diff --git a/README.md b/README.md
index 3833826..44b795c 100644
--- a/README.md
+++ b/README.md
@@ -1,11 +1,11 @@
 # PyTorch NEAT
 
 ## Background
-NEAT (NeuroEvolution of Augmenting Topologies) is a popular neuroevolution algorithm, one of the few such algorithms that evolves the architectures of its networks in addition to the weights.
+NEAT (NeuroEvolution of Augmenting Topologies) is a popular neuroevolution algorithm, one of the few such algorithms that evolves the architectures of its networks in addition to the weights. For more information, see this research paper: http://nn.cs.utexas.edu/downloads/papers/stanley.ec02.pdf.
 
-HyperNEAT is an extension to NEAT that indirectly encodes the weights of the network (called the substrate) with a separate network (called a CPPN, for compositional pattern-producing network).
+HyperNEAT is an extension to NEAT that indirectly encodes the weights of the network (called the substrate) with a separate network (called a CPPN, for compositional pattern-producing network). For more information on HyperNEAT, see this website: http://eplex.cs.ucf.edu/hyperNEATpage/.
 
-Adaptive HyperNEAT is an extension to HyperNEAT which indirectly encodes both the initial weights and an update rule for the weights such that some learning can occur during a network's "lifetime."
+Adaptive HyperNEAT is an extension to HyperNEAT which indirectly encodes both the initial weights and an update rule for the weights such that some learning can occur during a network's "lifetime." For more information, see this research paper: http://eplex.cs.ucf.edu/papers/risi_sab10.pdf.
 
 ## About
 PyTorch NEAT builds upon [NEAT-Python](https://github.com/CodeReclaimers/neat-python) by providing some functions which can turn a NEAT-Python genome into either a recurrent PyTorch network or a PyTorch CPPN for use in HyperNEAT or Adaptive HyperNEAT.