Hostile Neural Networks guide for FTB StoneBlock 4 1.1.1 to 1.3.0

Hostile Neural Networks lets you simulate mobs and turn the results into their drops, so you don't have to hunt them. It's based on Deep Mob Learning. The pack includes version 6.3.0 of the mod.^[1][2]

The core idea is the Data Model, an item that holds information about one entity. You craft a Model Framework, use it on the target entity to build the model, and then train it.^[1]

How the loop works

1. Train the model. Put it in a Deep Learner and the kills you make will train it. Every model starts out Faulty and has to reach Basic here.^[1]
2. Run simulations. Once a model is Basic or higher, the Simulation Chamber can simulate it. It needs energy and a catalyst item, which is usually a Prediction Matrix.^[1]
3. Collect predictions. Every simulation gives a generalized prediction, and it may also give a Mob Prediction. A higher model tier means better accuracy, up to 99.5% at Self Aware.^[1]
4. Make loot. The Loot Fabricator uses power to convert Mob Predictions into mob drops. You pick a target for each model type, and it remembers your choice.^[1]

Where it fits in the pack

The pack's resource_generation quest chapter follows this same order. a quest in "?" asks for the Model Framework and the Deep Learner. a quest in "?" comes after it and covers the Simulation Chamber, the Prediction Matrix and the Loot Fabricator.^[3]

A Data Model also shows up as one of the tasks in a quest in "?", in the creative chapter.^[4]

Datapacks can add custom Data Models, with one file per model in the data_models folder.^[1]

Hostile Neural Networks lets you simulate mobs and turn the results into their drops. The pack's quest book walks through the same first steps, starting with the Model Framework and the Deep Learner in quest a quest in "?". Here is the order I'd follow.^[1][3]

Start with the Model Framework

Everything starts with the Model Framework. You use it on the mob you want to copy, and that builds a Data Model for that mob. The Data Model is the core item of the mod, because it holds the information about an entity that you use to generate its drops.^[1][3]

Model FrameworkMod recipe

Crafting Shaped

Crafting Shaped

Clay Ball	Repeater	Clay Ball
any redstone dust	Smooth Stone	any redstone dust
Clay Ball	any gold ingot	Clay Ball

Recipe id: hostilenetworks:framework

Make the Deep Learner

A fresh model isn't useful yet, so you have to train it. The Deep Learner holds a few data models, and the kills you make start training them. Every model starts out Faulty, and you need to train it up to Basic in the Deep Learner before you can move on.^[1]

Deep LearnerMod recipe

Crafting Shaped

Crafting Shaped

any obsidian	Repeater	any obsidian
Repeater	any glass panes	Repeater
any obsidian	any redstone dust	any obsidian

Recipe id: hostilenetworks:deep_learner

Craft the Prediction Matrix

Once a model is Basic or higher, you can run simulations with it. Simulations need energy and a catalyst item, which is usually the Prediction Matrix.^[1]

16× Prediction MatrixMod recipe

Crafting Shaped

Crafting Shaped

any iron ingot	any glass panes	Empty slot
any glass panes	Clay Ball	any glass panes
Empty slot	any glass panes	any gold ingot

Recipe id: hostilenetworks:prediction_matrix

Simulation Chamber and Loot Fabricator

The Simulation Chamber runs the simulations. They always produce a generalized prediction, and they may also produce a Mob Prediction. The higher your model's tier, the better the accuracy, up to 99.5% at Self Aware. The Loot Fabricator then uses power to turn predictions into mob drops. You pick a target for each type of model, and the fabricator remembers your choice for each type.^[1]

The quest chapter follows the same order. The next quest, a quest in "?", asks for the Simulation Chamber, the Prediction Matrix and the Loot Fabricator together, and it comes after the first quest.^[3]

Handy extras

The Fabrication Directive stores selections from a Loot Fabricator. Right-click to copy them, and Shift-Right-click to apply them.^[5][6]

Fabrication DirectiveMod recipe

Crafting Shaped

Crafting Shaped

Empty slot	any generalized prediction	Empty slot
Paper	Prediction Matrix	Paper
Empty slot	Paper	Empty slot

Recipe id: hostilenetworks:fab_directive

The pack also has recipes that convert a Generalized Overworld Prediction into a Generalized Nether Prediction, and a Nether one into a Generalized Ender Prediction.^[6]

Generalized Nether PredictionMod recipe

Crafting Shaped

Crafting Shaped

Empty slot	Generalized Overworld Prediction	Empty slot
Generalized Overworld Prediction	Netherrack	Generalized Overworld Prediction
Empty slot	Generalized Overworld Prediction	Empty slot

