Conversation starter: AI
Image created using DALL-E

Conversation starter: AI

For my first post, I want to share with you an analogy that I recently used as an introduction to the vocabulary of artificial intelligence (AI). In just over a year, AI has penetrated every sector and is being utilized in new and creative ways every day. Many of us are grappling with the implications of this AI revolution: How can I harness this technology? What can it enable me to do better or faster than I could do before? Will it impact my job??

As with any tool, in order to use it you need to have a basic understanding of it. You can’t drive a car without a basic understanding of what the steering wheel, accelerator, and brake pedal do (at minimum). You don’t need to be able to build or even repair the car, but you do need to know the difference between the pedal on the right and the pedal on the left. Similarly, you don’t need to know how to write code or how to train a deep learning model in order to use powerful tools like ChatGPT. But you do need to know the difference between a neural network and generative AI. So let’s get started.

Have you ever seen or read about those advanced computer interfaces in sci-fi stories, like HAL-9000 in 2001: A Space Odyssey or J.A.R.V.I.S. in Iron Man? They incorporate many of the concepts we talk about in today’s AI landscape. Let’s take a look:

  • People just talk with these computer interfaces normally, like they would talk with a colleague or friend, without using any set commands or code — that’s natural language processing (NLP).?
  • These computers were trained on large collections of documents and other data that they can now access in the blink of an eye, like all the world's books or all the laws of the land — that’s a large language model (LLM). While language-based models were the first to capture the public’s imagination, they are not the only ones being developed. Many companies are working to develop similar models that can comb through collections of images, videos, and other types of information.?
  • When the inner workings of these computers are described, they are usually shown as a complex network with layer upon layer of connections. This architecture mimics the network of cells, called neurons, that encode information in our brain. That is why they are called neural networks. It doesn’t mean that they are made out of actual brain cells, just that they are set up in a similar way.
  • In these sci-fi stories, people can interact with these artificial intelligences and ask them to use what they know to create something new, like a plan of attack against the big bad enemy — that’s generative AI. When you ask a chatbot to write a weather report in the form of a sonnet, chances are that it is creating (generating) the response from scratch rather than sifting through a collection of existing sonnets and choosing the most appropriate one to show you.

What other terms have you heard in conversations about artificial intelligence? What descriptions or explanations have helped you understand them? Please add your comments, questions, and ideas below!

Stephanie Allen, PhD

Scientific Leadership | Community Impact

12 个月

Irit Rappley, thank you for your post, what a great conversation starter! I've been fascinated by Meta-learning - machine learning aimed at developing models or algorithms that can learn how to learn, and reading Andy Lee's post sparked my curiosity how are Meta-Learning and Agents different? There are similarities, like integrating environment feedback, the purpose and focus of the models are different though, meta-learning acquires knowledge or strategies to adapt to new tasks, agents interacts with the environment to take actions and improve decision-making overtime. Does that make meta-learning the introvert, and agents the extrovert AI? ??

Great insights! Looking forward to more posts on these topics! ??

Richard Parr

Futurist | Advisor | Speaker | Author | Educator Generative AI - AI Governance - Human Centered AI - Quantum ML - Quantum Cryptography - Quantum Robotics - Neuromorphic Computing - Space Innovation - Blockchain

1 年

Looking forward to the insightful discussions ahead!

Devan M. Monroe

Sr Development Officer, MIT Brain and Cognitive Sciences

1 年

Irit, thanks for posting this. You’ve written about this in a way that’s accessible for folks who aren’t familiar with the science but are bombarded with these buzzwords.

要查看或添加评论,请登录

Irit Rappley的更多文章

  • It's Not Rocket Science

    It's Not Rocket Science

    In research as well as management, we often aim to accomplish something really big. It’s often so big that it feels…

  • How can we deal with failure?

    How can we deal with failure?

    I was a finalist in an interview process but didn’t get the job. I was an executive at a startup that failed to close…

    12 条评论
  • An Introvert's Guide to Networking at JPM

    An Introvert's Guide to Networking at JPM

    JPM week is here, and tens of thousands of biopharma and healthcare professionals are descending on San Francisco for…

    12 条评论
  • The Future of AI in Drug Discovery

    The Future of AI in Drug Discovery

    As we wrap up 2024 and look ahead to 2025 and beyond, it is clear that artificial intelligence has become an…

    4 条评论
  • Some thoughts on leadership

    Some thoughts on leadership

    I’ve had a lot of conversations about leadership recently and wanted to share some thoughts here, as a starting point…

    1 条评论
  • What are chatbots?

    What are chatbots?

    You’ve heard the hype about chatbots, natural language processing (NLP), and generative AI. You’re ready to give it a…

社区洞察

其他会员也浏览了