Hostile Neural Networks in Project Infinity 0.1

Hostile Neural Networks lets you simulate mobs and turn the results into their drops, so you don't have to hunt them. This pack includes version 5.3.3 for 1.20.1. The mod is based on Deep Mob Learning.^[1][2]

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

How the loop works

1. Train the model in a Deep Learner. Kills you make while it is stored there train it. Every model starts out Faulty and has to reach Basic in the deep learner.^[1]
2. Once the model is Basic or higher, run simulations in a Simulation Chamber. This needs energy, which depends on the model type, and a catalyst item, usually a Prediction Matrix.^[1]
3. Each simulation always gives a generalized prediction. It may also give a Mob Prediction, depending on the model's accuracy. Accuracy rises with the model's tier, up to 99.5% at Self Aware.^[1]
4. Feed the predictions into a Loot Fabricator. It uses power to turn them into mob drops. You pick a target for each type of model, and it remembers your choice for each type, so one fabricator can take several kinds of prediction.^[1]

Where it sits in the quest book

The pack's quest book gives the mod its own chapter, Hostile Neural Networks, in the Resource Generation group (chapter 11 of 44 in the book). The chapter follows the same loop described above: it asks you to get the machines and items that train models, run simulations and turn predictions into loot.^[1][3]

The chapter opens with "Welcome to Hostile Neural Networks!" quest, a checkmark quest. Two quests follow it: "Learning Mobs" quest asks you to get the Deep Learner, which trains models, and "Modeling Mobs" quest asks you to get the Model Framework, which you use on a mob to build a model.^[1][3]

Both of those lead to "Simulating Deaths" quest, which asks you to get the Simulation Chamber, the machine that runs simulations. After that the chapter branches. Four quests ask for the Nether, Overworld, Twilight and End generalized predictions, such as "Generalized Nether Prediction" quest and "Generalized End Prediction" quest. "Creating Loot" quest asks for the Loot Fabricator, which turns predictions into mob drops.^[1][3]

Most of the rest of the chapter comes after Creating Loot. It is a long run of quests that each ask for a specific Data Model or Prediction item. Some of the chapter's quests aren't listed in the sources I have, so check the quest book itself for those.^[3]

Hostile Neural Networks turns mobs into loot, and the first steps always go in the same order: build a model of a mob, train it, simulate it, then turn the results into drops. The pack's quest book follows that order in its Hostile Neural Networks chapter (chapter 11 of 44, in the Resource Generation group). It opens with "Welcome to Hostile Neural Networks!" quest, a simple checkmark quest, and both "Learning Mobs" quest and "Modeling Mobs" quest come right after it.^[1][3]

Make a Model Framework

Your first item is the Model Framework, which the quest "Modeling Mobs" quest asks for. It is the starting point for a Data Model. Right click it on a mob to attune it to that mob. That works in this pack because [[config:config/hostilenetworks.cfg#models.Right Click To Attune]] is set to true, which is also the mod default.^[1][3][4]

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 a Deep Learner

Next is the Deep Learner (quest "Learning Mobs" quest, subtitled "The Learning HUD"). It holds a few data models, and the kills you make train them. Every model starts out Faulty and has to be trained up to Basic here before it can go into the Simulation Chamber. Kills upgrade models because [[config:config/hostilenetworks.cfg#models.Killing Upgrades Model]] is true in this pack, so leave that alone.^[1][3][4]

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

Build the Simulation Chamber

Once a model is Basic or higher, it can go into the Simulation Chamber. The quest "Simulating Deaths" quest comes after both of the quests above. Simulations always give a generalized prediction, and can also give a mob prediction depending on the model's accuracy. They cost energy and a catalyst item, which is usually a Prediction Matrix. The chamber stores up to 2,000,000 FE ([[config:config/hostilenetworks.cfg#power.Sim Chamber Power Cap]]), which is the mod default.^[1][3][4]

Simulation ChamberMod recipe

Crafting Shaped

Crafting Shaped

Empty slot	any glass panes	Empty slot
Ender Pearl	any obsidian	Ender Pearl
any lapi gems	Comparator	any lapi gems

Recipe id: hostilenetworks:sim_chamber

The chamber can also level up your model as it runs. [[config:config/hostilenetworks.cfg#models.Sim Chamber Upgrades Model]] is 1 in this pack (0 is no, 1 is yes, 2 is only up to tier boundaries).^[4]

Craft Prediction Matrices

One craft gives you 16 Prediction Matrix items, so make a stack before you start simulating.^[1][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

One craft makes 16 Prediction Matrix.

