Brandon Rohrer

@_brohrer_

Teaching. Writing. Machine learning. Data science. Algorithms. Robotics. Facebook. Microsoft. MIT PhD.

Boston
brohrer.github.io
ಆಗಸ್ಟ್ 2012 ಸಮಯದಲ್ಲಿ ಸೇರಿದ್ದಾರೆ
250 ಫೋಟೋಗಳು ಅಥವಾ ವೀಡಿಯೋಗಳು ಫೋಟೋಗಳು ಮತ್ತು ವೀಡಿಯೋಗಳು

ಟ್ವೀಟ್‌ಗಳು

ನೀವು @_brohrer_ ಅವರನ್ನು ತಡೆಹಿಡಿದಿರುವಿರಿ

ಈ ಟ್ವೀಟ್‌ಗಳನ್ನು ವೀಕ್ಷಿಸಲು ನೀವು ಖಚಿತವಾಗಿ ಬಯಸುವಿರಾ? ಟ್ವೀಟ್ ವೀಕ್ಷಣೆಯು @_brohrer_ ಅವರ ತಡೆತೆರವುಗೊಳಿಸುವುದಿಲ್ಲ

  1. ಪಿನ್ ಮಾಡಿದ ಟ್ವೀಟ್
    ಮಾರ್ಚ್ 2,2017

    How Work (including backpropagation!) is the latest addition to my video tutorial library. Enjoy.

    16 replies 108 ಮರುಟ್ವೀಟಿಸುವಿಕೆಗಳು 309ಇಷ್ಟಗಳು
    ರದ್ದುಗೊಳಿಸು
  2. ಡಿಸೆಂ 19

    More from the gallery

    3ಇಷ್ಟಗಳು
    ಈ ಥ್ರೆಡ್ ತೋರಿಸಿ
    ಈ ಥ್ರೆಡ್ ತೋರಿಸಿ
    ರದ್ದುಗೊಳಿಸು
  3. ಡಿಸೆಂ 19

    I’ve learned lots of cool things putting this together! Among the coolest: linear regression and logistic regression are just small, special types of neural networks.

    5 replies 23ಇಷ್ಟಗಳು
    ಈ ಥ್ರೆಡ್ ತೋರಿಸಿ
    ಈ ಥ್ರೆಡ್ ತೋರಿಸಿ
    ರದ್ದುಗೊಳಿಸು
  4. ಡಿಸೆಂ 19

    I'm working on a blog/video about What Neural Networks Learn. Here's a sneak peak.

    4 replies 5 ಮರುಟ್ವೀಟಿಸುವಿಕೆಗಳು 64ಇಷ್ಟಗಳು
    ಈ ಥ್ರೆಡ್ ತೋರಿಸಿ
    ಈ ಥ್ರೆಡ್ ತೋರಿಸಿ
    ರದ್ದುಗೊಳಿಸು
  5. ಡಿಸೆಂ 18

    This is a beautiful example of a curiosity-driven project. It also shows the versatility of logistic regression.

    "I would love to know how many minutes, hours, days there are of us talking" asked, in episode 100. I was curious how long each host had spoken in total, so I used machine learning to find out.
    ಈ ಥ್ರೆಡ್ ತೋರಿಸಿ
    6ಇಷ್ಟಗಳು
    ರದ್ದುಗೊಳಿಸು
  6. ಡಿಸೆಂ 17

    The migrating boundary between human and animal cognition reminds me of the migrating boundary between biological and machine cognition.

    Interesting discussion here about scientific tests of self-awareness. Whatever a mirror test does or doesn’t say, I find it fascinating how hard we humans work to try to convince ourselves we’re categorically separate from other animals, vs on a continuum.
    1 ಮರುಟ್ವೀಟಿಸುವಿಕೆ 4ಇಷ್ಟಗಳು
    ರದ್ದುಗೊಳಿಸು
  7. ಡಿಸೆಂ 15

    This is what exploring new directions in ML looks like. Acknowledging limitations of current methods. Referring back to biology. Mostly negative results. This is growth.

