Facebook has released an explainer on artificial intelligence (AI) with commentary from Yann Lecun, the company's head of AI research.
This consists of a written post outlining the principles of AI, and a video series briefly explaining key concepts like machine learning, gradient descent, deep learning, back propagation, and convolutional neural networks. Each of the above mentioned techniques are central to Facebook's progress in image-object, text, and speech recognition.
At the highest level, Facebook describes the three types of learning in AI:
- Reinforcement learning. Inspired by behavioral psychology, this area of machine learning focuses on reward-based decision-making — in other words, teaching machines to take actions in the pursuit of a reward. Reinforcement learning is frequently used when training machines to play and win at games like chess, Go, and video games. Its principal disadvantage is that an extremely large number of trials are needed for the machine to learn a simple task.
- Supervised learning. The most common mode of learning for a machine, supervised learning is like showing a child how to recognize an object by showing them a picture book, where the adult knows the answer and the child learns by observation and experience. At first, the machine won’t know how to distinguish the object, but will learn to after thousands or millions of trial runs with different labeled images. Eventually, the machine will be able to recognize the object even in images that it has never seen before, achieving what’s called “generalization ability.”
- Unsupervised/predictive learning. This refers to the kind of learning carried out naturally by humans and animals, which occurs spontaneously, in an unsupervised manner, through observation, experience and intuition, over the course of a lifetime. Humans don’t know how to teach machines at this level yet, not in a way that’s similar to humans or animals. Our dearth of understanding in this space – our lack of techniques for unsurprised or predictive AI learning – is a major stumbling block in AI right now.
AI is one of three areas in Facebook's 1o-year innovation roadmap, alongside greater connectivity initiatives, and virtual and augmented reality. Over 40 teams at the company and more than a quarter of its engineers use AI in the products they build. Facebook CEO Mark Zuckerberg touched on this in the Q3 earnings call:
- Linguistic understanding. This includes initiatives to read and understand articles and posts on the platform, messages between users and businesses to create appropriate auto-reply functions for chatbots, surfacing relevant and interesting content in news feed.
- Visual understanding. This involves understanding what’s in photos and videos, what people are doing inside the videos, and the objects in the scene, which can then be used to help visually impaired people, as well as create better rankings in the news feed.
Advancements in artificial intelligence, coupled with the proliferation of messaging apps, are fueling the development of chatbots — software programs that use messaging as the interface through which to carry out any number of tasks, from scheduling a meeting, to reporting weather, to helping users buy a pair of shoes.
Foreseeing immense potential, businesses are starting to invest heavily in the burgeoning bot economy. A number of brands and publishers have already deployed bots on messaging and collaboration channels, including HP, 1-800-Flowers, and CNN. While the bot revolution is still in the early phase, many believe 2016 will be the year these conversational interactions take off.
Laurie Beaver, research associate for BI Intelligence, Business Insider's premium research service, has compiled a detailed report on chatbots that explores the growing and disruptive bot landscape by investigating what bots are, how businesses are leveraging them, and where they will have the biggest impact.
The report outlines the burgeoning bot ecosystem by segment, looks at companies that offer bot-enabling technology, distribution channels, and some of the key third-party bots already on offer. The report also forecasts the potential annual savings that businesses could realize if chatbots replace some of their customer service and sales reps. Finally, it compares the potential of chatbot monetization on a platform like Facebook Messenger against the iOS App Store and Google Play store.
Here are some of the key takeaways:
- AI has reached a stage in which chatbots can have increasingly engaging and human conversations, allowing businesses to leverage the inexpensive and wide-reaching technology to engage with more consumers.
- Chatbots are particularly well suited for mobile — perhaps more so than apps. Messaging is at the heart of the mobile experience, as the rapid adoption of chat apps demonstrates.
- The chatbot ecosystem is already robust, encompassing many different third-party chat bots, native bots, distribution channels, and enabling technology companies.
- Chatbots could be lucrative for messaging apps and the developers who build bots for these platforms, similar to how app stores have developed into moneymaking ecosystems.
In full, the report:
- Breaks down the pros and cons of chatbots.
- Explains the different ways businesses can access, utilize, and distribute content via chatbots.
- Forecasts the potential impact chatbots could have for businesses.
- Looks at the potential barriers that could limit the growth, adoption, and use of chatbots.
To get your copy of this invaluable guide, choose one of these options:
- Subscribe to an ALL-ACCESS Membership with BI Intelligence and gain immediate access to this report AND over 100 other expertly researched deep-dive reports, subscriptions to all of our daily newsletters, and much more. >> START A MEMBERSHIP
- Purchase the report and download it immediately from our research store. >> BUY THE REPORT
The choice is yours. But however you decide to acquire this report, you’ve given yourself a powerful advantage in your understanding of chatbots.