Recipe id: hostilenetworks:living_matter/overworld_to_nether

Generalized Ender PredictionMod recipe

Crafting Shaped

Crafting Shaped

Empty slot	Generalized Nether Prediction	Empty slot
Generalized Nether Prediction	End Stone	Generalized Nether Prediction
Empty slot	Generalized Nether Prediction	Empty slot

Recipe id: hostilenetworks:living_matter/nether_to_ender

Hostile Neural Networks turns mob kills into simulated loot. Everything revolves around the Data Model, an item that holds information about one entity so you can generate that entity's drops later.^[1]

Building a Data Model

You start by crafting a Model Framework and using it on the target entity to build the model. After that, the model has to be trained.^[1][6]

Model FrameworkMod recipe

Crafting Shaped

Crafting Shaped

Clay Ball	Repeater	Clay Ball
any redstone dust	Smooth Stone	any redstone dust
Clay Ball	any gold ingot	Clay Ball

Recipe id: hostilenetworks:framework

Training in the Deep Learner

The Deep Learner stores a few data models, and the kills you make start training them. Every model starts in the Faulty state and has to be trained up to Basic in the Deep Learner.^[1]

Deep LearnerMod recipe

Crafting Shaped

Crafting Shaped

any obsidian	Repeater	any obsidian
Repeater	any glass panes	Repeater
any obsidian	any redstone dust	any obsidian

Recipe id: hostilenetworks:deep_learner

Running simulations

Once a model is Basic or higher, you can run simulations on it in the Simulation Chamber. A simulation always produces a generalized prediction. It may also produce a Mob Prediction, depending on the model's accuracy.^[1]

The higher the model's tier, the higher the accuracy, up to 99.5% at Self Aware. Each simulation needs energy, which depends on the model type, plus a catalyst item. The catalyst is usually a Prediction Matrix.^[1]

16× Prediction MatrixMod recipe

Crafting Shaped

Crafting Shaped

any iron ingot	any glass panes	Empty slot
any glass panes	Clay Ball	any glass panes
Empty slot	any glass panes	any gold ingot

Recipe id: hostilenetworks:prediction_matrix

Converting generalized predictions

Generalized predictions come in Overworld, Nether and End flavors. The pack has crafting recipes that convert a Generalized Overworld Prediction into a Generalized Nether Prediction, and a Nether one into a Generalized Ender Prediction.^[6]

Generalized Nether PredictionMod recipe

Crafting Shaped

Crafting Shaped

Empty slot	Generalized Overworld Prediction	Empty slot
Generalized Overworld Prediction	Netherrack	Generalized Overworld Prediction
Empty slot	Generalized Overworld Prediction	Empty slot

Recipe id: hostilenetworks:living_matter/overworld_to_nether

Generalized Ender PredictionMod recipe

Crafting Shaped

Crafting Shaped

Empty slot	Generalized Nether Prediction	Empty slot
Generalized Nether Prediction	End Stone	Generalized Nether Prediction
Empty slot	Generalized Nether Prediction	Empty slot

Recipe id: hostilenetworks:living_matter/nether_to_ender

The Loot Fabricator

Mob Predictions are processed in a Loot Fabricator, which uses power to convert predictions into mob drops. The available drops come from a pre-defined list of items, and you pick a target for each type of model.^[1]

The fabricator remembers your choice for each type. That means you can feed several prediction types into a single fabricator.^[1]

Copying fabricator settings

The Fabrication Directive stores selections from a Loot Fabricator. Right-click to copy a fabricator's selections, and Shift-Right-Click to apply them to another one. This is handy if you build more than one fabricator.^[5]

Fabrication DirectiveMod recipe

Crafting Shaped

Crafting Shaped

Empty slot	any generalized prediction	Empty slot
Paper	Prediction Matrix	Paper
Empty slot	Paper	Empty slot

Recipe id: hostilenetworks:fab_directive

Reading model tooltips

The mod's tooltip text shows that model info can include whether a model generates experience, how many loot table copies it rolls, and whether it produces no base drop. Holding Shift shows more info. The text also references an attunement action on entities, and a message saying attunement is disabled.^[5]

Unverified The sources don't cover the Simulation Chamber's or Loot Fabricator's recipes, exact energy costs, the tier thresholds between Faulty and Self Aware, or how these machines handle hoppers, pipes, and other automation. Check them in-game before building around them. It's also unconfirmed whether attunement is actually disabled in this pack.

