Everyone knows how important hand-labeled data is for machine learning and also how tedious it can be to collect large labeled data. This usually involves a lot of time and expertise. We actually know it because at idealo we have a lot of supervised machine learning projects where good labeled data is key to our models. So having any solution that can generate good labeled data quickly is beneficial.
Google AI now collaborated with Stanford and Brown University to explore a potential solution to this problem where they harness organizational knowledge to quickly label large training datasets. Their framework is called Snorkel DryBell which is based on the open-source Snorkel (Link: https://lnkd.in/dbspu-z) system. The idea is pretty simple instead of having humans labeling your data, Snorkel DryBell enables writing labeling functions that label training data programmatically. This way you can generated labeled data much faster but obviously it's also of much lower quality. It's a very interesting read! Check it out! #deeplearning #machinelearning
Article: https://lnkd.in/ddsKTsF
Paper: https://lnkd.in/d-cRdSz
✴️ @AI_Python_EN
Google AI now collaborated with Stanford and Brown University to explore a potential solution to this problem where they harness organizational knowledge to quickly label large training datasets. Their framework is called Snorkel DryBell which is based on the open-source Snorkel (Link: https://lnkd.in/dbspu-z) system. The idea is pretty simple instead of having humans labeling your data, Snorkel DryBell enables writing labeling functions that label training data programmatically. This way you can generated labeled data much faster but obviously it's also of much lower quality. It's a very interesting read! Check it out! #deeplearning #machinelearning
Article: https://lnkd.in/ddsKTsF
Paper: https://lnkd.in/d-cRdSz
✴️ @AI_Python_EN
Google AI Blog
Harnessing Organizational Knowledge for Machine Learning
Posted by Alex Ratner, Stanford University and Cassandra Xia, Google AI One of the biggest bottlenecks in developing machine learning (M...
The Matrix Calculus You Need For Deep Learning
By Terence Parr and Jeremy Howard : https://lnkd.in/dC5MqZM
#100DaysOfMLCode #ArtificialIntelligence #BigData #DeepLearning #MachineLearning #NeuralNetworks
✴️ @AI_Python_EN
By Terence Parr and Jeremy Howard : https://lnkd.in/dC5MqZM
#100DaysOfMLCode #ArtificialIntelligence #BigData #DeepLearning #MachineLearning #NeuralNetworks
✴️ @AI_Python_EN
Here's a list of ALL the #machinelearning and #deeplearning articles we have published in March so far. This is for both beginners and advanced #datascience enthusiasts!
• 11 Steps to Transition into #DataScience (for Reporting / MIS / BI Professionals) - https://buff.ly/2EPTCqB
• 5 Amazing Deep Learning Frameworks Every Data Scientist Must Know! (with Illustrated Infographic) - https://buff.ly/2TS3SXQ
• Top 5 Data Science GitHub Repositories and Reddit Discussions (February 2019) - https://buff.ly/2GYFvBW
• Hands-On Introduction to creditR: An Amazing R Package to Enhance Credit Risk Scoring and Validation - https://buff.ly/2CiGWa3
• A Step-by-Step #NLP Guide to Learn ELMo for Extracting Features from Text - https://buff.ly/2HCQjFo
• The Path to Artificial General Intelligence with Professor Melanie Mitchell - https://buff.ly/2HmDrnj
✴️ @AI_Python_EN
• 11 Steps to Transition into #DataScience (for Reporting / MIS / BI Professionals) - https://buff.ly/2EPTCqB
• 5 Amazing Deep Learning Frameworks Every Data Scientist Must Know! (with Illustrated Infographic) - https://buff.ly/2TS3SXQ
• Top 5 Data Science GitHub Repositories and Reddit Discussions (February 2019) - https://buff.ly/2GYFvBW
• Hands-On Introduction to creditR: An Amazing R Package to Enhance Credit Risk Scoring and Validation - https://buff.ly/2CiGWa3
• A Step-by-Step #NLP Guide to Learn ELMo for Extracting Features from Text - https://buff.ly/2HCQjFo
• The Path to Artificial General Intelligence with Professor Melanie Mitchell - https://buff.ly/2HmDrnj
✴️ @AI_Python_EN
⚠️ The most cliché things that some data scientists actually say and do
On deep learning:
✔️ Say: «Deep learning is overkill in most cases.»
