy between BRS and CRS, and also indicates the potential application of TiN/HfO2/Pt device for future ultra-dense resistive random access memory.

A Generative Adversarial Gated Recurrent Unit Model for Precipitation Nowcasting

Abstract: Precipitation nowcasting is an important task in operational weather forecasts. The key challenge of the task is the radar echo map extrapolation. The problem is mainly solved by an optical-flow method in existing systems. However, the method cannot model rapid and nonlinear movements. Recently, a convolutional gated recurrent unit (ConvGRU) method is developed, which aims to model such movements based on deep learning techniques. Despite the promising performance, ConvGRU tends to yield blurring extrapolation images and fails to multi-modal and skewed intensity distribution. To overcome the limitations, we propose in this letter a generative adversarial ConvGRU (GA-ConvGRU) model. The model is composed of two adversarial learning systems, which are a ConvGRU-based generator and a convolution neural network-based discriminator. The two systems are trained by playing a minimax game. With the adversarial learning scheme, GA-ConvGRU can yield more realistic and more accurate extrapolation. Experiments on real data sets have been conducted and the results demonstrate that the proposed GA-ConvGRU significantly outperforms state-of-the-art extrapolation methods ConvGRU and optical flow.

DSKmeans: A new kmeans-type approach to discriminative subspace clustering

Abstract: Most of kmeans-type clustering algorithms rely on only intra-cluster compactness, i.e. the dispersions of a cluster. Inter-cluster separation which is widely used in classification algorithms, however, is rarely considered in a clustering process. In this paper, we present a new discriminative subspace kmeans-type clustering algorithm (DSKmeans), which integrates the intra-cluster compactness and the inter-cluster separation simultaneously. Different to traditional weighting kmeans-type algorithms, a 3-order tensor is constructed to evaluate the importance of different features in order to integrate the aforementioned two types of information. First, a new objective function for clustering is designed. To optimize the objective function, the corresponding updating rules for the algorithm are then derived analytically. The properties and performance of DSKmeans are investigated on several numerical and categorical data sets. Experimental results corroborate that our proposed algorithm outperforms the state-of-the-art kmeans-type clustering algorithms with respects to four metrics: Accuracy, RandIndex, Fscore and Normal Mutual Information(NMI).

Quantum conductance in MoS2 quantum dots-based nonvolatile resistive memory device

Abstract: In this work, nonvolatile bipolar resistive switching behaviors based on the MoS2 quantum dots (QDs) embedded in the insulating polymethylmethacrylate (PMMA) were reported with the device configuration of Au/PMMA/PMMA:MoS2 QDs/PMMA/fluorine doped tin-oxide. The device exhibits the reversible switching performances with the excellent read endurance and data retention capability. The related carrier transport behaviors were predominated by Schottky emission and Ohmic conductions in OFF and ON states, respectively. Importantly, a conductance quantization effect was evidently observed in this MoS2 QD-based memory device. Combined with the energy band evolution, these phenomena were elucidated in views of electrons trapping/de-trapping and quantum tunneling effects of nanoscale MoS2 QDs. This work also suggests the potential application of MoS2 QDs in next generation ultra-high-density data storage.

Feature extraction for high-resolution imagery based on human visual perception

Abstract: Feature extraction is highly important for classification of remote-sensing RS images. However, extraction of comprehensive spatial features from high-resolution imagery is still challenging, leading to many misclassifications in various applications. To address the problem, a shape-adaptive neighbourhood SAN technique is presented based on human visual perception. The SAN technique is an adaptive feature-extraction method that not only considers spectral feature information but also the spatial neighbourhood as well as the shape of features. The distinct advantage of this approach is that it can be adjusted to different feature sizes and shapes. Assessment experiments on a Systeme Pour l'Observation de la Terre 5 SPOT-5 image were conducted to perform classification of land use/land cover. Results showed that improvement with SAN features is not significant for supervised classifiers due to the spectral confusion problem that resulted from similar spectral signatures between farmland and green areas, but a particularly significant improvement is observed for the unsupervised classifier. For the unsupervised classification, the SAN features noticeably improved the overall accuracy from 0.58 to 0.86, and the kappa coefficient from 0.45 to 0.80, indicating promise in the application of SAN features in the auto-interpretation of RS images.