    1 ಮರುಟ್ವೀಟಿಸುವಿಕೆ 17ಇಷ್ಟಗಳು
    ರದ್ದುಗೊಳಿಸು
  8. ಡಿಸೆಂ 15

    Also from Hinton: “Now if you send in a paper that has a radically new idea, there's no chance in hell it will get accepted .... Anything that makes the brain hurt is not going to get accepted.”

    Hinton on ML research: “We should be going for radically new ideas. We know a radically new idea in the long run is going to be much more influential than a tiny improvement. That’s the downside when we only have a few senior guys and a gazillion juniors.”
    3 replies 10 ಮರುಟ್ವೀಟಿಸುವಿಕೆಗಳು 49ಇಷ್ಟಗಳು
    ರದ್ದುಗೊಳಿಸು
  9. ಡಿಸೆಂ 14

    It’s a learning bonanza

    While our content on Pluralsight is always free, this weekend EVERYTHING on Pluralsight is unlocked. Take some time to learn something new this weekend! Thanks and team.
    2ಇಷ್ಟಗಳು
    ರದ್ದುಗೊಳಿಸು
  10. ಡಿಸೆಂ 14

    This ambiguity has been a source of confusion for me: "sigmoid" literally means "s-shaped". It includes both logistic and hyperbolic tangent functions, but sometimes it's used specifically to mean the logistic function.

    2 replies 8ಇಷ್ಟಗಳು
    ರದ್ದುಗೊಳಿಸು
  11. ಡಿಸೆಂ 13

    My new polynomial regression course on is project-based. It walks students through a realistic ML use case from start to finish. Educators, you are welcome to evaluate it (and all my other courses) at no cost. Just message me for the coupon codes.

    1 reply 9 ಮರುಟ್ವೀಟಿಸುವಿಕೆಗಳು 26ಇಷ್ಟಗಳು
    ರದ್ದುಗೊಳಿಸು
  12. ಡಿಸೆಂ 12

    Don’t let the catchy title fool you. This post is substantial and insightful. ML is easier and more enjoyable to learn when it’s tied to projects.

    You’re Doing it Wrong. Why Machine Learning Does Not Have to Be So Hard
    2 ಮರುಟ್ವೀಟಿಸುವಿಕೆಗಳು 14ಇಷ್ಟಗಳು
    ರದ್ದುಗೊಳಿಸು
  13. ಡಿಸೆಂ 12

    Straight talk from a pillar of the deep learning community.

    I wouldn’t have gotten anywhere in deep learning if it weren’t for Jupyter Notebooks. There are those super hackers who are naturally good at software engineering and they may not like coding with Jupyter, but notebooks are great for amateurs like me who don’t code for a living.
    3 ಮರುಟ್ವೀಟಿಸುವಿಕೆಗಳು 20ಇಷ್ಟಗಳು
    ರದ್ದುಗೊಳಿಸು
  14. ಡಿಸೆಂ 11

    I finally released my polynomial regression course into the world. It's the third in my End to End Machine Learning series. I'm relieved and nervous and proud. I hope you enjoy it.

    1 reply 4 ಮರುಟ್ವೀಟಿಸುವಿಕೆಗಳು 38ಇಷ್ಟಗಳು
    ರದ್ದುಗೊಳಿಸು
  15. ಡಿಸೆಂ 10

    It was a tenet of my (Prof Neville Hogan’s) grad robotics lab that it was easier to design robots around the limitations of control algorithms than to alter physics through aggressive control.

    Current robots are designed to work well with high precision control systems. Sometimes I wonder if it'll be easier to just learn better robot designs that work better with existing deep rl algorithms, compared to trying to make deep rl algorithms work better on existing robots.
    ಈ ಥ್ರೆಡ್ ತೋರಿಸಿ
    1 ಮರುಟ್ವೀಟಿಸುವಿಕೆ 5ಇಷ್ಟಗಳು
    ರದ್ದುಗೊಳಿಸು
  16. ಡಿಸೆಂ 10

    Spoiler: Bayesian analysis reveals an “Occam factor” favoring simpler models over memorization.

