Lu Chen
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Lu Chen has a total of 10 co-authors. Here are the top co-authors by number of shared publications:
Hideyuki Koto
3 shared publications- Masayuki Murata
Osaka University
3 shared publications Hidetoshi Yokota
2 shared publicationsNagao Ogino
2 shared publications- Shin'Ichi Arakawa
Osaka University
2 shared publications Masayuki Murata
Department of Information Science and Technology, Osaka University, Osaka, Japan
1 shared publication
Latest Publications
- Open Access
On invariant tori in some reversible systems
In the present paper, we consider the following reversible system \begin{equation*} \begin{cases} \dot{x}=\omega_0+f(x,y),\\ \dot{y}=g(x,y), \end{cases} \end{equation*} where $x\in\mathbf{T}^{d}$, $y\backsim0\in \mathbf{R}^{d}$, $\omega_0$ is Diophantine, $f(x,y)=O(y)$, $g(x,y)=O(y^2)$ and $f$, $g$ are reversible with respect to the involution G: $(x,y)\mapsto(-x,y)$, that is, $f(-x,y)=f(x,y)$, $g(-x,y)=-g(x,y)$. We study the accumulation of an analytic invariant torus $\Gamma_0$ of the reversible system with Diophantine frequency $\omega_0$ by other invariant tori. We will prove that if the Birkhoff normal form around $\Gamma_0$ is 0-degenerate, then $\Gamma_0$ is accumulated by other analytic invariant tori, the Lebesgue measure of the union of these tori being positive and the density of the union of these tori at $\Gamma_0$ being one. We will also prove that if the Birkhoff normal form around $\Gamma_0$ is $j$-degenerate ($1\leq j\leq d-1$) and condition (1.6) is satisfied, then through $\Gamma_0$ there passes an analytic subvariety of dimension $d+j$ foliated into analytic invariant tori with frequency vector $\omega_0$. If the Birkhoff normal form around $\Gamma_0$ is $d-1$-degenerate, we will prove a stronger result, that is, a full neighborhood of $\Gamma_0$ is foliated into analytic invariant tori with frequency vectors proportional to $\omega_0$.Open AccessOn invariant tori in some reversible systems
In the present paper, we consider the following reversible system \begin{equation*} \begin{cases} \dot{x}=\omega_0+f(x,y),\\ \dot{y}=g(x,y), \end{cases} \end{equation*} where $x\in\mathbf{T}^{d}$, $y\backsim0\in \mathbf{R}^{d}$, $\omega_0$ is Diophantine, $f(x,y)=O(y)$, $g(x,y)=O(y^2)$ and $f$, $g$ are reversible with respect to the involution G: $(x,y)\mapsto(-x,y)$, that is, $f(-x,y)=f(x,y)$, $g(-x,y)=-g(x,y)$. We study the accumulation of an analytic invariant torus $\Gamma_0$ of the reversible system with Diophantine frequency $\omega_0$ by other invariant tori. We will prove that if the Birkhoff normal form around $\Gamma_0$ is 0-degenerate, then $\Gamma_0$ is accumulated by other analytic invariant tori, the Lebesgue measure of the union of these tori being positive and the density of the union of these tori at $\Gamma_0$ being one. We will also prove that if the Birkhoff normal form around $\Gamma_0$ is $j$-degenerate ($1\leq j\leq d-1$) and condition (1.6) is satisfied, then through $\Gamma_0$ there passes an analytic subvariety of dimension $d+j$ foliated into analytic invariant tori with frequency vector $\omega_0$. If the Birkhoff normal form around $\Gamma_0$ is $d-1$-degenerate, we will prove a stronger result, that is, a full neighborhood of $\Gamma_0$ is foliated into analytic invariant tori with frequency vectors proportional to $\omega_0$. - Open Access
- 26 November 2020
- Published by Springer Nature in Applied Network Science
- Vol. 5 (1) , 1-24
Enhancing network modularity to mitigate catastrophic forgetting
Catastrophic forgetting occurs when learning algorithms change connections used to encode previously acquired skills to learn a new skill. Recently, a modular approach for neural networks was deemed necessary as learning problems grow in scale and complexity since it intuitively should reduce learning interference by separating functionality into physically distinct network modules. However, an algorithmic approach is difficult in practice since it involves expert design and trial and error. Kashtan et al. finds that evolution under an environment that changes in a modular fashion leads to the spontaneous evolution of a modular network structure. In this paper, we aim to solve the reverse problem of modularly varying goal (MVG) to obtain a highly modular structure that can mitigate catastrophic forgetting so that it can also apply to realistic data. First, we confirm that a configuration with a highly modular structure exists by applying an MVG against a realistic dataset and confirm that this neural network can mitigate catastrophic forgetting. Next, we solve the reverse problem, that is, we propose a method that can obtain a highly modular structure able to mitigate catastrophic forgetting. Since the MVG-obtained neural network can relatively maintain the intra-module elements while leaving the inter-module elements relatively variable, we propose a method to restrict the inter-module weight elements so that they can be relatively variable against the intra-module ones. From the results, the obtained neural network has a highly modular structure and can learn an unlearned goal faster than without this method.
