improve documentation

Author: henrykrumb

docs/source/index.rst

@@ -1,6 +1,14 @@
 Welcome to NCALab's documentation!
 ======================================
 
+NCALab is a framework designed to streamline the creation and analysis of Neural Cellular Automata (NCA) implementations.
+With NCALab, users can effortlessly explore various applications, including image segmentation, classification, and synthesis.
+Designed for researchers, developers, and enthusiasts alike, NCALab enables users to explore a variety of
+downstream tasks, including image segmentation, classification, and synthesis.
+
+This documentation serves as a comprehensive guide to understanding and utilizing NCALab's features.
+Here, you'll find detailed instructions, examples, and best practices to help you get started.
+
 .. toctree::
    :maxdepth: 2
    :caption: Contents:

docs/source/intro_nca.rst

@@ -3,11 +3,26 @@ Introduction to Neural Cellular Automata
 
 Neural Cellular Automata (NCA) are a new type of neural network models that merge the concepts
 of Cellular Automata and Artificial Neural Networks.
-In particular, they operate on a cellular grid or graph, applying a learned transition function
-in multiple time steps.
+These models operate on a cellular grid or graph, where each cell interacts with its neighbors
+according to a learned transition function over multiple time steps.
+This allows NCAs to evolve and adapt their states based on local interactions,
+leading to emergent behaviors that can be both interesting to study and useful in practical applications,
+such as medical imaging.
 
-They were originally proposed in 2020 by Mordvintsev et al. in an online publication titled
-`Growing Neural Cellular Automata <https://distill.pub/2020/growing-ca/>`__, which features an online demo
-of a Lizard emoji that emerges from a single black pixel.
+In NCAs, the transition function is typically parameterized by a neural network, which learns to update
+the state of each cell based on its current state and the states of its neighbors.
+This learning process allows NCAs to discover patterns and behaviors that may not be easily defined by
+traditional rule-based systems.
+The evolution of the cellular grid occurs over discrete time steps, where each update can lead to
+changes in the overall configuration.
 
-Recent work on NCA is displayed in a curated `Awesome List <https://github.com/MECLabTUDA/awesome-nca>`__.
+The concept of Neural Cellular Automata was first introduced in 2020 by Mordvintsev et al. in their
+online publication titled `Growing Neural Cellular Automata <https://distill.pub/2020/growing-ca/>`__,
+which features an online demo of a Lizard emoji that emerges from a single black pixel, illustrating
+the nature of NCAs to generate complex structures from minimal initial conditions.
+
+Since their introduction, NCAs have garnered significant interest in the research community.
+A curated collection of recent advancements and resources related to Neural Cellular Automata
+can be found in our `Awesome List <https://github.com/MECLabTUDA/awesome-nca>`__.
+This list includes papers, code repositories, and applications that highlight the versatility
+and potential of NCAs in various fields, including computer graphics, biology, and medical imaging.