Make a Loot Fabricator

Predictions are only useful once the Loot Fabricator processes them into mob drops (quest "Creating Loot" quest, which also follows Simulating Deaths). You pick a target for each type of model, and the fabricator remembers your choices, so one machine can handle several prediction types. In this pack it uses 256 FE/t ([[config:config/hostilenetworks.cfg#power.Loot Fab Power Cost]]) and stores up to 1,000,000 FE ([[config:config/hostilenetworks.cfg#power.Loot Fab Power Cap]]), both the mod defaults.^[1][3][4]

Loot FabricatorMod recipe

Crafting Shaped

Crafting Shaped

Empty slot	any netherite ingot	Empty slot
any diamond gem	any obsidian	any diamond gem
any gold ingot	Comparator	any gold ingot

Recipe id: hostilenetworks:loot_fabricator

After the basics

After Simulating Deaths, the quest book branches into the generalized predictions: "Generalized Overworld Prediction" quest, "Generalized Nether Prediction" quest, "Generalized Twilight Prediction" quest and "Generalized End Prediction" quest. You can also convert a prediction up a dimension with these two recipes.^[3][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

Turns a Generalized Overworld Prediction into a Generalized Nether Prediction.

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

Turns a Generalized Nether Prediction into a Generalized Ender Prediction.

Hostile Neural Networks revolves around one item: the Data Model. It holds information about one mob. You train it, run simulations with it, and turn the resulting predictions into that mob's drops. Here is how each step works in this pack.^[1]

Building and training a model

You start with a Model Framework (the Model Framework) and use it on the mob you want to copy. That attunes it to the mob. The pack keeps [[config:config/hostilenetworks.cfg#models.Right Click To Attune]] on, which is the mod default, so right clicking the mob works.^[1][4]

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

Every model starts out Faulty. You train it up to Basic in a Deep Learner. The Deep Learner stores a few models, and the kills you make begin to train them. The pack leaves [[config:config/hostilenetworks.cfg#models.Killing Upgrades Model]] on, so kills do upgrade a model.^[1][4]

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

The Simulation Chamber can also upgrade a model's data. [[config:config/hostilenetworks.cfg#models.Sim Chamber Upgrades Model]] takes 0 (no), 1 (yes) or 2 (only up to tier boundaries). The pack sets it to 1, the mod default. [[config:config/hostilenetworks.cfg#models.Continuous Accuracy]] is also on, so a model's accuracy rises gradually as it progresses toward the next tier instead of jumping only at tier changes.^[4]

Running simulations

Once a model is Basic or higher, you can put it in the Simulation Chamber. Every simulation produces a generalized prediction. Depending on the model's accuracy, it may also produce a Mob Prediction. Accuracy climbs with the model's tier, up to 99.5% at Self Aware.^[1]

A simulation needs energy, and the amount depends on the model. It also needs a catalyst item, which is usually the Prediction Matrix. Every model file shown from this pack uses the Prediction Matrix as its input. The chamber stores up to 2,000,000 FE, set by [[config:config/hostilenetworks.cfg#power.Sim Chamber Power Cap]], and the pack keeps the mod default.^{[1][4][7][8][9]}

Simulation ChamberMod recipe

Crafting Shaped

Crafting Shaped

Empty slot	any glass panes	Empty slot
Ender Pearl	any obsidian	Ender Pearl
any lapi gems	Comparator	any lapi gems

Recipe id: hostilenetworks:sim_chamber

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

This recipe makes 16 Prediction Matrices at once.

Each model is defined by a data file, and the pack changes many of them. Every file has a sim_cost and a base_drop. In this pack the base drop is a generalized Overworld, Nether or End prediction. Here are some sim_cost values:^{[7][8][9][10][11]}

• Cod: 32, with a base drop of Generalized Overworld Prediction.^[11]
• Creeper: 128, also an Overworld prediction.^[8]
• Blaze: 256, with a base drop of Generalized Nether Prediction.^[7]
• Dire Hound Leader: 4,096.^[12]
• Piglich and Gaia Spirit: 16,384 each, both with a base drop of Generalized Ender Prediction.^[13][14]
• Lich and several other boss models: 50,000.^{[10][15][16][17]}
• Draconic Guardian: 2,000,000, with an End prediction base drop.^[9]

