Anthony Goldbloom - How to Win Kaggle Competitions
Ғылым және технология
Anthony Goldbloom is the founder and CEO of Kaggle. In 2011 & 2012, Forbes Magazine named Anthony as one of the 30 under 30 in technology. In 2011, Fast Company featured him as one of the innovative thinkers who are changing the future of business.
He and Lukas discuss the differences in strategies that do well in Kaggle competitions vs academia vs in production. They discuss his 2016 Ted talk through the lens of 2020, frameworks, and languages.
Topics Discussed:
0:00 Sneak Peek
0:20 Introduction
0:45 methods used in kaggle competitions vs mainstream academia
2:30 What are people doing to win competitions/feature engineering
3:55 Kaggle Competitions now vs 10 years ago
8:35 Data augmentation strategies
10:06 Overfitting in Kaggle Competitions
12:53 How to not overfit
14:11 Kaggle competitions vs the real world
18:15 Getting into ML through Kaggle
22:03 Other Kaggle products
25:48 Favorite under appreciated kernel or dataset
28:27 Python & R
32:03 Frameworks
35:15 2016 Ted talk though the 2020 frame
37:54 Reinforcement Learning
38:43 What’s the topic in ML that people don’t talk about enough?
42:02 Where are the biggest bottlenecks in deploying ML software?
Check out Kaggle: www.kaggle.com/
Follow Anthony on Twitter: / antgoldbloom
Watch his 2016 Ted Talk: www.ted.com/talks/anthony_gol...
Visit our podcasts homepage for transcripts and more episodes!
www.wandb.com/podcast
🎙 Get our podcasts on these platforms:
Soundcloud: wandb.me/soundcloud
Apple Podcasts: wandb.me/apple-podcasts
Spotify: wandb.me/spotify
Google: wandb.me/gd_google
KZread: wandb.me/youtube
We started Weights and Biases to build tools for Machine Learning practitioners because we care a lot about the impact that Machine Learning can have in the world and we love working in the trenches with the people building these models. One of the most fun things about these building tools has been the conversations with these ML practitioners and learning about the interesting things they’re working on. This process has been so fun that we wanted to open it up to the world in the form of our new podcast called Gradient Dissent. We hope you have as much fun listening to it as we had making it!
Weights and Biases:
We’re always free for academics and open source projects. Email carey@wandb.com with any questions or feature suggestions.
* Blog: www.wandb.com/articles
* Gallery: See what you can create with W&B - app.wandb.ai/gallery
* Join our community of ML practitioners working on interesting problems - www.wandb.com/ml-community
Host: Lukas Biewald - / l2k
Producer: Lavanya Shukla - / lavanyaai
Editor: Cayla Sharp - caylasharp.com/
Пікірлер: 53
Each of your questions was so meaningful and pertinent. Thank you so much for such an incredibly insightful video.
Thank you very much for this amazing podcast. Helped me and gave me better ideas on how to proceed ❤️
awesome, and thanks for the timestamps! really helpful
Thank you. This is very useful to people who are beginners in ML filed. Keep doing such brilliant insightful content
Thank you for such an insightful interview! Your questions perfectly mirrored my own curiosities, and it's the first video I've watched till the end. It provided a wealth of knowledge and left me feeling incredibly satisfied.
really great insights, thanks for creating and sharing this.
This was really insightful! Thanks for having this interview!
@WeightsBiases
3 жыл бұрын
Glad you enjoyed it!
Great interview Lukas, I didn't think I'd get good advices in winning Kaggle from someone at that top-level like the CEO but turns out he is well versed in the business and was generous with the information provided, thank you for prompting and leading the conversation with questions that really matter.
Thank you, really insightful content
Is Anthony practising dance moves while answering the questions? :)
@v.r.santhosh7201
3 жыл бұрын
i think he is walking on treadmill
@patrickadonis7930
2 жыл бұрын
sorry to be offtopic but does any of you know a trick to get back into an instagram account? I was stupid lost my account password. I love any tricks you can offer me.
@davianjabari6697
2 жыл бұрын
@Patrick Adonis instablaster ;)
@patrickadonis7930
2 жыл бұрын
@Davian Jabari I really appreciate your reply. I got to the site through google and im in the hacking process atm. I see it takes a while so I will get back to you later with my results.
@patrickadonis7930
2 жыл бұрын
@Davian Jabari it did the trick and I now got access to my account again. I'm so happy! Thank you so much, you saved my account !
It´s working great. Thank you so much for sharing the information
love the work guys! keep it up fella xxx
Phenomenal interview!
Really love who you bring on to these interviews. Very insightful questions as well. Great channel! Keep it coming!
@WeightsBiases
3 жыл бұрын
Thank you! Let us know if you have any ideas or requests on who we could talk to :)
That wire made me think he was plugged in! Amazing talk though, WandB if you see this comment please make a video series or something on the clever techniques kind of case studies. That data augmentation trick was really cool...
@WeightsBiases
3 жыл бұрын
Lol, thank you! We'll take note, we're designing another set of tutorials now and will take this into consideration
Loved listening to this talk! I am about to finish my masters soon and am excited to start my professional journey in the field of ML. I often hear people say that kaggling is quite different from working in the industry because the datasets you see on kaggle have already been cleaned. Any advice for a noob who's looking to bridge this gap and also gain experience in data cleaning and wrangling?
@WeightsBiases
Жыл бұрын
Thank you! You may find this Kaggle course on data cleaning useful www.kaggle.com/learn/data-cleaning
I was just yesterday entering a new Kaggle competition and was looking for some more insights into the field. I don't know how you did it, but the algorithm on choosing your guests works!
@WeightsBiases
3 жыл бұрын
Glad we were able to help, happy Kaggle-ing!
Can anyone hold & fix the position of the guy on the right
His (Anthony's) dance moves are hypnotising my mind and I keep on anticipating his next move...yet the discussion is particularly fertile!
@JohnSmith-bb1sv
Жыл бұрын
He's walking on a treadmill.
Really nice questions and answers, this was gold! Btw..why is he walking? lol
@WeightsBiases
3 жыл бұрын
He was on a treadmill, thank you!
Come on people, he's on his treadmill.
Around 13:20 he mentions a Google researcher that I didn't catch. Does anyone know the name of the researcher he mentions? Regarding data sets that are too small to CV.
@lbiewald
3 жыл бұрын
I believe it's Moritz Hardt
Thanks
thanks
Great video
@WeightsBiases
2 жыл бұрын
Thanks, glad you enjoyed it!
this cross-validates a road-map to the journey of AI challenge
Niceeeeeeee
He says cross validation can be used to prevent overfitting..?? Its used to check overfitting right...overfitting is reduced by limiting features or adding more data.
@lbiewald
3 жыл бұрын
This is kind of like meta-overfitting. In a kaggle competition people will typically try a ton of different techniques, so with a fixed test set it's possible to overfit hyperparameters to the specific test set, and then performance drops on kaggle's held out test set.
Is he cycling while the interview?
anthony going left-right-left right. maybe he's on a treadmill.
Dont move
Last years review of Kaggle winners.. in case you are interested.. kzread.info/dash/bejne/ooBtmKuJir2TaJs.html
is he walking on a treadmill ?
@WeightsBiases
2 жыл бұрын
Yep!
Why Anthony cant stop moving? Is he in a walking machine?
@colocolo49
Жыл бұрын
looked like he was on a slow-moving treadmill.