❌ Do: Use deep learning for everything.
On Excel:
✔️ Say: «Excel is such a useless and outdated tool.»
❌ Do: Use Excel every day.
On statistics:
✔️ Say: «My biggest strength is math and statistics.»
❌ Do: Google the definition of standard deviation.
On correlation:
✔️ Say: «Correlation is not causation.»
❌ Do: Base feature selection entirely on correlation.
On big data
✔️ Say: «We should use a big data store for this.»
❌ Do: Put everything in an SQL database.
On careers
✔️ Say: «I truly believe in the mission of this company. We’re going to change the world.»
❌ Do: Change jobs every year or whenever someone offers slightly more money.
✔️ Say: «I’m a senior data scientist and machine learning expert.»
❌ Do: Still, haven’t shipped a model to production.
On science
✔️ Say: «We use the scientific method. Every hypothesis needs to be tested.»
❌ Do: Deploy a model straight to production because it converged on the training set.
On academic papers
✔️ Say: «I read a lot of papers.»
❌ Do: Read the abstract of a DeepMind paper once.
On p-values
✔️ Say: «The p-value is very often misunderstood.»
❌ Do: Offer a flawed explanation of p-values.
❇️ @AI_Python
🗣 @AI_Python_arXiv
✴️ @AI_Python_EN
On deep learning:
✔️ Say: «Deep learning is overkill in most cases.»
❌ Do: Use deep learning for everything.
On Excel:
✔️ Say: «Excel is such a useless and outdated tool.»
❌ Do: Use Excel every day.
On statistics:
✔️ Say: «My biggest strength is math and statistics.»
❌ Do: Google the definition of standard deviation.
On correlation:
✔️ Say: «Correlation is not causation.»
❌ Do: Base feature selection entirely on correlation.
On big data
✔️ Say: «We should use a big data store for this.»
❌ Do: Put everything in an SQL database.
On careers
✔️ Say: «I truly believe in the mission of this company. We’re going to change the world.»
❌ Do: Change jobs every year or whenever someone offers slightly more money.
✔️ Say: «I’m a senior data scientist and machine learning expert.»
❌ Do: Still, haven’t shipped a model to production.
On science
✔️ Say: «We use the scientific method. Every hypothesis needs to be tested.»
❌ Do: Deploy a model straight to production because it converged on the training set.
On academic papers
✔️ Say: «I read a lot of papers.»
❌ Do: Read the abstract of a DeepMind paper once.
On p-values
✔️ Say: «The p-value is very often misunderstood.»
❌ Do: Offer a flawed explanation of p-values.
❇️ @AI_Python
🗣 @AI_Python_arXiv
✴️ @AI_Python_EN
New documentation about how differentiable programming works in Swift: • Differentiable Functions and Differentiation APIs: https://github.com/tensorflow/swift/blob/master/docs/DifferentiableFunctions.md … • Differentiable Types: https://github.com/tensorflow/swift/blob/master/docs/DifferentiableTypes.md … With language integration, autodiff is just a compiler implementation detail.
