AutoGluon: AutoML for Image, Text, Time Series, and Tabular Data

AutoML for Image, Text, Time Series, and Tabular Data.AutoGluon automates machine learning tasks enabling you to easily achieve strong predictive performance in your applications. With just a few lines of code, you can train and deploy high-accuracy machine learning and deep learning models on image, text, time series, and tabular data.

Transcript

2. Share your knowledge. It is a way to achieve immortality. ~ Dalai Lama XIV~ Learning is not attained by chance, it must be sought for with ardor and attended to with diligence ~ Abigail Adams~
3. Disclaimer* ★ The views expressed are solely those of the speaker and do not reflect the views of their employer or affiliated organizations. ★ Information presented is for educational purposes only and should not be taken as professional advice. ★ No endorsement of any products, services, or organizations is intended by their mention. ★ Content is accurate to the best of the speaker's knowledge as of the presentation date; future changes are not accounted for. ★ Content is recorded , deck for sharing after consent for speakers
4. AutoGluon - ML Get started with training your prediction models today!
5. Speaker : Paul Antony Cyber Security Engineer @ Amazon 12+ Years Experience Hobby: Badminton, Driving, Walking, FPV LinkedIn : Paul Antony @LinkedIn Speaker Photo
6. Who is this meant for? - Python - ML
7. Agenda - HighLevel - ML and AI - When to use ML for projects? - Modalities - General steps - Simplified - AutoGluon - Demo - Bonus - Q&A
8. Artificial Intelligence vs Machine Learning AI ML
9. Types of ML 1) Supervised a) Regression b) Classification 2) Unsupervised a) Clusters
10. When to use ML for projects? - Cannot code the rules - Cannot scale - Needs to change with data
11. Steps - Preparing data - Separate dataset into: - Train data - Test data - Train and create a model using Train data - Test the model with test data - Adjust model or Original data set - Repeat
12. Modalities - We mentioned data earlier for training. This data can be: - Text - Images - Time series based - Tabular
13. AutoGluon - Simplified - Different Modalities - Different models
14. Demo
15. Summary for code #1 load D_tr = TabularDataset(data=”data.csv”) D_te = TabularDataset(data=”test.csv”) #2 create model Predictor = TabularPredictor(label=”to_predict”, eval_metric=”mean_squared_error”).fit(tra in_data=D_tr) #3 predict Predict_out = Predictor.predict(D_te)
16. Whats next? - Visit link to download demo code discussed - You can go look for datasets on web at: - Kaggle - Huggingface - Train your models by replacing train and test data I also have a blog version of this presentation here https://www.thesynonymsofopinion.com/post/machine-learning-for-newbies
17. Sources - Autogluon github - Kaggle - Coursera - Machine learning University
18. Q & A
19. I want to be a future speaker Please reach out with your topic @ pinnaclepoint2024@gmail.com @ http://tinyurl.com/techenthus2024 (WhatsApp) Social Media Handles @ Youtube : https://www.youtube.com/@NextGenTechEnthusiasts @ Twitter : @MeetTechEnthus @ Code demos: https://github.com/nextgen-tech-enthusiasts
21. Content Sharing Consent: Speaker (Fill in) May we share this deck with the Tech Enthusiast Group Yes ( Paul Antony) This is live Streamed on Youtube.Permission to use this video and other social media handles to promote Tech Enthusiast and Sharing Yes ( Paul Antony)