Other articles


  1. Best things of 2015

    Here are a few things I appreciated in 2015:

    arXiv papers:

    • Batch Normalization- a simple conceptual idea beautifully executed. We've known for years that neural networks respond better to input data that follow certain distributions. Since the output of each layer is the input to the next, it's a good …

    read more
  2. Overfitting, Regularization, and Hyperparameters

    (This lays the ground work for the next post, which was getting too long to be effective.)

    Overfitting

    One of the goals of machine learning is generalizability. A model that only works on the exact data it was trained on is effectively useless. Let's say you're tasked with creating a …

    read more
  3. Talking Talking Machines

    Talking Machines is currently THE machine learning podcast. Last fall when it started out, I was intrigued that someone was starting a podcast focusing specifically on machine learning. But I didn't really listen until this summer, which was my loss. Now that I'm catching up on it, I think it's …

    read more
  4. Why Blog?

    A question anyone starting a blog in 2015 should answer:

    Why?

    Like a lot of people, I think things through more thoroughly and with more discipline when writing them down. If I just consider a topic, I can tend to get into habits of thought. So the act of writing …

    read more

links

social