Some most interesting open-source projects in artificial intelligence and machine learning

 

• OpenCog Foundation
  In their words: We’re undertaking a serious effort to build a thinking machine.
  This organisation is the best I have ever encountered so far. I am currently contributing to this organisation. I am working on the language learning project and I hope to be selected for GSoC 2015 through this organisation.
  They plan to teach a machine a language. They plan to make it intelligent enough to play games. They are working on Natural Language Processing and many other areas related to AI.
  You should check out their roadmap to see what they plan to do. It is certainly impressive.
• Battle for Wesnoth
  This is an open source game for both Linux and Windows. 
  It is a strategy game and it is very interesting. The graphics side can do with a bit of improvement but they are still impressive considering that they are open source.
  This game has a strong AI component. In fact, coding an AI for this game is notoriously difficult given the flexibility which this game gives to the players.
• Natural Language Toolkit
  This Natural Language Toolkit is written in Python. It has dependency parsing, part-of-speech tagging and sentiment analysis. They plan to integrate Twitter too in this project.
  I would certainly recommend you this project if you are interested in NLP.
• Institute for Artificial Intelligence
  This organisation researches in Artificial Intelligence. It has many projects involving robotics which actually involve modifying Robot Operating Systems (ROS). They got selected for GSoC 2015 and their ideas page has some interesting projects.
• Theano  (http://deeplearning.net/software…) is a CPU and GPU compiler  for mathematical expressions in Python. It combines the convenience of  NumPy with the speed of optimized native machine language. For  gradient-based machine learning algorithms (like training an MLP or  convolutional net), Theano is from 1.6x to 7.5x faster than competitive alternatives (including those in C/C++, NumPy, SciPy, and Matlab) when compiled for the CPU and between 6.5x and 44x faster  when compiled for the GPU. You can read more about it in this SciPy  2010 paper:  http://www.iro.umontreal.ca/~lis…

• TensorRT – The fastest inference library on GPU, developed by NVIDIA and using highly-optimized on the lowest level algoritms.

  More info: https://developer.nvidia.com/ten…

 

AbnAsia.org Software. Faster. Better. More Reliable. +84945924877 (Asia# Mobile, WhatsApp, Telegram, Viber, Zalo); +16699996606 (US# Mobile, WhatsApp, Telegram) [email protected]