arXiv subjects

Bin Lu

Publications and source records attributed to Bin Lu.

32 records · Page 2Linked to original sources

arXiv2023

Capacity Enhancement of n-GHZ State Super-dense Coding Channels by Purification and Quantum Neural Network

A super-dense coding protocol based on the n-GHZ state is proposed to enable the two communicating parties to choose the number of transmitted code words according to their demand and to adapt the quantum super-dense coding protocol to multiple transmitted code word scenarios. A method is proposed that combines entanglement purification and Quantum Neural Network (QNN) to improve the channel capacity of super-dense coding. By simulating a realistic quantum communication noise environment in the Cirq platform, the effect of purification and QNN on the enhancement of fidelity and channel capacity in super-dense coding communication scenarios with different dimensions under unitary and non-unitary noise conditions is analyzed. The experimental results show that the channel capacity of super-dense coding is improved in different degrees when purification and QNN are applied separately, and the combination of purification and QNN has a superimposed effect on the channel capacity enhancement of super-dense coding, and the enhancement effect is more significant in different dimensions.

quant-ph↗

arXiv2023

Graph Out-of-Distribution Generalization with Controllable Data Augmentation

Graph Neural Network (GNN) has demonstrated extraordinary performance in classifying graph properties. However, due to the selection bias of training and testing data (e.g., training on small graphs and testing on large graphs, or training on dense graphs and testing on sparse graphs), distribution deviation is widespread. More importantly, we often observe \emph{hybrid structure distribution shift} of both scale and density, despite of one-sided biased data partition. The spurious correlations over hybrid distribution deviation degrade the performance of previous GNN methods and show large instability among different datasets. To alleviate this problem, we propose \texttt{OOD-GMixup} to jointly manipulate the training distribution with \emph{controllable data augmentation} in metric space. Specifically, we first extract the graph rationales to eliminate the spurious correlations due to irrelevant information. Secondly, we generate virtual samples with perturbation on graph rationale representation domain to obtain potential OOD training samples. Finally, we propose OOD calibration to measure the distribution deviation of virtual samples by leveraging Extreme Value Theory, and further actively control the training distribution by emphasizing the impact of virtual OOD samples. Extensive studies on several real-world datasets on graph classification demonstrate the superiority of our proposed method over state-of-the-art baselines.

cs.LG↗

arXiv2022

Geometer: Graph Few-Shot Class-Incremental Learning via Prototype Representation

With the tremendous expansion of graphs data, node classification shows its great importance in many real-world applications. Existing graph neural network based methods mainly focus on classifying unlabeled nodes within fixed classes with abundant labeling. However, in many practical scenarios, graph evolves with emergence of new nodes and edges. Novel classes appear incrementally along with few labeling due to its newly emergence or lack of exploration. In this paper, we focus on this challenging but practical graph few-shot class-incremental learning (GFSCIL) problem and propose a novel method called Geometer. Instead of replacing and retraining the fully connected neural network classifer, Geometer predicts the label of a node by finding the nearest class prototype. Prototype is a vector representing a class in the metric space. With the pop-up of novel classes, Geometer learns and adjusts the attention-based prototypes by observing the geometric proximity, uniformity and separability. Teacher-student knowledge distillation and biased sampling are further introduced to mitigate catastrophic forgetting and unbalanced labeling problem respectively. Experimental results on four public datasets demonstrate that Geometer achieves a substantial improvement of 9.46% to 27.60% over state-of-the-art methods.

