Hostile Neural Networks

Hostile Neural Networks (HNN for short) is a mod based around simulating mob kills for loot! To do this, you will be collecting "data" on mobs, then running simulations from the data to create their drops!

Asks for: Redstone

The Deep Learner is one of the core components of HNN. With this item, you can right click to open up a HUD and store Data Models to gather data while killing mobs. Each Learner can hold 4 Data Models, and as long as you have the Learner anywhere in your inventory, it will collect data for the models placed inside! Note: You can have multiple Deep Learners in your inventory, and it will count the kills for all Data Models inside all of the Learners.

Asks for: Deep Learner

Model Frameworks are the starting point for creating Data Models. To start collecting data on a specific mob, you'll want to make one of these then right click the mob you'd like to gather data on. When done properly, the Model Framework will turn into the Data Model of the mob you clicked on. To gather data for that mob, place the newly created Data Model inside of a Deep Learner, then start killing that specific mob with the Deep Learner in your inventory.

Asks for: Model Framework

The Simulation Chamber is the main machine within HNN. When given power, it will run simulations based on the Data Model that is placed inside. The power required also depends on the Data Model placed inside, and can be seen within the Model's tooltip. The machine also requires at least one Prediction Matrix to run a simulation. Simulations have two modes. Training: The Data Model gains data, without producing any Generalized Predictions or Mob Predictions. Costs more Forge Energy compared to Inference. Inference: The result of the simulation will be output onto the Matrix, without the Data Model gaining any Data. Simulations will always produce some kind of Generalized Prediction, which can be used in various crafting recipes. They also vary depending on the origin of the mob in the Data Model. If the simulation succeeds, you will receive a Mob Prediction based off of the model. The higher the tier of the Data Model used, the higher the success rate is for a simulation. Faster compared to Training.

Asks for: Prediction Matrix, Simulation Chamber

The Loot Fabricator is the machine that gets you that sweet, sweet loot from Mob Predictions. When given power and a successful Mob Prediction, you will be able to choose which item you'd like to get from that mob's Data Model. The machine will remember your choice, so it can be automated. Examples of Mob Predictions: - Ender Dragons for Dragon Eggs - Zombies for Iron - Withers for Nether Stars

Asks for: Loot Fabricator

Once you make your first Data Model, you'll need to level it up to be useful. Data Models start out at the first tier: Faulty. You'll need to at least level it to Basic if you'd like to run some simulations on it! Once placed inside of the Deep Learner, the HUD will tell you how many kills are required to level it up. You can also see this info in the tooltip of the Model itself.

Asks for: Data Model

The Data Center is a 7x7x7 multiblock combining the functionalities of the Simulation Chamber and the Loot Fabricator in a unique way. It is capable of handling 25 Self Aware Data Models concurrently, with the same effects as Inference simulations, with the following difference: Instead of producing Mob Predictions, the Data Center produces loot directly, with the same possible modes as the Loot Fabricator. To build the multiblock, the floor (first layer) must be made of Obsidian (check the Obsidians task to see what forms of Obsidian are accepted by the Data Center). The walls and ceiling of the multiblock must be made of Black Stained Glass. The controller may substitute one bottom-wall block, and IO Ports can substitute Black Stained Glass.

Asks for: Connecting Black Stained Glass, Clear Black Stained Glass, Scratched Black Stained Glass, Connecting Tinted Black Stained Glass, Data Center, Data Center IO Port