Does this increase the power of a neural network?

If we use a stack with a DFA (Finite Automata), that is, a PDA on each node of a neural network, can we solve problems that can not be solved by a normal neural network?

Lightning Network Daemon – How to get the outgoing and incoming capacity of a node with the help of LND?

Battery Exchange Network

The Stack Exchange network includes 176 question and answer communities, including Stack Overflow, the largest and most reliable online community on which developers can learn, share knowledge and build their careers.

Visit Stack Exchange

Network – Dynamic Gaming Server Architecture

I'm trying to find a way to design a gaming server architecture for a mobile game in real time. The design should be profitable, but still to some extent scalable (later).

Mobile client feeds:

Connects to a distribution server that has been given by DNS LB and is grouped with x players. If it is reached, the distribution server must contact a control server that starts or uses a running game server. The control server returns the IP address and port to the distribution server, which then sends it to the grouped drives.
Players then connect to the specified game server. After the match, players will be sent together to a new server or distribution server.

I wonder how the real server (VM) of game servers can be managed dynamically.
I've looked at AWS EC2 instances, but if we had 500 players with 6 players per group playing 6 hours a day, it would take about 84 EC2 instances t3.small (2 vCPU, 2GiB RAM). It will cost around $ 1,160.00 per month. With this price, it would be beneficial to have real virtual machines and not dynamic cloud instances, I guess?

enter the description of the image here

Explanation of the components


Mobile clients will connect to these to be grouped with other players. It's like a lobby where players will sit up to reach X players or for example. a countdown has elapsed.

Control server:

Need to support the booted server (VM) and manage the boot requests of the game server of the distributors.

Game server:

Is an author like server server with physics, etc. A game server (software) for X players in a group. Players will connect from the distributor after the "phase" of the lobby.


How can this be dynamic, scalable / profitable and is it a usable architecture?
I see a problem with the control server because it has to deal with N game server vms and N distributors. But if we scale the control server, each control server must know the same process.

(If something is missing or can be improved, please let me know.) English is not my first language, so it's a bit difficult to express my goal in a subject as complicated as this one.)

American Network vs AakashPereira vs GamingForum

American Network –

machine learning – Splitting of a large matrix into a neural network

I'm trying to create a neural network that takes a 16x100x100 matrix and divides it into every 100×100 part, places them through a dense layer, and then reassembles them to form a 16x10x10 matrix. I have a little trouble creating the NetModel.

Is there a simpler way to send the entry to 16 PartLayers and then each PartLayer to LinearLayer?

IP Network New Cryptocurrency (beta phase) | NewProxyLists

So, what is the IP network?
Pi is a secure mobile cryptocurrency and smart contract platform owned and operated by ordinary citizens.
To achieve this goal, we give everyone convenient access to the most proven and convincing cryptocurrency application to date: mining (earn cryptocurrency rewards to help secure a large ledger distributed).
More specifically, our goal is to enable more people to help validate transactions and earn cryptocurrency rewards from the comfort of their mobile phones.

The only way to join the IP network is to be invited by an existing member like me. ?
For a limited time, you can join the beta to earn Pi and help grow the network.


The official website is He's still in beta, so you have to be invited. Use the code (Lordvik16), its capital L
This is not yet registered on public exchanges, but it will happen soon. So, it's better to extract Pi as much as we can.
If you have a phone lying around, it will be easier.
Let Hope Pi go to big games.

American network

American Network –

linux networking – The Docker overlay host to the container network does not work

I'm trying to create a docker clock network without Swarm Link by using multi host consul in an AMI instance of ecs instance. I have created a network of recovery and able to communicate with the containers. However, when I try to reach a specific port on a host, the port receives a refused connection error. The host does not have a firewall. How can I reach the container port on the overlay network.

Strange fact

  1. If I put the container without overlay network I can ping the container port from the same host machine as well as from another machine connected to the overlay network.
[ec2-user@ip-172-31-3-222 ~]$ sudo docker run --name mynginx1 -p 5080: 80 -d nginx

Job :

[ec2-user@ip-172-31-3-222 ~]$ telnet 5080
Try ...
Connected to
The escape character is & ^;

& # 39 ;.
  1. If I put the container superimposed network I can not ping the container from the host machine or from another host machine. .
[ec2-user@ip-172-31-3-222 ~]$ sudo docker run --net my-overlay --name mynginxoverlay -p 5081: 80 -d nginx
[ec2-user@ip-172-31-3-222 ~]$ telnet 5081
Try ...

My overlay network:

network docker create -d overlay --subnet = / 20 my-overlay


[ec2-user@ip-172-31-7-118 ~]$ sudo docker inspects my overlay
        "Name": "my-overlay",
        "Id": "f4f2c67649ee108c9c88f224687a0ac1590a63f7ac06544cfe6dbd88aa2d773a",
        "Created": "2019-06-22T15:09:13.739759915Z",
        "Scope": "global",
        "Driver": "overlay",
        "EnableIPv6": false,
        "IPAM": {
            "Driver": "default",
            "Options": {},
            "Config": [
                    "Subnet": ""
"Internal": false,
"Attachable": false,
"Ingress": false,
"ConfigFrom": {
"Network": ""
"ConfigOnly": false,
"Containers": {
"ep-2bf37ede7b867f3b4059755d482da6e98557896d5c64c06c7e8655b4892e01a1": {
"Name": "mynginxoverlay",
"EndpointID": "2bf37ede7b867f3b4059755d482da6e98557896d5c64c06c7e8655b4892e01a1",
"MacAddress": "02: 42: ac: 1f: 00: 02",
"IPv4Address": "",
"IPv6Address": ""
"Options": {},
"Labels": {}

American Network vs Bayside Gamers

American Network –

Network – Hyper-V Server 2016: Can not Automatically Enter IP Addresses of Virtual Machines on a VLAN Configured with IP Helps

I am pretty new to this field, so I'm sorry if my understanding of virtual networking is a bit lacking.

I have a Hyper-V cluster (Server 2016) and I would like to be able to configure the virtual machines that I have configured on this cluster via DHCP. I use an external vswitch with these virtual machines configured with a specific VLAN ID. This particular VLAN is configured with DHCP scope and IP wizards.

The problem is that these virtual machines do not automatically recover IP addresses even after DHCP activation. I can manually assign IP addresses from this subnet and they work perfectly, but the virtual machines do not "capture" them. If I do not manually assign an IP address from this range, an APIPA address will simply be assigned to it. Again, the subnet is configured with IP helpers, so in theory I should not need to configure a DHCP server on this VLAN (or at least that's what I understand).

What would you suggest I try next? Or what additional information can I provide?

Thank you!