Questions related to Network Architecture
centralized or decentralized? using service brokers or Kerberos? Blockchain? infrastructures? softwares/tools? recommended papers to read? please discuss...
I am running a survey regarding metro-area network architecture and have some analytics to share through this link: https://docs.google.com/forms/d/1kJnEjukNDGC4JARuhgBI0HUUrQ8UNA-G71Aa6heaK8U/viewanalytics
The link will be active today only, until 8 pm CET.
Can you help me find a nice solution to plot different CNN architectures automatically?
At the moment, I have a 3 head 1D-CNN, with 2 convolutional layers, 2 max-pooling layers, and 2 fully connected layers.
I used 3 heads to have different resolutions (kernel size) on the same signals.
Hello, i am searching for some new innovation points in 6G about networking slice recently.
I want to know what updates will be made to network slicing in 6G, which cannot do well in 5G.
(Maybe about architectures, methodologies, algorithms, and combination with other 6G new technologies.)
Can you recommend me some good ideas, innovation points or papers( and white papers), thank you.
I have created a neural network architecture by considering the data for one study area. Have performed grid-search to tune almost all the hyperparameters of the model. I want to use the same network architecture (without any hyperparameter tuning) on the data of another geographical area for the same topic/ problem.
What I am intended to do is, take the same architecture and train the model with the data of the new area. Later I want to show that even without changing the model architecture the same can be applied to other geographical areas also.
Is this valid and understandable work?
I could not find any resource or paper online that has done a similar thing. All I could find is transfer learning. But I don't think this can be termed as transfer learning, or is it?
A repository to aggregate a set of reproducible and reconfigurable codes and notebooks for testing various task placement policies for edge and fog server networks.
The simulation models are based on two of the most powerful python-based simulation modelling frameworks, namely salabim and simpy.
The systems that are modelled cover the basic types of task placement problems in edge computing servers. The models are useful for managing a network of edge computing servers.
There are animation and GUI options which are indispensable in simulation modelling.
The uploaded models are templates for building simulation models for a variety of edge network policies. Researchers who research task and server placement problems in edge computing would find the models useful.
What kind of network architecture is used if i have multiple X,Y,Z coorinates(for ex: 4 coordinates (x1,y,z1), (x2,y2,z2),(x3,y3,z3), (x4,y4,z4) ) & few other numeric features and the target is numeric.
The coordinates are related to spatial information(XYZ drone coordinates are related to multiple drone coordinates). Is there any specific network that preserves the spatial meaning or just a simple regression network is fine? I am using the PyTorch framework.
Dear fellow scientists,
if there are some among you who need help with small to medium-sized Deep Learning projects in some kind of medical field (e.g. setting up network architecture, defining proper loss functions, training and optimizing, etc.): Please feel free to get in touch with me.
I am a fourth-year medical student with experience in medical (physiological) research and substantial training and experience in designing and conducting small/medium Deep Learning projects in a medical context.
I am looking forward to hearing from you!
I am trying to model a system which consists of a second order linear part (for example a set of mass spring damper with variable parameters for spring and damper) and a non-linear part which is a cascade of a derivative, a time delay, a static non-linearity (e.g. half-wave rectifier) and a low-pass system. I want to use neural networks for this purpose. I have tried time delay neural networks which were not successful. I am now trying to use RNNs (Recurrent Neural Networks). Since I am relatively new to this subject, I was wondering which network architecture are suitable for this purpose? (I have to use a limited number of parameters)
I am working with extinction models in interaction networks using the bipartite R-package.
I can determine the order in which species will be removed based on abundance and degree or random.
abundance <- second.extinct(Safariland, participant = "lower", method = "abundance", nrep = 10,
details = FALSE, ext.row=NULL, ext.col=NULL)
However, I need and am not able to elaborate a vector to determine a different extinction sequence, using the ‘external’ method.
Nowadays there are different types of available access networks which are used by end users to connect to the internet. Thus, the users must be provided with seamless network connectivity to stay connected while moving around from one place to another. This seamless network connectivity is achieved by connecting different types of networks which is called heterogeneous internetworking. By integrating different network technologies into one common heterogeneous network architecture, they can coexist and interoperate with each other and improve network performance in term of Quality of Service (QoS).
is it possible to use deep learning with a small size of samples ?
and what is the state-of-the-art neural network architecture when using it with microarray dataset?
Our team is developing a keyword spotting(kws) model used in detecting keywords spoken by humans. Our main approaches include some state-of-the-art neural network architectures such as CNN, Depthwise CNN and ResNet. In our experiments, these models achieve >98% accuracies in detecting specific keywords in both training and testing dataset(randomly selected). However, when we place a mircophone to collect sounds under office environment to evaluate real-time performance of the models, all of them show high mistriggered rate, i.e., the models believe a target keyword is spoken while it is not. How do we optimize the models to reduce the mistriggered rate in kws? Any papers/documents/tutorials are definitely welcomed to share!
