Practical AI avatar

Dexa/Practical AI

Learn more
Published May 26, 2021

Elixir meets machine learning

Explore the fusion of Elixir with machine learning as José Valim unveils advancements in collaborative notebooks and the development of Axon, positioning Elixir as a formidable competitor to Python with innovative tools like Livebook and Numerical Elixir enhancing data science workflows.
Episode Highlights
Practical AI logo

Topics covered

Popular Clips

Episode Highlights

  • Building Axon

    The development of Axon, a machine learning library for Elixir, showcases a strategic approach to building complex systems. explains that the foundation of Axon is built on simple functions, allowing for rapid development and easy composition of layers and operations 1. This layered approach contrasts with other libraries that require extensive low-level coding, enabling Axon to support various layers and optimizers quickly. highlights the power of solid abstractions, noting that Axon's API is designed to be familiar to users of other frameworks, which accelerates development 2.

    The foundation is just functions. You're just building functions on top of functions, so it's very easy to compose.

    ---

       

    Community Involvement

    Community involvement plays a crucial role in the evolution of Axon. emphasizes the importance of community interaction, with discussions happening in the Airline Ecosystem Foundation and a machine learning working group 3. This open collaboration fosters innovation and allows for the sharing of ideas and resources. notes that Elixir's community-driven approach has been instrumental in expanding its capabilities beyond web development, into areas like data processing and machine learning 4.

    We have a monthly meeting where we meet and discuss and exchange ideas. So that's definitely the place.

    ---

       

    API Features

    Axon's API is designed to be intuitive for users familiar with other machine learning frameworks. describes how Axon offers a high-level API similar to Keras or PyTorch, making it accessible for those transitioning from other platforms 5. This familiarity encourages experimentation and adoption among developers. highlights the importance of interoperability, noting that Axon's design allows for easy serialization and integration with other tools, which is crucial in the diverse landscape of AI frameworks 6.

    The same level of API convenience that you would expect from Keras or PyTorch is there in Axon.

    ---

Related Episodes