cs.AI↗

arXiv2022

Spatio-Temporal Graph Few-Shot Learning with Cross-City Knowledge Transfer

Spatio-temporal graph learning is a key method for urban computing tasks, such as traffic flow, taxi demand and air quality forecasting. Due to the high cost of data collection, some developing cities have few available data, which makes it infeasible to train a well-performed model. To address this challenge, cross-city knowledge transfer has shown its promise, where the model learned from data-sufficient cities is leveraged to benefit the learning process of data-scarce cities. However, the spatio-temporal graphs among different cities show irregular structures and varied features, which limits the feasibility of existing Few-Shot Learning (\emph{FSL}) methods. Therefore, we propose a model-agnostic few-shot learning framework for spatio-temporal graph called ST-GFSL. Specifically, to enhance feature extraction by transfering cross-city knowledge, ST-GFSL proposes to generate non-shared parameters based on node-level meta knowledge. The nodes in target city transfer the knowledge via parameter matching, retrieving from similar spatio-temporal characteristics. Furthermore, we propose to reconstruct the graph structure during meta-learning. The graph reconstruction loss is defined to guide structure-aware learning, avoiding structure deviation among different datasets. We conduct comprehensive experiments on four traffic speed prediction benchmarks and the results demonstrate the effectiveness of ST-GFSL compared with state-of-the-art methods.

cs.LG↗

arXiv2022

PP-ShiTu: A Practical Lightweight Image Recognition System

In recent years, image recognition applications have developed rapidly. A large number of studies and techniques have emerged in different fields, such as face recognition, pedestrian and vehicle re-identification, landmark retrieval, and product recognition. In this paper, we propose a practical lightweight image recognition system, named PP-ShiTu, consisting of the following 3 modules, mainbody detection, feature extraction and vector search. We introduce popular strategies including metric learning, deep hash, knowledge distillation and model quantization to improve accuracy and inference speed. With strategies above, PP-ShiTu works well in different scenarios with a set of models trained on a mixed dataset. Experiments on different datasets and benchmarks show that the system is widely effective in different domains of image recognition. All the above mentioned models are open-sourced and the code is available in the GitHub repository PaddleClas on PaddlePaddle.

cs.CV↗

arXiv2021

PP-OCRv2: Bag of Tricks for Ultra Lightweight OCR System

Optical Character Recognition (OCR) systems have been widely used in various of application scenarios. Designing an OCR system is still a challenging task. In previous work, we proposed a practical ultra lightweight OCR system (PP-OCR) to balance the accuracy against the efficiency. In order to improve the accuracy of PP-OCR and keep high efficiency, in this paper, we propose a more robust OCR system, i.e. PP-OCRv2. We introduce bag of tricks to train a better text detector and a better text recognizer, which include Collaborative Mutual Learning (CML), CopyPaste, Lightweight CPUNetwork (LCNet), Unified-Deep Mutual Learning (U-DML) and Enhanced CTCLoss. Experiments on real data show that the precision of PP-OCRv2 is 7% higher than PP-OCR under the same inference cost. It is also comparable to the server models of the PP-OCR which uses ResNet series as backbones. All of the above mentioned models are open-sourced and the code is available in the GitHub repository PaddleOCR which is powered by PaddlePaddle.

cs.CV↗

arXiv2021

PP-LCNet: A Lightweight CPU Convolutional Neural Network

We propose a lightweight CPU network based on the MKLDNN acceleration strategy, named PP-LCNet, which improves the performance of lightweight models on multiple tasks. This paper lists technologies which can improve network accuracy while the latency is almost constant. With these improvements, the accuracy of PP-LCNet can greatly surpass the previous network structure with the same inference time for classification. As shown in Figure 1, it outperforms the most state-of-the-art models. And for downstream tasks of computer vision, it also performs very well, such as object detection, semantic segmentation, etc. All our experiments are implemented based on PaddlePaddle. Code and pretrained models are available at PaddleClas.

cs.CV↗

arXiv2021

100% renewable electricity in Japan

Japan has committed to carbon neutrality by 2050. Emissions from the electricity sector amount to 42% of the total. Solar photovoltaics (PV) and wind comprise three quarters of global net capacity additions because of low and falling prices. This provides an opportunity for Japan to make large reductions in emissions while also reducing its dependence on energy imports. This study shows that Japan has 14 times more solar and offshore wind resources than needed to supply 100% renewable electricity. A 40 year hourly energy balance model is presented of Japan's electricity system using historical data. Pumped hydro energy storage, high voltage interconnection and dispatchable capacity (hydro, biomass and hydrogen energy) are included to balance variable generation and demand. Differential evolution is used to find the least-cost solution under various constraints. The levelized cost of electricity is found to be USD 86 per MWh for a PV-dominated system, and USD 110 per MWh for a wind-dominated system. These costs can be compared with the average system prices on the spot market in Japan of USD 102 per MWh. In summary, Japan can be self-sufficient for electricity supply at competitive costs.

