Difficulty Parsing Language? Check Out Google's New AI

In general, humans by nature are good at parsing sentences which means having the ability to identify and comprehend individual parts and functions of a sentence to understand the whole meaning of a statement. This ambiguity in spoken and written language makes parsing a challenge for artificial intelligence (AI) programs. Here comes Google attempting to give AI more human-like attributes- the ability to grasp human speech and text.

Recently, the Silicon Valley search giant is offering the source code for parsing called SyntaxNet and Parsey McParseface- the open-source software programming toolkits Google researchers have initially developed. These pieces of software Google shared for free will enable other developers to create programs or apps that will better understand the complexities of English, the way humans use it.

"Our hope is that people will just use this instead of building their own," said Dave Orr, SyntaxNet's product manager as quoted by the Wall Street Journal. "They don't have to reinvent the wheel."

Google's latest AI innovation for natural language understanding (NLU) has the ability to parse sentences with more than 94% accuracy which is quite a milestone in Silicon Valley's seemingly relentless pursuit for developing human-like AI.

Parsey McParseface has its foundations in SyntaxNet, the neural network framework considered as the first main element for NLU. By open-sourcing the code, non-Google developers would now be entitled to using the framework to enhance the ability of computer models to comprehend human language.

"Because Parsey McParseface is the most accurate such model in the world, we hope that it will be useful to developers and researchers interested in automatic extraction of information, translation, and other core applications of NLU," Google senior staff research scientist Slav Petrov wrote on Google blog as quoted by CIO Today

This is not the first time Google open sourced its AI technology. Last year, it released the programming code for TensorFlow which allows anyone to link computing devices into a neural network that can process data the way human brains work. Likewise, Facebook is venturing into AI and open-sourced its Torch software to enhance machine learning and vision.

In a nutshell, artificial intelligence is the scientific quest of making computers or machines think and act more like humans. Just recently, Google made a breakthrough when Google's other AI software called DeepMind managed to defeat a South Korean grandmaster in the board game called 'Go'. For quite some time, AI found it difficult to defeat humans since the oriental board game is far more complex and extremely unpredictable than Western chess.

DeepMind's Go victory has broader implications in terms of how AI will make a difference in the world to come.

"This is another important step toward creating artificial intelligence that can help us in everything from accomplishing our daily tasks and travels, to eventually tackling even bigger challenges like climate change and cancer diagnosis," Google CEO Sundar Pichai said as quoted by CNET.