✴️ @AI_Python_EN
✴️ @AI_Python_EN
Machine Learning with Python, Jupyter, KSQL and TensorFlow https://ift.tt/2FbgQq6 #python #tensorflow #jupyter #ksql https://ift.tt/2TFCqNC
✴️ @AI_Python_EN
✴️ @AI_Python_EN
Discover Computer Vision Datasets with this search engine
#dataset #image #visual #search #engine #vision
https://www.visualdata.io
✴️ @AI_Python_EN
#dataset #image #visual #search #engine #vision
https://www.visualdata.io
✴️ @AI_Python_EN
MIT DeepLearning Basics — Introduction and Overview with TensorFlow #robotics #game #games
bit.ly/2E4xnx6
✴️ @AI_Python_EN
bit.ly/2E4xnx6
✴️ @AI_Python_EN
Deep Learning Project Building with Python and Keras ☞ http://bit.ly/2HADriH #DeepLearning #ai
✴️ @AI_Python_EN
✴️ @AI_Python_EN
A Step-by-Step Guide to Machine Learning Problem Framing: Diving into Machine Learning (ML) without knowing what you’re trying to achieve is a recipe for disaster. #MachineLearning #DeepLearning #DataScience
https://medium.com/thelaunchpad/a-step-by-step-guide-to-machine-learning-problem-framing-6fc17126b981
✴️ @AI_Python_EN
https://medium.com/thelaunchpad/a-step-by-step-guide-to-machine-learning-problem-framing-6fc17126b981
✴️ @AI_Python_EN
Rust: Programming Language Cheat Sheet.
#BigData #Analytics #DataScience #AI #MachineLearning #IoT #IIoT #PyTorch #Python #RStats #Rust #TensorFlow #Java #JavaScript #ReactJS #GoLang #CloudComputing #Serverless #Linux #Programming
http://bit.ly/2HEVJzl
✴️ @AI_Python_EN
#BigData #Analytics #DataScience #AI #MachineLearning #IoT #IIoT #PyTorch #Python #RStats #Rust #TensorFlow #Java #JavaScript #ReactJS #GoLang #CloudComputing #Serverless #Linux #Programming
http://bit.ly/2HEVJzl
✴️ @AI_Python_EN
It was interesting to work on classifying duplicate questions on quora. Just uploaded the code to git.
Approach 1: Siamese Network with manhattan distance as the objective function.
Code: https://lnkd.in/fhzV_HU
Approach 2: XGBoost + TF-iDF + NLP feature engineering.
Code: https://lnkd.in/f3zqm37
Competition: https://lnkd.in/fYHtwJq
✴️ @AI_Python_EN
Approach 1: Siamese Network with manhattan distance as the objective function.
Code: https://lnkd.in/fhzV_HU
Approach 2: XGBoost + TF-iDF + NLP feature engineering.
Code: https://lnkd.in/f3zqm37
Competition: https://lnkd.in/fYHtwJq
✴️ @AI_Python_EN
Deep Learning Drizzle
"Read enough so you start developing intuitions and then trust your intuitions and go for it!" - Geoffrey Hinton
By Marimuthu K.: https://lnkd.in/e6BBDVJ
#artificialintelligence #deeplearning #machinelearning
✴️ @AI_Python_EN
"Read enough so you start developing intuitions and then trust your intuitions and go for it!" - Geoffrey Hinton
By Marimuthu K.: https://lnkd.in/e6BBDVJ
#artificialintelligence #deeplearning #machinelearning
✴️ @AI_Python_EN
Curated list of awesome ****DEEP LEARNING**** tutorials, projects and communities.
Github Link - https://lnkd.in/fJdpFMn
#deeplearning #machinelearning #datascience
✴️ @AI_Python_EN
Github Link - https://lnkd.in/fJdpFMn
#deeplearning #machinelearning #datascience
✴️ @AI_Python_EN
This guide gives a complete understanding about various #machinelearning algorithms along with R & Python #codes to run them. These #algorithms can be applied to any data problem:
Linear Regression,
Logistic Regression,
Decision Tree,
SVM,
Naive Bayes,
kNN,
K-Means,
#Random Forest.
If you are keen to master machine learning, start right away.
Link : bit.ly/2CpWIjH
#machinelearning #deeplearning #python #coding #linkedin #decisiontrees #logisticregression #linearregression #forest #analytics #randomization #computervision
✴️ @AI_Python_EN
Linear Regression,
Logistic Regression,
Decision Tree,
SVM,
Naive Bayes,
kNN,
K-Means,
#Random Forest.
