chuanenlin

[Article] 8 Simple Techniques to Prevent Overfitting (in Machine Learning)

Posted by chuanenlin on Sunday June 07, 2020 @01:05PM
from the self-post dept.

Hi!Overfitting is a very common problem in Machine Learning.I wrote a short article introducing '8 Simple Techniques to Prevent Overfitting' for beginners.Hope some of you may find it helpful!https://towardsdatascience.com/8-simple-techniques-to-prevent-overfitting-4d443da2ef7d

( Read More... | 1 comment | on HN )

Re:Ask HN: Best Functionality for Character Animation Tool? on Sunday October 27, 2019 @05:44PM (#21372630)

Any ideas welcomed!

Ask HN: Best Functionality for Character Animation Tool?

Posted by chuanenlin on Sunday October 27, 2019 @04:30PM
from the ask-hn dept.

Hi! I'm conducting research on enhancing the character animation process (more specifically, to complement and augment traditional hand-drawn animation). I’m interested if there’s any particular functionality you like about existing tools or if there are any exciting new functionalities you would love to see? Thank you!

( Read More... | 1 comment | 5 points | on HN )

Re:Motion Estimation with Optical Flow on Tuesday April 30, 2019 @03:31AM (#19785835)

@Darkphibre, wonderful background! Yes, and optical flow + recent advancements in deep learning is also certainly something exciting to see flourish! Action recognition seems like a promising research area (https://research.google.com/ava/ CVPR '18).

Re:Motion Estimation with Optical Flow on Tuesday April 30, 2019 @02:45AM (#19785662)

Hi @adzm, I'm the author of the article - thanks for reading! I think many of the implementations involve video processing techniques such as motion-compensated frame interpolation... this seems interesting (https://www.mia.uni-saarland.de/Publications/raket-isvc12.pd...).

Re:Motion Estimation with Optical Flow on Tuesday April 30, 2019 @02:35AM (#19785631)

Hi @haxiomic, I'm the author of the article - thanks for reading! Several more recent approaches of computing optical flow with CNNs such as FlowNet2 (CVPR '17) (https://github.com/lmb-freiburg/flownet2) should also be runnable on video or live webcam feed.

Re:Motion Estimation with Optical Flow on Tuesday April 30, 2019 @02:13AM (#19785530)

Hi @bitL, I'm the author of the article. Yes, I think deep learning is a promising solution that replaces the need for manual fine-tuning and is certainly driving momentum in optical flow research. Something you may want to look into is sequence-to-sequence (seq2seq) learning (https://google.github.io/seq2seq/).