The files also carry tier_data, a list of four numbers. Most models in this pack use 3, 54, 354 and 1,254. Lord Pumpkinhead has 6 as its first number, Gaia Spirit has 4, and Cod uses 6, 54, 354 and 1,254 together with a data_per_kill list of 1, 4, 10 and 18. The Draconic Guardian is far steeper: 1, 12,500, 25,000 and 32,767, with data_per_kill of 1, 1,000, 1,000 and 1,000.^{[9][11][13][14][18]}

Converting prediction types

Overworld, Nether and End predictions can be upgraded by crafting. An Overworld prediction becomes a Nether prediction, and a Nether prediction becomes an End 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

Loot Fabricator

Mob Predictions are processed in the Loot Fabricator, which spends power to turn them into mob drops. You pick a target drop for each type of model, and the fabricator remembers your choice per type. That means one fabricator can take several kinds of predictions.^[1]

It stores up to 1,000,000 FE ([[config:config/hostilenetworks.cfg#power.Loot Fab Power Cap]]) and costs 256 FE/t ([[config:config/hostilenetworks.cfg#power.Loot Fab Power Cost]]). The pack keeps both at the mod defaults.^[4]

Loot FabricatorMod recipe

Crafting Shaped

Crafting Shaped

Empty slot	any netherite ingot	Empty slot
any diamond gem	any obsidian	any diamond gem
any gold ingot	Comparator	any gold ingot

Recipe id: hostilenetworks:loot_fabricator

The available drops come from each model's fabricator_drops list, and the pack edits these lists. The Blaze model offers 16 Blaze Rod, 2 Blaze Shard and 4 Blaze Heart. The Creeper model offers 32 Gunpowder, 4 Creeper Head and 4 Creeper Heart. The Draconic Guardian model lists only Air.^[7][8][9]

Pack recipes for data models

The pack adds two extra recipes that output a Data Model. One is a shapeless recipe that takes a Wither Token. The other is a shaped recipe that takes a Data Model plus a set of Born in Chaos items, including Soulbane and Pumpkin Pistol.^[6]

Data ModelAdded by the pack

Shapeless (kubejs)

• Wither Token

Recipe id: kubejs:kjs/hostilenetworks_data_model

Data ModelAdded by the pack

Crafting Shaped

Crafting Shaped

Bag of Candy	Pumpkin Bullet	Lord Pumpkinhead's Hat
Smoldering Infernal Ember	Pumpkin Pistol	Soulbane
Magical Holiday Candy	Data Model	Magical Holiday Candy

Recipe id: minecraft:kjs/hostilenetworks_data_model

Hostile Neural Networks has its own quest chapter in the pack's book, in the "Resource Generation" group. Here is the path the book lays out.^[3]

The core path

The chapter opens with a welcome checkmark quest. Two quests follow it: "Learning Mobs" quest asks for the Deep Learner, and "Modeling Mobs" quest asks for the Model Framework.^[3][6]

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

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

After those, "Simulating Deaths" quest asks for the Simulation Chamber. "Leveling Up Data Models" quest is a checkmark quest that also comes after the first two.^[3][6]

Simulation ChamberMod recipe

Crafting Shaped

Crafting Shaped

Empty slot	any glass panes	Empty slot
Ender Pearl	any obsidian	Ender Pearl
any lapi gems	Comparator	any lapi gems

Recipe id: hostilenetworks:sim_chamber

Simulating Deaths leads to the next batch of quests: "Creating Loot" quest asks for the Loot Fabricator, and there is one quest for each prediction type: Overworld, Nether, Twilight and End.^[3][6]

Loot FabricatorMod recipe

Crafting Shaped

Crafting Shaped

Empty slot	any netherite ingot	Empty slot
any diamond gem	any obsidian	any diamond gem
any gold ingot	Comparator	any gold ingot

Recipe id: hostilenetworks:loot_fabricator

• "Generalized Overworld Prediction" quest asks you to hand in Generalized Overworld Prediction, which is the starting point of the netherrack recipe below.^[3][6]
• "Generalized Nether Prediction" quest asks you to hand in Generalized Nether Prediction, which the netherrack recipe makes and the end stone recipe uses.^[3][6]
• The book's Twilight step, "Generalized Twilight Prediction" quest, only asks you to hand in Generalized Twilight Prediction. The recipes here do not show how to make it.^[3][6]
• "Generalized End Prediction" quest asks you to hand in Generalized Ender Prediction, which the end stone recipe makes.^[3][6]