If you are keen to master machine learning, start right away.
Link : bit.ly/2CpWIjH
#machinelearning #deeplearning #python #coding #linkedin #decisiontrees #logisticregression #linearregression #forest #analytics #randomization #computervision
✴️ @AI_Python_EN
Forwarded from DLeX: AI Python (Meysam Asgari)
Have you heard of SuperTML?
Two-Dimensional Word Embedding and Transfer Learning Using ImageNet Pretrained CNN Models for the Classifications on Tabular Data
SuperTML: Two-Dimensional Word Embedding and Transfer Learning Using ImageNet Pretrained CNN Models for the Classifications on Tabular Data
Tabular data is the most commonly used form of data in industry. Gradient Boosting Trees, Support Vector Machine, Random Forest, and Logistic Regression are typically used for classification tasks on tabular data.
DNN models using categorical embeddings are also applied in this task, but all attempts thus far have used one-dimensional embeddings. The recent work of Super Characters method using two-dimensional word embeddings achieved the state of art result in text classification tasks, showcasing the promise of this new approach.
The SuperTML method, which borrows the idea of Super Characters method and two-dimensional embeddings to address the problem of classification on tabular data. It has achieved state-of-the-art results on both large and small datasets.
Here’s the paper: https://lnkd.in/djGFf63
❇️ @AI_Python
🗣 @AI_Python_arXiv
✴️ @AI_Python_EN
Two-Dimensional Word Embedding and Transfer Learning Using ImageNet Pretrained CNN Models for the Classifications on Tabular Data
SuperTML: Two-Dimensional Word Embedding and Transfer Learning Using ImageNet Pretrained CNN Models for the Classifications on Tabular Data
Tabular data is the most commonly used form of data in industry. Gradient Boosting Trees, Support Vector Machine, Random Forest, and Logistic Regression are typically used for classification tasks on tabular data.
DNN models using categorical embeddings are also applied in this task, but all attempts thus far have used one-dimensional embeddings. The recent work of Super Characters method using two-dimensional word embeddings achieved the state of art result in text classification tasks, showcasing the promise of this new approach.
The SuperTML method, which borrows the idea of Super Characters method and two-dimensional embeddings to address the problem of classification on tabular data. It has achieved state-of-the-art results on both large and small datasets.
Here’s the paper: https://lnkd.in/djGFf63
❇️ @AI_Python
🗣 @AI_Python_arXiv
✴️ @AI_Python_EN
Data-Driven Careers Dechipered
Seems Different with what I believed, but still worth it to discuss
#business #technology #datascience
❇️ @AI_Python
🗣 @AI_Python_arXiv
✴️ @AI_Python_EN
Seems Different with what I believed, but still worth it to discuss
#business #technology #datascience
❇️ @AI_Python
🗣 @AI_Python_arXiv
✴️ @AI_Python_EN
Yann LeCun Tweeted:
I gave two talks in Harvard's Mind, Brain, & Behavior Distinguished Lecture Series last week.
The slides are here:
- 2019-03-13 "The Power and Limits of Deep Learning":...
https://t.co/Pp4flTGlZ8
✴️ @AI_Python_EN
I gave two talks in Harvard's Mind, Brain, & Behavior Distinguished Lecture Series last week.
The slides are here:
- 2019-03-13 "The Power and Limits of Deep Learning":...
https://t.co/Pp4flTGlZ8
✴️ @AI_Python_EN
Our Computer Vision and Deep Learning Group:
https://t.me/joinchat/ECtp7VVFvEwjIrMrdgI-2w
🗣 @AI_Python_Arxiv
✴️ @AI_Python_EN
❇️ @AI_Python
https://t.me/joinchat/ECtp7VVFvEwjIrMrdgI-2w
🗣 @AI_Python_Arxiv
✴️ @AI_Python_EN
❇️ @AI_Python