Businesses are increasingly talking to their customers using instant messaging and social media. These conversations contain valuable insight into the underlying causes of user behaviour. However, this channel is significantly underused because of the challenges posed by analysing informal, poorly spelt, user generated text. Traditional natural language processing is based on rules and hand engineered features which makes it too inflexible. On-boarding new languages requires human expert knowledge making it also prohibitively expensive.

Advancements in AI were recently spotlighted by AlphaGo, a computer program developed by Google's DeepMind team. The program -- which relies on decision-making algorithms and neural networks -- defeated a human European champion at the board game Go in a feat previously believed to be years away. On the investor side, Jim Breyer of Breyer Capital has said AI will deliver massive returns for investors betting on applications for industries including healthcare and entertainment. With this in mind, we used CB Insights' database to look at funding to artificial intelligence startups since 2010. Our artificial intelligence category covers startups primarily focused on developing AI, across areas including image processing, natural language processing, machine learning, deep learning, and predictive APIs, among other core applications.

MNIST is the "Hello World" of machine learning. With this tutorial, you can train on MNIST with a single-layer neural network. By applying a more complex algorithm, Lenet, to MNIST, you can achieve 99 percent accuracy. Lenet is a deep convolutional network. Word2vec is a popular natural-language processing algorithm capable of creating neural word embeddings, which in turn can be fed into deeper neural networks.

Elon Musk is develop artificial intelligence which will enable robots that can do housework, have conversations and play games. OpenAI's mission is to build safe AI, and ensure AI's benefits are as widely and evenly distributed as possible. OpenAI will measure intelligence using a metric which consists of a variety of OpenAI Gym environments with a unified action and observation space (so a single agent can run across all of them), including games, robotics, and language-based tasks. Their implementation will evolve over time, and they'll keep the community updated along the way They are working to enable a physical robot (off-the-shelf; not manufactured by OpenAI) to perform basic housework. There are existing techniques for specific tasks, but we believe that learning algorithms can eventually be made reliable enough to create a general-purpose robot.

If J.K. Rowling was ever concerned that artificial intelligence could do her job for her, she let out a sigh of relief this week. A fan of the Harry Potter series in San Francisco had an artificial intelligence computer algorithm write a few chapters for the iconic series, after teaching the computer to read earlier novels. The results are far from ideal. Max Deutsch, a product manager at Intuit, trained a a Long Short Term Memory (LSTM) Recurrent Neural Network computer by teaching it to read the first four Harry Potter novels. A LSTM computer is trained to notice patterns (say, in genomes or handwriting), which makes it a great test subject for writing patterns.

Video has taken the world by storm with a myriad of intelligent devices continuously capturing vast amounts of data about how people live and what they do. At the end of 2014, IHS Technology estimated over 245 million operational cameras were active globally. London alone has 500,000 cameras dotted throughout the city, which works out at about one camera for every 16 people. Thanks to smart cameras, CCTV devices, and even drones mounted with intelligent cameras, users are able to record videos at an unprecedented scale and pace. This vast store of data-rich content is used for a range of purposes – from gaming and law enforcement to crowd management at large events.

THE TERM "artificial intelligence" has been associated with hubris and disappointment since its earliest days. It was coined in a research proposal from 1956, which imagined that significant progress could be made in getting machines to "solve kinds of problems now reserved for humans…if a carefully selected group of scientists work on it together for a summer". That proved to be rather optimistic, to say the least, and despite occasional bursts of progress and enthusiasm in the decades that followed, AI research became notorious for promising much more than it could deliver. Researchers mostly ended up avoiding the term altogether, preferring to talk instead about "expert systems" or "neural networks". But in the past couple of years there has been a dramatic turnaround.

Google plans to use more than one million anonymized eye scans to teach computers how to diagnose ocular disease. The Menlo Park, Calif.-based company has signed a deal with a British eye hospital to use artificial intelligence to learn from the medical records of 1.6 million patients in London hospitals. The goal is to teach a computer program to recognize the signs of two common types of eye disease, diabetic retinopathy and age-related macular degeneration. That's something humans are surprisingly imperfect at. Physicians diagnose these ailments by analyzing medical charts and interviewing patients, yet still get it wrong 10 to 20% of the time.

The artificial intelligence software is learning how to recognize early signs of two eye diseases.Video provided by Newsy Newslook A scan of a human eye. SAN FRANCISCO -- Google plans to use more than one million anonymized eye scans to teach computers how to diagnose ocular disease. The Menlo Park, Calif.-based company has signed a deal with a British eye hospital to use artificial intelligence to learn from the medical records of 1.6 million patients in London hospitals. The goal is to teach a computer program to recognize the signs of two common types of eye disease, diabetic retinopathy and age-related macular degeneration. That's something humans are surprisingly imperfect at.

One of the difficult problems in using machine learning to generate sequences, such as melodies, is creating long-term structure. Long-term structure comes very naturally to people, but it's very hard for machines. Basic machine learning systems can generate a short melody that stays in key, but they have trouble generating a longer melody that follows a chord progression, or follows a multi-bar song structure of verses and choruses. Likewise, they can produce a screenplay with grammatically correct sentences, but not one with a compelling plot line. Without long-term structure, the content produced by recurrent neural networks (RNNs) often seems wandering and random.