Prediction tiers

You can step a prediction up a tier by crafting. Overworld becomes Nether with netherrack, and Nether becomes Ender with end stone.^[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

Every model file shown for this pack takes a Prediction Matrix as its input. A matrix recipe makes 16 at once.^{[6][7][10][12][13][15][16][17][19]}

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

Each model's base drop is one of the prediction tiers, and the pack's model files set it. Blaze drops Nether, Dark Vortex and Dire Hound Leader drop Overworld, and the Piglich drops End. The Piglich model costs 16384 per simulation and its fabricator drops include 4 Piglich Hearts. The Blaze costs 256 and the Dark Vortex only 50. Dire Hound Leader costs 4096.^{[7][12][13][19]}

The boss models from Bosses of Mass Destruction all cost 50000. The Lich and Void Blossom drop Overworld predictions, the Gauntlet drops Nether, and the Obsidilith drops End.^{[10][15][16][17]}

Data models

After Creating Loot, the chapter has a very long run of quests that each ask you to hand in a data model or a prediction. Their titles are placeholders in the book. The run ends with a quest in "Hostile Neural Networks", a quest that asks you to hand in the Ultimate Prediction item.^[3]

The pack also adds two crafting recipes that produce a data model. One is a shapeless recipe and the other is a shaped one that takes in an existing data model.^[6]

Data ModelAdded by the pack

Shapeless (kubejs)

• Wither Token

Recipe id: kubejs:kjs/hostilenetworks_data_model

Data ModelAdded by the pack

Crafting Shaped

Crafting Shaped

Bag of Candy	Pumpkin Bullet	Lord Pumpkinhead's Hat
Smoldering Infernal Ember	Pumpkin Pistol	Soulbane
Magical Holiday Candy	Data Model	Magical Holiday Candy

Recipe id: minecraft:kjs/hostilenetworks_data_model

Several of the pack's data files (Piglich, Gauntlet, Lich, Obsidilith, Void Blossom) set the tier_data list to 3, 54, 354 and 1254. The Blaze, Dark Vortex and Dire Hound Leader files shown here have no tier_data entry.^{[10][13][15][16][17]}

The Extra HNN branch

After Simulating Deaths, the book also has a quest titled Simulator Modeling, with the subtitle "Models upgrading too slow?". Two quests follow it: a quest in "Hostile Neural Networks" asks you to hand in a speed upgrade item and a quest in "Hostile Neural Networks" asks for a data upgrade item. a quest in "Hostile Neural Networks" asks for a tier upgrade item and comes after the data one.^[3]

The book places a quest in "Hostile Neural Networks" (subtitled "4x your models!") after both Simulating Deaths and Creating Loot. It asks you to hand in a merger camera and a blank extra data model. From there the book continues to "Wither Skeleton, Wither, Ender Dragon, Warden Data Model" quest, which asks for an extra data model, and to two matching upgrade chains. Each step comes after the last, and its subtitle carries a processing multiplier.^[3]

1. Simulation Chamber line: the book asks you to hand in the items for a quest in "Hostile Neural Networks" (labelled 4x), a quest in "Hostile Neural Networks" (16x), a quest in "Hostile Neural Networks" (32x), then a quest in "Hostile Neural Networks" (64x).^[3]
2. Loot Fabricator line: the book asks you to hand in the items for a quest in "Hostile Neural Networks" (labelled 4x), a quest in "Hostile Neural Networks" (16x), a quest in "Hostile Neural Networks" (32x), then a quest in "Hostile Neural Networks" (64x).^[3]

Where it leads next

The V4 quests connect to the Masterful Machinery chapter, in the "Production" group. a quest in "Masterful Machinery" comes after Ultimate Simulation Chamber V4 and asks for a prediction chamber controller together with the Prediction Matrix. a quest in "Masterful Machinery" comes after Ultimate Loot Fabricator V4 and asks for a loot fabrication plant controller together with a Piglich Heart. Both also come after the Priority Setter quest.^[6][20]

In this pack, Prediction and Data Model items show up in recipes from a lot of other mods. The Loot Fabricator and Simulation Chamber are also the starting points for upgraded machines. Here is what the sources confirm.^[6]