    9ಇಷ್ಟಗಳು
    ಈ ಥ್ರೆಡ್ ತೋರಿಸಿ
    ಈ ಥ್ರೆಡ್ ತೋರಿಸಿ
    ರದ್ದುಗೊಳಿಸು
  17. ಡಿಸೆಂ 10

    A longstanding mystery of neural networks is that they are so flexible that they can memorize pure noise, but instead they extract repeatable patterns. No one knew why. Until now.

    A Bayesian Perspective on Generalization and Stochastic Gradient Descent – Google AI
    3 replies 17 ಮರುಟ್ವೀಟಿಸುವಿಕೆಗಳು 58ಇಷ್ಟಗಳು
    ಈ ಥ್ರೆಡ್ ತೋರಿಸಿ
    ಈ ಥ್ರೆಡ್ ತೋರಿಸಿ
    ರದ್ದುಗೊಳಿಸು
  18. ಅವರು ಮರುಟ್ವೀಟಿಸಿದ್ದಾರೆ
    ಡಿಸೆಂ 9

    This video by has good tips about using multiple resources to learn concepts that are new to you, but the thing that caught my attention is a good intuitive explanation of what is meant by the (not-helpfully-named) "kernel trick".

    1 reply 10 ಮರುಟ್ವೀಟಿಸುವಿಕೆಗಳು 32ಇಷ್ಟಗಳು
    ರದ್ದುಗೊಳಿಸು
  19. ಡಿಸೆಂ 8

    Albon wisdom. Seriously. Not sarcastically. Thread.

    More and more I think the best grand strategy for real success is focus. Cut out people who try to tear you down and media that distracts you from what matters.
    ಈ ಥ್ರೆಡ್ ತೋರಿಸಿ
    6ಇಷ್ಟಗಳು
    ರದ್ದುಗೊಳಿಸು
  20. ಡಿಸೆಂ 8

    Watching this right now! It's a pleasure to hear from one of the founders of the R community.

    In one hour (at 4:30pm ET) I’ll look back at 10 years of history of the R community on the 10th anniversary of the Revolutions blog at DC. Livestream here:
    5ಇಷ್ಟಗಳು
    ರದ್ದುಗೊಳಿಸು
  21. ಡಿಸೆಂ 8

    I just solved television. Every twenty-two minute show could be shortened to five if two characters were honest with each other in the opening scene.

    2 replies 7 ಮರುಟ್ವೀಟಿಸುವಿಕೆಗಳು 38ಇಷ್ಟಗಳು
    ರದ್ದುಗೊಳಿಸು

@_brohrer_ ಅವರು ಇನ್ನೂ ಟ್ವೀಟ್ ಮಾಡಿಲ್ಲ.

ಲೋಡಿಂಗ್ ಸಮಯ ಸ್ವಲ್ಪ ತೆಗೆದುಕೊಳ್ಳುತ್ತಿರುವಂತೆನಿಸುತ್ತದೆ.

Twitter ಸಾಮರ್ಥ್ಯ ಮೀರಿರಬಹುದು ಅಥವಾ ಕ್ಷಣಿಕವಾದ ತೊಂದರೆಯನ್ನು ಅನುಭವಿಸುತ್ತಿರಬಹುದು. ಮತ್ತೆ ಪ್ರಯತ್ನಿಸಿ ಅಥವಾ ಹೆಚ್ಚಿನ ಮಾಹಿತಿಗೆ Twitter ಸ್ಥಿತಿಗೆ ಭೇಟಿ ನೀಡಿ.

    ಇದನ್ನೂ ಸಹ ನೀವು ಇಷ್ಟಪಡಬಹುದು

    ·