10 Years of the PCG workshop: Past and Future Trends
Abstract: As of 2020, the
international workshop on Procedural Content Generation enters its second
decade. The annual workshop, hosted by the international conference on the
Foundations of Digital Games, has collected a corpus of 95 papers published in
its first 10 years. This paper provides an overview of the workshop’s activities
and surveys the prevalent research topics emerging over the years.
Procedural Generation of
Interactive Stories using Language Models
Abstract: In this paper we introduce an architecture, an
implementation and an evaluation of a system for the automatic creation of
interactive stories for games. Our goal is to algorithmically create a branched
story for the entire game; in each game run a different variant is generated.
The architecture is based on natural language processing (NLP) to generate
meaningful stories. For NLP we use a statistical language model based on a
neural network (Generative Pretrained Transformer, GPT-2). The basic stories
generated in this way tend to have too many different persons, and they tend to
get incoherent for longer texts, so we add a component restricting the number of
persons and improving the consistency. The system is initialized with a
hand-written game introduction that defines the main characters and the
inventory; it also sets the goals for the game. From that text the remainder of
the game story is generated algorithmically. We have fully implemented our
system, and we report initial, encouraging experimental results.
Procedural generation using quantum computation
Abstract: Quantum computation is an emerging technology
that promises to be a powerful tool in many areas. Though some years likely
still remain until significant quantum advantage is demonstrated, the
development of the technology has led to a range of valuable resources. These
include publicly available prototype quantum hardware, advanced simulators for
small quantum programs and programming frameworks to test and develop quantum
software. In this provocation paper we seek to demonstrate that these resources
are sufficient to provide the first useful results in the field of procedural
generation. This is done by introducing a proof-of-principle method: a quantum
generalization of a blurring process, in which quantum interference is used to
provide a unique effect. Through this we hope to show that further developments
in the technology are not required before it becomes useful for procedural
generation. Rather, fruitful experimentation with this new technology can begin
now.
Procedural Content Generation of Puzzle Games using Parameterized Generative Adversarial Networks
Abstract: In
this article, we present a preliminary analysis of using parameterized
Generative Adversarial Networks (GANs) to produce levels for the puzzle game
Lily’s Garden\footnote{https://tactilegames.com/lilys-garden/}, based on an
existing body of human-made levels. This could be greatly beneficial in the
context of Procedural Content Generation (PCG) and possibly assist
level-designers create new levels more efficiently.
As Lily’s Garden has many different unique board pieces (henceforth pieces) and further combinations of those pieces, we reduce the data for the levels to a representation with 8 unique pieces and extract two condition-vectors from the real levels in an effort to give a potential designer control over the output of the GANs. The conditions are level-shape and piece-distribution.
The GANs successfully approximates the level shape given to it, even on shapes that it has not been trained on, but fails to approximate the piece-distribution vector. The GANs are also evaluated on two tests, more specific to the puzzle-game genre: testing for color-islands (pieces with the same color are connected, making them clickable), and testing for broken pieces (which means that there are color- or blocker pieces outside of the level shape).
While the GANs performs well in the first test it struggles greatly with the latter and we suggest that this might be improved by trying out alternative architectures for both the Generator and Discriminator of the GANs. Further, we hypothesize that the GANs might better pick up on the piece-distribution vector if we enrich the dataset with more channels than the current reduced problem.
As Lily’s Garden has many different unique board pieces (henceforth pieces) and further combinations of those pieces, we reduce the data for the levels to a representation with 8 unique pieces and extract two condition-vectors from the real levels in an effort to give a potential designer control over the output of the GANs. The conditions are level-shape and piece-distribution.
The GANs successfully approximates the level shape given to it, even on shapes that it has not been trained on, but fails to approximate the piece-distribution vector. The GANs are also evaluated on two tests, more specific to the puzzle-game genre: testing for color-islands (pieces with the same color are connected, making them clickable), and testing for broken pieces (which means that there are color- or blocker pieces outside of the level shape).
While the GANs performs well in the first test it struggles greatly with the latter and we suggest that this might be improved by trying out alternative architectures for both the Generator and Discriminator of the GANs. Further, we hypothesize that the GANs might better pick up on the piece-distribution vector if we enrich the dataset with more channels than the current reduced problem.