Twilight Forest items from predictions

Generalized Twilight Prediction can be crafted into Twilight Forest items: Cicada, Firefly, Liveroot, Moonworm, Raven Feather, Steeleaf and Torchberries. These are the mod's own recipes, not ones the pack added. Here are three of them.^[6]

8× CicadaMod recipe

Crafting Shapeless

• Generalized Twilight Prediction
• Generalized Twilight Prediction
• Note Block

Recipe id: hostilenetworks:living_matter/twilight/cicada

4× Raven FeatherMod recipe

Crafting Shapeless

• Generalized Twilight Prediction
• Feather

Recipe id: hostilenetworks:living_matter/twilight/raven_feather

8× SteeleafMod recipe

Crafting Shapeless

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

Recipe id: hostilenetworks:living_matter/twilight/steeleaf

Extended Crafting upgrades

The pack changed the recipes for the Ultimate machines so they are built on Extended Crafting tables and chain upward from the Hostile Neural Networks machines. There are four Ultimate Loot Fabricator tiers, from Ultimate Loot Fabricator V1 to Ultimate Loot Fabricator V4, and two Ultimate Simulation Chamber tiers, Ultimate Simulation Chamber V1 and Ultimate Simulation Chamber V2. The V1 recipes use a table, and the later tiers build on the earlier ones.^[6]

Ultimate Loot Fabricator V1Changed by the pack

Shaped Table (extendedcrafting)

Shaped Table (extendedcrafting)

Prediction Matrix	Netherite Ingot	Prediction Matrix	Netherite Ingot	Prediction Matrix
Netherite Ingot	Loot Fabricator	Loot Fabricator	Loot Fabricator	Netherite Ingot
Prediction Matrix	Diamond	Prediction Matrix	Diamond	Prediction Matrix
Netherite Ingot	Diamond	Redstone Block	Diamond	Netherite Ingot
Prediction Matrix	Netherite Ingot	Redstone Block	Netherite Ingot	Prediction Matrix

Recipe id: extrahnn:ultimate_loot_fabricator_v1

Ultimate Simulation Chamber V1Changed by the pack

Shaped Table (extendedcrafting)

Shaped Table (extendedcrafting)

Diamond	Netherite Ingot	Ender Eye	Netherite Ingot	Diamond
Netherite Ingot	Simulation Chamber	Ender Eye	Simulation Chamber	Netherite Ingot
Diamond	Extra Model Framework	Ender Eye	Extra Model Framework	Diamond
Netherite Ingot	Simulation Chamber	Ender Eye	Simulation Chamber	Netherite Ingot
Diamond	Netherite Ingot	Ender Eye	Netherite Ingot	Diamond

Recipe id: extrahnn:ultimate_sim_chamber_v1

The pack also changed the recipe for the Extra Model Framework, which upgrades the Model Framework.^[6]

Extra Model FrameworkChanged by the pack

Crafting Shaped

Crafting Shaped

any diamond gem	Quartz	any diamond gem
Redstone Block	Model Framework	Redstone Block
any diamond gem	Netherite Scrap	any diamond gem

Recipe id: extrahnn:extra_data_model

Extended Crafting tables are also used for the pack's combined predictions. Shapeless table recipes added by the pack turn 25 predictions into Ultimate Boss Prediction, 14 into Ultimate Misc Prediction and 30 into Ultimate Vanilla Prediction.^[6]

Ultimate Boss PredictionAdded by the pack

Shapeless Table (extendedcrafting)

• Prediction
• Prediction
• Prediction
• Prediction
• Prediction
• Prediction
• Prediction
• Prediction
• Prediction
• Prediction
• Prediction
• Prediction
• Prediction
• Prediction
• Prediction
• Prediction
• Prediction
• Prediction
• Prediction
• Prediction
• Prediction
• Prediction
• Prediction
• Prediction
• Prediction

Recipe id: kubejs:ultimate_boss_prediction

Ultimate Misc PredictionAdded by the pack

Shapeless Table (extendedcrafting)

• Prediction
• Prediction
• Prediction
• Prediction
• Prediction
• Prediction
• Prediction
• Prediction
• Prediction
• Prediction
• Prediction
• Prediction
• Prediction
• Prediction

Recipe id: kubejs:ultimate_misc_prediction

Ultimate Vanilla PredictionAdded by the pack

Shapeless Table (extendedcrafting)