Hostile Neural Networks progression in this pack follows a short chain in the quest book's resource generation chapter, and the recipes below are the in-pack versions.^[3][6]

Step 1: Model Framework and Deep Learner

The first quest, a quest in "?", asks you to get a Model Framework and a Deep Learner. Both have crafting recipes in this pack.^[3][6]

Model FrameworkMod recipe

Crafting Shaped

Crafting Shaped

Clay Ball	Repeater	Clay Ball
any redstone dust	Smooth Stone	any redstone dust
Clay Ball	any gold ingot	Clay Ball

Recipe id: hostilenetworks:framework

Deep LearnerMod recipe

Crafting Shaped

Crafting Shaped

any obsidian	Repeater	any obsidian
Repeater	any glass panes	Repeater
any obsidian	any redstone dust	any obsidian

Recipe id: hostilenetworks:deep_learner

Step 2: The machines

Once the first quest is done, a quest in "?" follows it. It asks you for a Simulation Chamber, a Prediction Matrix and a Loot Fabricator.^[3]

The Prediction Matrix is a plain crafting recipe, so you can make it right away.^[6]

16× Prediction MatrixMod recipe

Crafting Shaped

Crafting Shaped

any iron ingot	any glass panes	Empty slot
any glass panes	Clay Ball	any glass panes
Empty slot	any glass panes	any gold ingot

Recipe id: hostilenetworks:prediction_matrix

Fabrication Directives and prediction tiers

The Fabrication Directive is also crafted by hand in this pack. Its recipe uses the Prediction Matrix, so craft that first.^[6]

Fabrication DirectiveMod recipe

Crafting Shaped

Crafting Shaped

Empty slot	any generalized prediction	Empty slot
Paper	Prediction Matrix	Paper
Empty slot	Paper	Empty slot

Recipe id: hostilenetworks:fab_directive

The generalized predictions are organised by dimension: Overworld, Nether and Ender. In this pack a Generalized Overworld Prediction can be crafted into a Generalized Nether Prediction, and a Nether prediction into an Generalized Ender Prediction.^[6]

Generalized Nether PredictionMod recipe

Crafting Shaped

Crafting Shaped

Empty slot	Generalized Overworld Prediction	Empty slot
Generalized Overworld Prediction	Netherrack	Generalized Overworld Prediction
Empty slot	Generalized Overworld Prediction	Empty slot

Recipe id: hostilenetworks:living_matter/overworld_to_nether

Generalized Ender PredictionMod recipe

Crafting Shaped

Crafting Shaped

Empty slot	Generalized Nether Prediction	Empty slot
Generalized Nether Prediction	End Stone	Generalized Nether Prediction
Empty slot	Generalized Nether Prediction	Empty slot

Recipe id: hostilenetworks:living_matter/nether_to_ender

The creative chapter

The Data Model also shows up as one of the tasks in quest a quest in "?" in the creative chapter. It sits next to creative items from several other mods, and it only comes after a list of other quests in that chapter. It isn't part of the normal progression path above.^[4]

The clearest link between Hostile Neural Networks and another mod in this pack is Twilight Forest. The pack has a set of recipes that turn Generalized Twilight Prediction into Twilight Forest materials.^[6]

The pack changed every one of these recipes, so don't assume they match the mod's defaults. They give you these Twilight Forest items: Cicada, Firefly, Liveroot, Moonworm, Raven Feather, Steeleaf and Torchberries.^[6]

Twilight Forest recipes

8× CicadaChanged by the pack

Crafting Shapeless

• Generalized Twilight Prediction
• Generalized Twilight Prediction
• Note Block

Recipe id: hostilenetworks:living_matter/twilight/cicada

8× FireflyChanged by the pack

Crafting Shapeless

• Generalized Twilight Prediction
• Generalized Twilight Prediction
• Torch

Recipe id: hostilenetworks:living_matter/twilight/firefly

8× LiverootChanged by the pack

Crafting Shapeless

• Generalized Twilight Prediction
• Generalized Twilight Prediction
• Generalized Twilight Prediction
• Generalized Twilight Prediction
• Stick

Recipe id: hostilenetworks:living_matter/twilight/liveroot

8× MoonwormChanged by the pack

Crafting Shapeless

• Generalized Twilight Prediction
• Generalized Twilight Prediction
• Soul Torch

Recipe id: hostilenetworks:living_matter/twilight/moonworm

4× Raven FeatherChanged by the pack

Crafting Shapeless

• Generalized Twilight Prediction
• Feather

Recipe id: hostilenetworks:living_matter/twilight/raven_feather

8× SteeleafChanged by the pack

Crafting Shapeless

• Generalized Twilight Prediction
• Generalized Twilight Prediction
• Generalized Twilight Prediction
• Generalized Twilight Prediction
• any leave

Recipe id: hostilenetworks:living_matter/twilight/steeleaf

The second ingredient here is a tag, so any item in the leaves tag works.