- 26 November 2020
- Published by Springer Nature in Applied Network Science
- Vol. 5 (1) , 1-24
Open AccessEnhancing network modularity to mitigate catastrophic forgetting
Catastrophic forgetting occurs when learning algorithms change connections used to encode previously acquired skills to learn a new skill. Recently, a modular approach for neural networks was deemed necessary as learning problems grow in scale and complexity since it intuitively should reduce learning interference by separating functionality into physically distinct network modules. However, an algorithmic approach is difficult in practice since it involves expert design and trial and error. Kashtan et al. finds that evolution under an environment that changes in a modular fashion leads to the spontaneous evolution of a modular network structure. In this paper, we aim to solve the reverse problem of modularly varying goal (MVG) to obtain a highly modular structure that can mitigate catastrophic forgetting so that it can also apply to realistic data. First, we confirm that a configuration with a highly modular structure exists by applying an MVG against a realistic dataset and confirm that this neural network can mitigate catastrophic forgetting. Next, we solve the reverse problem, that is, we propose a method that can obtain a highly modular structure able to mitigate catastrophic forgetting. Since the MVG-obtained neural network can relatively maintain the intra-module elements while leaving the inter-module elements relatively variable, we propose a method to restrict the inter-module weight elements so that they can be relatively variable against the intra-module ones. From the results, the obtained neural network has a highly modular structure and can learn an unlearned goal faster than without this method.
- 1 February 2016
- Published by Elsevier in Computer Communications
- Vol. 76, 101-110
An evolvable network design approach with topological diversity
- 1 February 2016
- Published by Elsevier in Computer Communications
- Vol. 76, 101-110
An evolvable network design approach with topological diversity
- 1 December 2014
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 1-7
Separation of Background and Foreground Traffic Based on Periodicity Analysis
This paper proposes a novel approach to separating background (BG) and foreground (FG) traffic based on periodicity analysis. As BG traffic is commonly periodically generated by applications, this trait is leveraged to effectively detect BG traffic. Concretely, the Period Candidate Array (PCA) approach is proposed to extract only necessary information from long and sparse traffic flows, hence quickly detects the flows' periodicity with low computational cost. The PCA works directly with "on-site'' traffic without depending on historical data as in machine learning methods. As a result, the proposed approach can be immediately applied to the real world traffic management systems. In addition, the PCA properly works with latency-included traffic affected by network delays. Experimental results reveal the effectiveness and efficiency of the PCA compared to the conventional methods in terms of computational cost, memory usage, and independence to historical data.
- 1 December 2014
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 1-7
Separation of Background and Foreground Traffic Based on Periodicity Analysis
This paper proposes a novel approach to separating background (BG) and foreground (FG) traffic based on periodicity analysis. As BG traffic is commonly periodically generated by applications, this trait is leveraged to effectively detect BG traffic. Concretely, the Period Candidate Array (PCA) approach is proposed to extract only necessary information from long and sparse traffic flows, hence quickly detects the flows' periodicity with low computational cost. The PCA works directly with "on-site'' traffic without depending on historical data as in machine learning methods. As a result, the proposed approach can be immediately applied to the real world traffic management systems. In addition, the PCA properly works with latency-included traffic affected by network delays. Experimental results reveal the effectiveness and efficiency of the PCA compared to the conventional methods in terms of computational cost, memory usage, and independence to historical data.
- 1 April 2014
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 789-794
Designing an evolvable network with topological diversity
As environments surrounding the Internet become more changeable, a design approach is needed that requires less equipment to scale up networks against traffic growth resulting from various environmental changes. Here, we propose an evolvable network design approach in which network equipment is deployed without predetermined purpose rather than for a preplanned purpose. We use mutual information on node degree to measure the topological diversity of networks, and maximize topological diversity in the network design by minimizing the mutual information. Evaluations show that, compared to networks based on an ad hoc design method, networks constructed by our design approach can re-use already installed equipment in different environments.
- 1 April 2014
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 789-794
Designing an evolvable network with topological diversity
As environments surrounding the Internet become more changeable, a design approach is needed that requires less equipment to scale up networks against traffic growth resulting from various environmental changes. Here, we propose an evolvable network design approach in which network equipment is deployed without predetermined purpose rather than for a preplanned purpose. We use mutual information on node degree to measure the topological diversity of networks, and maximize topological diversity in the network design by minimizing the mutual information. Evaluations show that, compared to networks based on an ad hoc design method, networks constructed by our design approach can re-use already installed equipment in different environments.