Nonvolatile Memories Based on Graphene and Related 2D Materials.

Abstract: The pervasiveness of information technologies is generating an impressive amount of data, which need to be accessed very quickly. Nonvolatile memories (NVMs) are making inroads into high-capacity storage to replace hard disk drives, fuelling the expansion of the global storage memory market. As silicon-based flash memories are approaching their fundamental limit, vertical stacking of multiple memory cell layers, innovative device concepts, and novel materials are being investigated. In this context, emerging 2D materials, such as graphene, transition metal dichalcogenides, and black phosphorous, offer a host of physical and chemical properties, which could both improve existing memory technologies and enable the next generation of low-cost, flexible, and wearable storage devices. Herein, an overview of graphene and related 2D materials (GRMs) in different types of NVM cells is provided, including resistive random-access, flash, magnetic and phase-change memories. The physical and chemical mechanisms underlying the switching of GRM-based memory devices studied in the last decade are discussed. Although at this stage most of the proof-of-concept devices investigated do not compete with state-of-the-art devices, a number of promising technological advancements have emerged. Here, the most relevant material properties and device structures are analyzed, emphasizing opportunities and challenges toward the realization of practical NVM devices.

Self‐Assembled Networked PbS Distribution Quantum Dots for Resistive Switching and Artificial Synapse Performance Boost of Memristors

Abstract: With the advent of the era of big data, resistive random access memory (RRAM) has become one of the most promising nanoscale memristor devices (MDs) for storing huge amounts of information. However, the switching voltage of the RRAM MDs shows a very broad distribution due to the random formation of the conductive filaments. Here, self-assembled lead sulfide (PbS) quantum dots (QDs) are used to improve the uniformity of switching parameters of RRAM, which is very simple comparing with other methods. The resistive switching (RS) properties of the MD with the self-assembled PbS QDs exhibit better performance than those of MDs with pure-Ga2 O3 and randomly distributed PbS QDs, such as a reduced threshold voltage, uniformly distributed SET and RESET voltages, robust retention, fast response time, and low power consumption. This enhanced performance may be attributed to the ordered arrangement of the PbS QDs in the self-assembled PbS QDs which can efficiently guide the growth direction for the conducting filaments. Moreover, biosynaptic functions and plasticity, are implemented successfully in the MD with the self-assembled PbS QDs. This work offers a new method of improving memristor performance, which can significantly expand existing applications and facilitate the development of artificial neural systems.

Combining Instance-Based Learning and Logistic Regression for Multilabel Classification.

Abstract: Multilabel classification is an extension of conventional classification in which a single instance can be associated with multiple labels. Recent research has shown that, just like for conventional classification, instance-based learning algorithms relying on the nearest neighbor estimation principle can be used quite successfully in this context. However, since hitherto existing algorithms do not take correlations and interdependencies between labels into account, their potential has not yet been fully exploited. In this paper, we propose a new approach to multilabel classification, which is based on a framework that unifies instance-based learning and logistic regression, comprising both methods as special cases. This approach allows one to capture interdependencies between labels and, moreover, to combine model-based and similarity-based inference for multilabel classification. As will be shown by experimental studies, our approach is able to improve predictive accuracy in terms of several evaluation criteria for multilabel prediction.

Semiconductor Quantum Dots for Memories and Neuromorphic Computing Systems

Abstract: The continued growth in the demand of data storage and processing has spurred the development of high-performance storage technologies and brain-inspired neuromorphic hardware. Semiconductor quantum dots (QDs) offer an appealing option for these applications since they combine excellent electronic/optical properties and structural stability and can address the requirements of low-cost, large-area, and solution-based manufactured technologies. Here, we focus on the development of nonvolatile memories and neuromorphic computing systems based on QD thin-film solids. We introduce recent advances of QDs and highlight their unique electrical and optical features for designing future electronic devices. We also discuss the advantageous traits of QDs for novel and optimized memory techniques in both conventional flash memories and emerging memristors. Then, we review recent advances in QD-based neuromorphic devices from artificial synapses to light-sensory synaptic platforms. Finally, we highlight major challenges for commercial translation and consider future directions for the postsilicon era.