• Prediction
• Prediction
• Prediction
• Prediction
• Prediction
• Prediction
• Prediction
• Prediction
• Prediction
• Prediction
• Prediction
• Prediction
• Prediction
• Prediction
• Prediction
• Prediction
• Prediction
• Prediction
• Prediction
• Prediction
• Prediction
• Prediction
• Prediction
• Prediction
• Prediction
• Prediction
• Prediction
• Prediction
• Prediction
• Prediction

Recipe id: kubejs:ultimate_vanilla_prediction

Other recipes that use predictions and data models

• The pack added an Advanced AE reaction recipe that makes a Nether Star from 2 predictions plus a Nitro Crystal.^[6]
• A prediction is one of the inputs for Blacker Lotus, a pack-added recipe that also uses Actually Additions crystals, Botania's Black Lotus and Crystal Dragonstone.^[6]
• Wither Skeleton Seeds is crafted on an Extended Crafting table in a pack-added recipe that takes a data model among its inputs.^[6]
• A prediction can be crafted into a Chaos Guardian Prediction, a pack-added shapeless recipe.^[6]
• A data model and Wither Token can be swapped back and forth with two pack-added shapeless recipes.^[6]
• A pack-added recipe turns a data model into another data model using a set of Born in Chaos items, including Pumpkin Pistol and Soulbane.^[6]

Nether StarAdded by the pack

Reaction (advanced_ae)

• 2× Prediction
• Nitro Crystal

Recipe id: minecraft:nether_star_reaction_chamber

Blacker LotusAdded by the pack

Crafting Shaped

Crafting Shaped

Empowered Enori Crystal	Prediction	Empowered Diamatine Crystal
Crystal Dragonstone	Black Lotus	Crystal Dragonstone
Empowered Emeradic Crystal	Crystal Dragonstone	Empowered Void Crystal

Recipe id: minecraft:kjs/botania_blacker_lotus

Data models for modded mobs

The pack has its own data model files for mobs from other mods. The sources show these examples, and each takes a Prediction Matrix as its simulation input.^{[10][13][15][16][17]}

• The All the Modium Piglich has a simulation cost of 16,384, and its base drop is Generalized Ender Prediction.^[13]
• The Bosses of Mass Destruction Gauntlet has a simulation cost of 50,000, and its base drop is Generalized Nether Prediction.^[15]
• The Lich has a simulation cost of 50,000, and its base drop is Generalized Overworld Prediction.^[10]
• The Obsidilith has a simulation cost of 50,000, and its base drop is Generalized Ender Prediction.^[16]
• The Void Blossom has a simulation cost of 50,000, and its base drop is Generalized Overworld Prediction.^[17]
• The pack also changed the vanilla Blaze data model, with a simulation cost of 256. Its Loot Fabricator drops include items from other mods.^[7]

The pack has dozens more data model files for other mods' mobs, and their contents are not covered here.^[13]

In the quest book

The Hostile Neural Networks chapter sits in the Resource Generation group, as chapter 11 of 44, and has 164 quests. The visible quests ask for this mod's own items, starting with "Welcome to Hostile Neural Networks!" quest, and many of them come after "Creating Loot" quest. Around 93 quests are not shown here, so check the quest book itself for what they cover.^[3]

The pack's quest book gives Hostile Neural Networks its own chapter, also called Hostile Neural Networks, in the Resource Generation group. Two quests in the Masterful Machinery chapter also ask for items from this mod. The chapter is long, and most of it is one repeated pattern, covered below.^[3][20]

Getting started

1. "Welcome to Hostile Neural Networks!" quest is the opener. Its task is a checkmark, with no item to hand in.^[3]
2. "Learning Mobs" quest asks you to hand in the Deep Learner. Its subtitle is The Learning HUD, and the book lists the welcome quest before it.^[3]
3. "Modeling Mobs" quest asks you to hand in the Model Framework, with the subtitle For 'Gathering Data'. The book also lists the welcome quest before this one.^[3]
4. "Simulating Deaths" quest asks you to hand in the Simulation Chamber. The book lists Learning Mobs and Modeling Mobs before it.^[3]
5. "Leveling Up Data Models" quest is another checkmark quest, listed after Modeling Mobs and Learning Mobs.^[3]