I am using HNOCS omnet++. I made basic connection changes in topology and I am getting this error
<!> Error in module (cCompoundModule) Mesh (id=1) during network setup: No submodule `router' to be connected, at G:\software\hnoc\hnocs.release\hnocs.01_02_13\src\ topologies\Mesh.ned:93.
Can anyone explain what does this error mean and how to remove it?
Thanks in advance
I came across a neural network architecture which can train the dataset on two different CNN's of varying input shapes and layers and finally merging them together. The architecture looks similar to the attached image. Any useful links or codes could be helpful .
I'm looking to train a neural network to map between two different vector spaces. I have a set of mappings between vectors in both spaces so this would be a supervised training problem. The output layer would be d nodes where d is the dimensionality of the output vector space. But I'm not sure if there are certain architectures I need to use to perform this regression task.
Hi everyone, hope you are doing well. Please guide me, I want to set my network like this: that sink node on receiving a packet start sending packets back on the same route but when it receives a packet from other route then sink node starts stop its prior transmission and start transmission again on new route.
As far as I've seen, MANN architectures employ external memories in formats other than neural networks. Some of the advantages in these architectures stem exactly from this hybridization, compensating for NN drawbacks (eg. 'catastrophic forgeting'). On the other hand, multi-NNs models (eg. Generative Adversarial Networks, GANs) also appear very promising. I am curious whether there are models out there which use NNs, not only for processing, but for memory too?
I have collected flows on a machine connected in parallel to a switch by mirroring the port of the switch. All the incoming and outgoing traffic of the enterprise passes through this switch. Flows are collected using Softflowd and NfDump tools and Nfcapd files are created where each file captures flows of five minutes duration. When i checked the flows for DNS flows, i found no record with source port 53 or some dns server ip as server ip. However there are number of flows with destination port 53 and some dns server ip as destination ip. Does it mean that NfDump is collecting only outgoing traffic? If yes, what can be the reason?
In a service chain, the data rate between two successive VNFs changes based on the type of VNFs. This is referred to as compression factors in the paper linked below. In such paper they do not provide any reference nor numbers for this. Is anyone aware of real data about compression factors?
I am trying to run an ANN model using IBM SPSS. After every run I observe that architecture of the network and hence the synaptic weights change. May I know what is the probable reason for this or is it the nature of ANN?
I would like to calculate total power consumption in a network architecture.
Is there any simulation tool we can use to find power consumption on these layers (Application, presentation, session, transport) of OSI model?
Also power consumption by middleware (e.g., RMI-Java, ICE, DDS).
my project is 5G mobile network and I have some problems in find ta way to solve the small cells interference, therefore, I found that CACP is the solution I needed the mathematical equation of this aspect in any paper or reference
Hi there ,
I wish to get prominent inputs from respective scholars in software and networking , how your school of thought on architecture , components , diagrams and etc. Is it possible all of that in one layout and how to ensure it is understandable to the layman reader / assessor.
Importing a network or graph is very common which can be visualized and analyzed using various software like Gephi etc but to design a family of networks or graphs that have say a desired set of network properties e.g. 5-8 links per node with node density dependent on space variables or something like that? If it is also done, please mention some reference or web-links. Thanks in advance!
DNS is in Layer 7 protocol right? Why not 3/4? its job is nothing but name to ip conversion.? what about DHCP where its job to distribute ip.?
waiting 4 ur response??
Architecture styles are
Data flow architectures
Call and return architectures
I have a network of static nodes but now I want to add a mobile node which will move from a node to other according to a certain condition. Please help me
As I use Agent-based approach to simulate a multi-Layer Network, could someone of you please help me about this topic by giving me an example of datasets which can be useful to validate our architecture?
Thank you in advance,
Have a Nice day..
For a FPGA (PROASIC3, Libero IDE v9.1) based digital design, I want to increase fan-out count of a net without changing the original functionality. So i instantiated a combination of AND and XOR gates with enable input with that particular net (say net1) as mentioned below.
Gate1: AND2 port map (A=> net1, B=> enable, Y => net2);
Gate2: XOR2 port map (A=> net1, B=>net2, c=>net_next_circuit);
So, whenever enable=0, net1 value is resumed. i instantiated the same logic thrice to make fan-out count of net1 as 4 (along with already existing instantiation) but these modules are optimized in synthesis. How to resume these modules in the design?
