(cover photo by Andy Kelly)
Welcome to the very first installment of what I plan to be a weekly round-up of everything I've found relevant and noteworthy on the topics of AI, machine learning and chatbots. Expect these posts to highlight news from Microsoft products and services, and also include general, vendor-agnostic information.
- Microsoft creates AI that can read a document and answer questions about it as well as a person: For the first time, an AI model has beaten human performance on the SQuAD test, which assesses how well a system (or a person) can understand Wikipedia articles and answer questions about them.
- New hire or new bot?: It's a bold idea to create a bot and integrate it in your team as a normal coworker, but it's even more audacious to do that without telling the team that their new teammate is actually a bot! Someone at Microsoft did just that and there's a lot to learn about the outcome of the experiment.
- The Future Computed: Artificial Intelligence and its role in society: A free e-book discussing the growing importance of AI in our lives and the key challenges AI developers are going to face as this technology integrates in our society.
- Slack hopes its AI will keep you from hating Slack: Slack has been widely adopted in our industry but managing enterprise communications as chats can become messy pretty quickly... That's why a team of 20 people is applying machine learning and natural language processing techniques to help you navigate your many chatrooms and identify the content that's important and relevant to you.
- What most people don’t understand about AI and the state of machine learning: We can't deny that AI is both overhyped (as a buzzword, overused and often misused) and misunderstood (most people have limited understanding of what machine learning is capable of). This article tries to reset the definitions of AI and machine learning by taking an objective look at what they are and what they can achieve today.
- Microsoft researchers build a bot that draws what you tell it to: What happens when you ask an AI model to draw "a red double decker bus floating on top of a lake"? Surprising results for sure, and some already dare to call it Artificial Imagination!
- The 8 neural network architectures Machine Learning researchers need to learn: A technical, yet very accessible article covering 8 core machine learning techniques, from the basic to the cutting-edge. Each technique is explained in great details and illustrated with practical examples. Lengthy but great read!
- What you need to create great customer journeys (with tech): A short article discussing the key benefits of using a chatbot to interact with your customers.
- Google’s new cloud service lets you train your own AI tools, no coding knowledge required: Google has released Cloud AutoML Vision, an online service that lets you build image classifiers in just a couple of clicks (very similar to Microsoft's Custom Vision Service).
- LUIS quick start with list entities: A step-by-step example showing how to build a LUIS model using list entities.
- Registration for the ML4ALL conference is open: Happening from April 8th to 10th in Portland, Oregon.