ASpiReNN is a little C library (with Python bindings) which provides support for simple (leaky integrate-and-fire) spiking neural networks. It is primarily designed for highly recurrent networks, but it can also be used with multi-layer nets, though performance won't be the same. Though only Leaky integrate-and-fire (for the neurons) and Spike-Timing Dependent Plasticity (for learning rules) are currently implemented, adding new models shouldn't be too difficult.
Trad4 is a fully concurrent, thread safe, graph based programming language that scales linearly on multiple cores. It is initially intended for deployment in the financial industry to model real-time risk. Trad4 is proposed as an alternative to the Von Neumann model of computer architecture. It is a new way of arranging programs in memory and a new style of flow-of-control.
PyMVPA (Python MultiVariate Pattern Analysis) is a Python module intended to ease pattern classification analyses of large data sets. It provides high-level abstraction of typical processing steps and a number of implementations of some popular algorithms. While it is not limited to the neuroimaging domain, it is eminently suited for such data sets. It requires nothing but free software to run.
Evolving Games for Unnatural Intelligence is a Java package for unsupervised machine learning based on Evolutionary Game Theory on directed graphs. It is able to segment data without any previuos information on the number of segments. It has no GUI, but implements generalizations of the original method proposed by Li, Chen, He and Jiang in the arxiv paper "A Novel Clustering Algorithm Based Upon Games on Evolving Network", published on 30 Dec 2008.
PSIworld (Programmable Scalable Interactive World) is a framework that provides a set of libraries, utilities, and applications to ease the implementation of dynamic artificial intelligence environments. Specifically, the development of multi-agent applications is targeted. It is designed in a generic manner so that various kinds of applications can use it. This framework not only includes pure algorithm libraries for AI computation tasks, but also a C/C++ library for distributed computation. A server-client model also involves various, concurrent visualization methods of distinct Agents or Societies.