M.I.N.U.E.T.:
Procedural Musical Accompaniment for Textual Narratives
Abstract: Extensive research has been conducted on using
procedural music generation in real-time applications such as accompaniment to
musicians, visual narratives, and games. However, less attention has been paid
to the enhancement of textual narratives through music. In this paper, we
present Mood Into Note Using Extracted Text (MINUET), a novel system that can
procedurally generate music for textual narrative segments using sentiment
analysis. Textual analysis of the flow and sentiment derived from the text is
used as input to condition accompanying music. Music generation systems have
addressed variations through changes in sentiment. By using an ensemble
predictor model to classify sentences as belonging to particular emotions,
MINUET generates text-accompanying music with the goal of enhancing a reader’s
experience beyond the limits of the author’s words. Music is played via the
JMusic library and a set of Markov chains specific to each emotion with mood
classifications evaluated via stratified 10-fold cross validation. The
development of MINUET affords the reflection and analysis of features that
affect the quality of generated musical accompaniment for text. It also serves
as a sandbox for further evaluating sentiment-based systems on both text and
music generation sides in a coherent experience of an implemented and extendable
experiential artifact.
Spatial Layout of Procedural Dungeons Using Linear Constraints and
SMT Solvers
Abstract: Dungeon
generation is among the oldest problems in procedural content generation.
Creating the spatial aspects of a dungeon requires three steps: random
generation of rooms and sizes, placement of these rooms inside a fixed area, and
connecting rooms with passageways. This paper uses a series of integer linear
constraints, solved by an SMT solver, to perform the placement step. Separation
constraints ensure dungeon rooms do not intersect and maintain a minimum fixed
separation. Designers can specify control lines, and dungeon rooms will be
placed within a fixed distance of these control lines. Generation times vary
with number of rooms and constraints, but are often very fast. Spatial
distribution of solutions tend to have hot spots, but is surprisingly uniform
given the underlying complexity of the solver. The approach demonstrates the
effectiveness of a declarative approach to dungeon layout generation, where
designers can express desired intent, and the SMT solver satisfies this if
possible.
Sequential Segment-based Level
Generation and Blending using Variational Autoencoders
Abstract: Existing methods of level generation using
latent variable models such as VAEs and GANs do so in segments and produce the
final level by stitching these separately generated segments together. In this
paper, we augment such methods by training VAEs to learn a sequential model of
segment generation such that generated segments logically follow from prior
segments. By further combining the VAE with a classifier that determines whether
to place the generated segment to the top, bottom, left or right of the previous
segment, we obtain a pipeline that enables the generation of arbitrarily long
levels that progress in any of these four directions and are composed of
segments that logically follow one another. In addition to generating more
coherent levels of non-fixed length, this method also enables implicit blending
of levels from separate games that do not have similar orientation. We
demonstrate our approach using levels from Super Mario Bros., Kid Icarus and
Mega Man, showing that our method produces levels that are more coherent than
previous latent variable-based approaches and are capable of blending levels
across games.
Tabletop Roleplaying
Games as Procedural Content Generators
Abstract: Tabletop roleplaying games (TTRPGs) and
procedural content generators can both be understood as systems of rules for
producing content. In this paper, we argue that TTRPG design can usefully be
viewed as procedural content generator design. We present several case studies
linking key concepts from PCG research — including possibility spaces,
expressive range analysis, and generative pipelines — to key concepts in TTRPG
design. We then discuss the implications of these relationships and suggest
directions for future work uniting research in TTRPGs and PCG.
Multi-Objective level generator generation with
Marahel
Abstract: Procedural level
generation is a hard problem. Game designers usually design their constructive
level generators using game knowledge. This paper introduces a new system to
design constructive level generators by searching the space of constructive
level generators defined by Marahel language. We use NSGA-II, a multi-objective
optimization algorithm, to search for generators for three different problems
(Binary, Zelda, and Sokoban). We restrict the representation to a subset of
Marahel language to push the evolution to find more efficient generators. The
results show that the generated generators were able to achieve a good
performance on most of the fitness function over these three problems.