8× TorchberriesChanged by the pack

Crafting Shapeless

• Generalized Twilight Prediction
• Sweet Berries

Recipe id: hostilenetworks:living_matter/twilight/torchberries

Quests

In the resource generation chapter, a quest in "?" asks for Model Framework and Deep Learner. a quest in "?" comes after it and asks for Simulation Chamber, Prediction Matrix and Loot Fabricator. Both quests only ask for items from this mod, so there is no cross-mod requirement there.^[3]

In the creative chapter, a quest in "?" lists Data Model as a task item next to creative items from several other mods, including Ars Nouveau, Create and Mekanism Lasers. That is a shared quest rather than a crafting link.^[4]

Hostile Neural Networks shows up in two quest chapters in this pack. The main one is resource_generation, where the mod gets a short two-step path. It also appears once in the creative chapter.^[3][4]

Chapter: resource_generation

Both quests here are untitled in the quest file, so you'll recognise them by their IDs and their tasks. They run in this order:^[3]

1. a quest in "?" is the starting point. Its tasks ask you to obtain a Model Framework and a Deep Learner.^[3]
2. a quest in "?" unlocks after the first quest. It asks for three machines: the Simulation Chamber, the Prediction Matrix and the Loot Fabricator.^[3]

So the pack has you start with the model framework and the Deep Learner, then move on to the machines that use your data models.^[3]

Chapter: creative

a quest in "?" is also untitled. Its task list includes the Data Model alongside items from many other mods. It only unlocks after a long list of earlier quests in that chapter. The mod is just one entry among many here, so this isn't a Hostile Neural Networks progression quest.^[4]

Same as in the FTB Skies 2 1.22.0 guide

Everything for Hostile Neural Networks lives in one file, config/hostilenetworks.cfg. Most of the options you will care about are power sizes and how models level up, so here is exactly how this pack sets them.^[7]

The value the pack flags as changed

[[config:config/hostilenetworks.cfg#models.Data Center Sim Cost Multiplier]] is set to 1.5. It scales the per-tick power cost the Data Center pays for each active model, as a multiplier on the model's base simCost, so 1.5 means 150% of normal. The allowed range is 0.01 to 100.0. The config listing marks this option as changed by the pack, but the mod default it shows is also 1.5, so 1.5 is the value you get in this pack.^[7]

Power

The power options below all match the mod defaults in this pack.^[7]

• [[config:config/hostilenetworks.cfg#power.Data Center Power Cap]]: 8,000,000 FE. The config describes this as sized for 25 concurrent self-aware simulations.^[7]
• [[config:config/hostilenetworks.cfg#power.Sim Chamber Power Cap]]: 2,000,000 FE stored in the Simulation Chamber.^[7]
• [[config:config/hostilenetworks.cfg#power.Loot Fab Power Cap]]: 1,000,000 FE stored in the Loot Fabricator.^[7]
• [[config:config/hostilenetworks.cfg#power.Loot Fab Power Cost]]: 256 FE/t for the Loot Fabricator.^[7]

Model training and attuning

• [[config:config/hostilenetworks.cfg#models.Action Upgrades Model]] is true, so killing mobs and breaking blocks upgrades the data on a matching model.^[7]
• [[config:config/hostilenetworks.cfg#models.Right Click To Attune]] is true, so right clicking a blank data model on a mob or block attunes it to that target.^[7]
• [[config:config/hostilenetworks.cfg#models.Continuous Accuracy]] is true, so a model's accuracy rises as it gains progress toward the next tier, instead of staying at the base accuracy of its current tier.^[7]
• [[config:config/hostilenetworks.cfg#models.Sim Chamber Upgrades Model]] is 1, meaning the Simulation Chamber's Training Mode upgrades model data. The options are 0 (disabled, which also disables Training Mode entirely), 1 (yes) and 2 (only up to tier boundaries).^[7]
• [[config:config/hostilenetworks.cfg#models.Enable Block Data Models]] is false. It is described as experimental and covers block data models such as the built-in ore models.^[7]

Deep Learner HUD

The HUD is anchored at top_left with an X offset of 0 and a Y offset of 0, which are the mod defaults. Change [[config:config/hostilenetworks.cfg#client.Deep Learner HUD Anchor Point]], [[config:config/hostilenetworks.cfg#client.Deep Learner HUD X Offset]] and [[config:config/hostilenetworks.cfg#client.Deep Learner HUD Y Offset]] to move it. Each offset can go from -1000 to 1000.^[7]