I used syn_keep or syn_preserve attributes, even then tool does optimization.
Please give me any solution.
Thank you in advance.
I want to upgrade a network infrastructure to include 1000Mb switches and 10/100/1000 auto sensing network interface cards in each desktop. Recommend the cabling type that you would use in this network. Also, I have 100BaseTX cabling, and I need to provide access to a machine that is located 400 meters from the switch. What strategy could be used to accomplish this task?
Different solutions have been employed but packet loss ratio is very high. The best up till now has been wire-shark. Kindly guide as to how we can increase the capture ratio and what software are commonly used in the community for this purpose.
Thanks in advance for your help.
Do I need any special software for grid computing or is my optimizing software (for example Cplex) is enough?
I'm going to prepare 5 computers for this aim, what other kinds of requirements I need? Does it work with ordinary computer sharing network? Give me your suggestions please.
I have read the documents available on google and also gone through the Wikipedia's definition and formula about calculating the modularity for a particular network. I am using r studio for implementation.For example if my network is like following.Then the value of modularity being calculated in r studio as an output for this network is 0.08. How the answer 0.08 is being achieved that i am not getting.Can any one explain the mathematical formulation of modularity for networks.How the formula that is given in Wikipedia is being applied in this case that i am not getting. I hope my question make sense. Thanks in advance.
g <- graph_from_literal( 1--2, 4--5, 3--1, 5--2, 4--3 )
membership <- c(1,2,1,2,1)
I have found references to Webgraph, WebGraph++, and PowerGrASP as well as others research papers that describe work directed by S. Vigna. Are there any other tools that take in a graph (edgelist, gml, or any other format) and compress the graph into some compact data-structure and can be reconstructed to its original state that you can recommend?
To throw some light on Following areas specifically,
KPI’s for NPO;
DTH vs. IPTV
ADSL vs. HSPA
VDSL/FTT x vs. LTE
Macro rollout vs. Femto deployment
This question related to SDN and it is based on the fact the I wanna control the flux of my network using a Ryu controller, but to do so I have to know what are the nodes in the network. I know something about link layer discovery protocol, but it is not clear. I really appreciate your help.Thanks
By theory ,
It can be stated that UDP will faster as its connection less doesn’t required receiver side bothering,
but by my recent experiment I find TCP is some what giving 50% more throughput .
Theory which I formulated is, Overhead of TCP is one time. After that TCP get a connected way to pass data ,whereas UDP need to find it every time.
Also by Nagle effect which can make TCP packet immediate departure have same effect as UDP.
If so are ,then Why UDP is still in use in which condition UDP will be more useful then TCP.
Is there any available works on cloud resource discovery that use an hybrid architecture, specially master-slave or super-peer architectures?
I mean, is this a feasible idea since the network performance and resource
availability can be the tightest bottleneck for any cloud.
In fact I need some references on this subject.
Suppose our TCP SYN packet goes directly to the attacker. Can he generate fake response (TCP SYN-ACK) packet and send it back to us?
I am going through the literature survey for DHCP attacks. I would be thankful if someone provide me some reference for the same.
Determining the network architecture is one of the most important and difficult tasks in ANN model development, what is the best way to fined the the optimum number of layers and the number of nodes in each of these nodes?
I've recently looked into VLSI implementations of neural network architectures in which threshold logic circuit technologies are used. I'd appreciate keywords for such applications or publication links.
I need to extract f x,y,z signals, then combine them and re-extract them from the same signal without loss of data.
Can anyone construct a network whose diameter is more than X times as large as its average distance, where X is an arbitrarily given integer?
Is it possible to configure switches' forwarding logic, packet format, communication protocols etc. in cloudsim? Is it possible to simulate new/custom topologies, specifically recursive architectures like BCUBE? Does it support specifying communication protocols such as IP, TCP, and UDP with the simulation.
Which network software design, for example network architecture, can incorporate agents, host databases?
Wikipedia article about Peer-to-Peer computing says - "Peer-to-peer (P2P) computing or networking is a distributed application architecture that partitions tasks or work loads between peers. Peers are equally privileged, equipotent participants in the application. They are said to form a peer-to-peer network of nodes."
While Wikipedia article for Multi-agent systems says- "A multi-agent system (M.A.S.) is a computerized system composed of multiple interacting intelligent agents within an environment. Multi-agent systems can be used to solve problems that are difficult or impossible for an individual agent or a monolithic system to solve. Intelligence may include some methodic, functional, procedural or algorithmic search, find and processing approach."
Although having difference in definition, both architectures involve cooperation between distributed entities.
How does one define the suitability of having a P2P system based architecture or a MAS architecture for a research problem?