physics.soc-ph↗

arXiv2020

A zero-carbon, reliable and affordable energy future in Australia

Australia has one of the highest per capita consumption of energy and emissions of greenhouse gases in the world. It is also the global leader in rapid per capita annual deployment of new solar and wind energy, which is causing the country's emissions to decline. Australia is located at low-moderate latitudes along with three quarters of the global population. These factors make the Australian experience globally significant. In this study, we model a fully decarbonised electricity system together with complete electrification of heating, transport and industry in Australia leading to an 80% reduction in greenhouse gas emissions. An energy supply-demand balance is simulated based on long-term (10 years), high-resolution (half-hourly) meteorological and energy demand data. A significant feature of this model is that short-term off-river energy storage and distributed energy storage are utilised to support the large-scale integration of variable solar and wind energy. The results show that high levels of energy reliability and affordability can be effectively achieved through a synergy of flexible energy sources; interconnection of electricity grids over large areas; response from demand-side participation; and mass energy storage. This strategy represents a rapid and generic pathway towards zero-carbon energy futures within the Sunbelt.

eess.SY↗

arXiv2017

Secure Numerical and Logical Multi Party Operations

We derive algorithms for efficient secure numerical and logical operations using a recently introduced scheme for secure multi-party computation~\cite{sch15} in the semi-honest model ensuring statistical or perfect security. To derive our algorithms for trigonometric functions, we use basic mathematical laws in combination with properties of the additive encryption scheme in a novel way. For division and logarithm we use a new approach to compute a Taylor series at a fixed point for all numbers. All our logical operations such as comparisons and large fan-in AND gates are perfectly secure. Our empirical evaluation yields speed-ups of more than a factor of 100 for the evaluated operations compared to the state-of-the-art.

cs.CR↗

arXiv2013

A 2-chain can interlock with an open 10-chain

It is an open problem, posed in \cite{SoCG}, to determine the minimal $k$ such that an open flexible $k$-chain can interlock with a flexible 2-chain. It was first established in \cite{GLOSZ} that there is an open 16-chain in a trapezoid frame that achieves interlocking. This was subsequently improved in \cite{GLOZ} to establish interlocking between a 2-chain and an open 11-chain. Here we improve that result once more, establishing interlocking between a 2-chain and a 10-chain. We present arguments that indicate that 10 is likely the minimum.

cs.CG↗

arXiv2006

The Kauffman Polynomials of 2-bridge Knots

The 2-bridge knots are a family of knots with bridge number 2. In this paper, we compute the Kauffman polynomials of 2-bridge knots using the Kauffman skein theory and linear algebra techniques. Our calculation can be easily carried out using Mathematica, Maple, Mathcad, etc.

math.GT↗

arXiv2001

The Kauffman polynomials of generalized Hopf links

Following the recent work by Chan, and by Morton and Hadji on the Homflypt polynomials of some generalized Hopf links, we investigate the Kauffman polynomials of generalized Hopf links. By studying the Kauffman skein module of the solid torus S^1\times D^2, we establish a similar skein map on the Kauffman skein module of S^1\times D^2 which has distinct eigenvalues. Furthermore we are able to calculate the Kauffman polynomials of some specific generalized Hopf links.

math.GT↗

arXiv2001

On the Kauffman skein modules

Let k be a subring of the field of rational functions in α, s which contains α^{1}, α^{-1}, s^{1}, s^{-1}, . Let M be a compact oriented 3-manifold, and let K(M) denote the Kauffman skein module of M over k. Then K(M) is the free k-module generated by isotopy classes of framed links in M modulo the Kauffman skein relations. In the case of k={Q}(α, s), the field of rational functions in α, s, we give a basis for the Kauffman skein module of the solid torus and a basis for the relative Kauffman skein module of the solid torus with two points on the boundary. We then show that K(S^{1} \times S^2) is generated by the empty link, i.e., K(S^{1} \times S^2)=k.

math.GT↗