Paper In A Nutshell

Debayan Bhattacharya

0.0(0개의 평가)
과학 기술

If you are in a hurry and you want to know the core ideas of a research paper in Deep Learning/Artificial Intelligence, this podcast is for you. In each episode, I take a scientific paper and explain in simple words its main ideas. Hope you enjoy it!

2021. 04. 04.

Capsule Networks

After a long hiatus, I am back with a new episode. In today’s episode, I talk about Capsule Networks. What excites me the most about this network is that it takes the concepts of linear algebra and applies it to visual recognition. Vectors, a concept which any high school student is well aware of, is used to capture spatial relationships of objects with respect to one another. It attempts to redress a fundamental flaw of Convolutional Neural Network in the process. This paper is a brain child of Geoffrey Hinton - the pioneer of Deep Learning. I always enjoy reading his papers as they are heavily inspired by biology. This paper is nothing short of amazing! I hope you enjoy this episode. Link to the paper: https://arxiv.org/pdf/1710.09829.pdf Link to Max Pechyonkin’s blog: https://pechyonkin.me/capsules-1/
2020. 12. 12.

Dropout: A Simple Way to Prevent Neural Networks from Overfitting

In today’s paper I talk about one of the most important papers in Deep Learning. Dropouts were introduced to make neural networks understand the data better. What’s fascinating is that the idea of dropping out units in a neural network was inspired from Darwanian theory of evolution. It is absolutely amazing when people derive ideas from the natural world and translate those ideas into different scientific disciplines. I had a really wonderful time reading the paper and understanding its core ideas. For this episode, I have changed the format a bit. You will understand what I mean by that if you have followed my episodes thus far. Paper: https://jmlr.org/papers/volume15/srivastava14a/srivastava14a.pdf Please follow me on Spotify if you like this podcast. Suggest to your friends and colleagues if you think I am worth their time. That would mean a lot to me. Also, I would really like it if you gave me feedback or suggestions. If not, write me a mail nonetheless! I would love to get to know you. Email: paperinanutshell@gmail.com If you like my episodes, please follow me on Facebook, Instagram and Twitter. Follow me on Spotify and Apple Pocasts. Facebook: https://www.facebook.com/Paper-In-A-Nutshell-101609351791726 Instagram: https://www.instagram.com/paperinanutshell/ Twitter: https://twitter.com/NutshellPaper Streaming links: https://anchor.fm/debayan-bhattacharya
2020. 11. 28.

U-Nets

In today’s paper, we discuss U Nets-the revolutionary neural network that took research in Deep Learning based Medical Imaging Analysis to a new level. We discuss what U net is , what it is used for and why it works specifically in Medical Imaging Analysis. Paper discussed: https://arxiv.org/pdf/1505.04597.pdf Additional Reading Material: A Survey on Deep Learning in Medical Image Analysis : https://arxiv.org/pdf/1702.05747.pdf What is Pixel, Warping and Rand error: https://ashm8206.github.io/2018/04/08/Segmentation-Metrics.html Deep Neural Networks Segment Neuronal Membranes in Electron Microscopy Images: https://papers.nips.cc/paper/2012/file/459a4ddcb586f24efd9395aa7662bc7c-Paper.pdf If you like my episodes, please follow me on Facebook, Instagram and Twitter. Follow me on Spotify and Apple Pocasts. Facebook: https://www.facebook.com/Paper-In-A-Nutshell-101609351791726 Instagram: https://www.instagram.com/paperinanutshell/ Twitter: https://twitter.com/NutshellPaper Streaming links: https://anchor.fm/debayan-bhattacharya
2020. 11. 21.

YOLO: You Only Look Once

In this episode, we discuss what is YOLO and why it is so good! We break down the paper into its core ideas. I hope you enjoy it. Link to the papers: V1: https://arxiv.org/pdf/1506.02640.pdf V2: https://arxiv.org/pdf/1612.08242.pdf V3: https://pjreddie.com/media/files/papers/YOLOv3.pdf V4: https://arxiv.org/pdf/2004.10934.pdf My favorite YOLO implementations: Tensorflow: https://github.com/YunYang1994/tensorflow-yolov3 Pytorch: https://github.com/eriklindernoren/PyTorch-YOLOv3 If you enjoy this episode, head over to my instagram channel and follow me. https://www.instagram.com/paperinanutshell/ Thank you! Dhonnobaad!
2020. 11. 20.

Pilot

Hello everyone! Welcome to the pilot episode of my podcast “Paper In A Nutshell”. My name is Debayan Bhattacharya, a master student at Techniche Üniversität Hamburg. I have more than 2 years of programming experience and another 2 years of experience in the field of deep learning. The core of my research is Deep Learning and Artificial Intelligence for Computer Vision Applications. I have read a lot of research papers in Deep Learning ranging from Convolutional Networks to Natural Language Processing. One thing which I have always wanted is an explanation of difficult papers in simple terms. This podcast is especially for that. I will take one paper and break it down to its core ideas.I believe that unless you can explain complex ideas in clean and simple terms, you have not clearly understood the idea in the first place. Furthermore, extrapolation of simple ideas to real world scenarios leads to breakthroughs in technology. I hope through my podcasts I can do some “transfer learning” from my brain to yours. Apart from that, I am genuinely fascinated by this amazing field of Deep Learning. Artificial Intelligence is all around us. It is changing the very way we interact with machines. We are surrounded by countless machines trying to recommend to us what food we want to eat, what movie we want to watch, what should we buy next and so on. Therefore, it is imperative that we understand the algorithms that power these predictions. I believe that we can do tremendous good to the world if we consciously build technology that helps mankind. Thank you all! I will see you in the first episode!

5개의 에피소드

제작진

Debayan Bhattacharya
방송 연도

2020년 - 2021년
에피소드

5
등급

전체 연령 사용가
웹사이트 보기

Paper In A Nutshell

Paper In A Nutshell

에피소드

Capsule Networks

Dropout: A Simple Way to Prevent Neural Networks from Overfitting

U-Nets

YOLO: You Only Look Once

Pilot

소개

정보