After Simulating Deaths

Several quests are listed after Simulating Deaths. Four of them ask for a generalized prediction: "Generalized Overworld Prediction" quest asks for the Generalized Overworld Prediction, "Generalized Nether Prediction" quest for the Generalized Nether Prediction, "Generalized End Prediction" quest for the Generalized Ender Prediction, and "Generalized Twilight Prediction" quest for the Generalized Twilight Prediction.^[3]

"Creating Loot" quest asks you to hand in the Loot Fabricator. a quest in "Hostile Neural Networks" starts the upgrade branch, with the subtitle Models upgrading too slow?^[3]

Upgrades and the Extra Model Framework

The upgrade branch runs in a short line. After Simulator Modeling the book lists the Upgrade: Speed quest a quest in "Hostile Neural Networks" and the Upgrade: Data quest a quest in "Hostile Neural Networks". The Upgrade: Tier quest a quest in "Hostile Neural Networks" is listed after Upgrade: Data.^[3]

a quest in "Hostile Neural Networks" has the subtitle 4x your models! and asks for a merger camera and a blank extra data model. The book lists Simulating Deaths and Creating Loot before it. "Wither Skeleton, Wither, Ender Dragon, Warden Data Model" quest is listed after it, with the subtitle Try a combo!^[3]

a quest in "Hostile Neural Networks" is listed after Creating Loot. It asks for the setting card and has the subtitle Copy/Paste.^[3]

Ultimate machine tiers

After Extra Model Framework the book lists two parallel lines of quests, one for the Simulation Chamber and one for the Loot Fabricator. Each quest is listed after the previous tier, and the subtitles read as processing multipliers.^[3]

• Simulation Chamber: a quest in "Hostile Neural Networks" (subtitle 4x Processing), then a quest in "Hostile Neural Networks" (16x), then a quest in "Hostile Neural Networks" (32x), then a quest in "Hostile Neural Networks" (64x).^[3]
• Loot Fabricator: a quest in "Hostile Neural Networks" (subtitle 4x Processing), then a quest in "Hostile Neural Networks" (16x), then a quest in "Hostile Neural Networks" (32x), then a quest in "Hostile Neural Networks" (64x).^[3]

The big collection

Most of the chapter is a long run of quests listed after Creating Loot. Each one asks for a single item, either a data model (Data Model) or a prediction (Prediction). The book shows them with a placeholder title, so they appear as %s Data Model and %s Prediction. a quest in "Hostile Neural Networks" is one data model quest and a quest in "Hostile Neural Networks" is one prediction quest.^[3]

a quest in "Hostile Neural Networks" asks for an ultimate prediction item. The book lists all of those prediction and data model quests before it.^[3]

Masterful Machinery chapter

Two quests in the Masterful Machinery chapter (Production group) ask for items from this mod or its quest line. Both are listed after the Priority Setter quest.^[20]

• a quest in "Masterful Machinery" asks for the prediction chamber controller and the Prediction Matrix. It is listed after the Ultimate Simulation Chamber V4 quest.^[20]
• a quest in "Masterful Machinery" asks for the loot fabrication plant controller and the Piglich Heart. It is listed after the Ultimate Loot Fabricator V4 quest.^[20]

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.^[21]

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.^[21]

Power

The power options below all match the mod defaults in this pack.^[21]

• [[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.^[21]
• [[config:config/hostilenetworks.cfg#power.Sim Chamber Power Cap]]: 2,000,000 FE stored in the Simulation Chamber.^[21]
• [[config:config/hostilenetworks.cfg#power.Loot Fab Power Cap]]: 1,000,000 FE stored in the Loot Fabricator.^[21]
• [[config:config/hostilenetworks.cfg#power.Loot Fab Power Cost]]: 256 FE/t for the Loot Fabricator.^[21]

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.^[21]
• [[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.^[21]
• [[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.^[21]
• [[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).^[21]
• [[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.^[21]

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.^[21]

Hostile Neural Networks in Project Infinity 0.1 0.0.51.4-hotfix-2 is version 5.3.3 (the file HostileNeuralNetworks-1.20.1-5.3.3.jar).^[2]

The mod was first included in Project Infinity 0.1 0.0.45.4 or earlier, so it has been part of the pack for a while.^[2]

Unverified The details of what changed between Hostile Neural Networks versions, and any pack changelog entries for this mod, could not be confirmed from the available sources. Check the mod's own changelog before assuming a behaviour changed.