OpenAI announced the final staged release of its 1.5 billion parameter language model GPT-2, along with all associated code and model weights

https://medium.com/syncedreview/openai-releases-1-5-billion-parameter-gpt-2-model-c34e97da56c0

❇️ @AI_Python_EN

NeurIPS 2019: Adversarial music that prevents Amazon Alexa from waking up

https://www.profillic.com/paper/arxiv:1911.00126

They target the attack on the wake-word detection system, jamming the model with some inconspicuous background music to deactivate the VAs while the audio adversary is present

❇️ @AI_Python_EN

The Reinforcement-Learning Methods that Allow AlphaStar to Outcompete Almost All Human Players at StarCraft II

https://bit.ly/2NjezOU

What are the three types of error in a #ML model?

👉 1. Bias - error caused by choosing an algorithm that cannot accurately model the signal in the data, i.e. the model is too general or was incorrectly selected. For example, selecting a simple linear regression to model highly non-linear data would result in error due to bias.

👉 2. Variance - error from an estimator being too specific and learning relationships that are specific to the training set but do not generalize to new samples well. Variance can come from fitting too closely to noise in the data, and models with high variance are extremely sensitive to changing inputs. Example: Creating a decision tree that splits the training set until every leaf node only contains 1 sample.

👉 3. Irreducible error - error caused by noise in the data that cannot be removed through modeling. Example: inaccuracy in data collection causes irreducible error.

❇️ @AI_Python_EN

François Chollet (Google, Creator of Keras) just released a paper on defining and measuring intelligence and a GitHub repo that includes a new #AI evaluation dataset, ARC – "Abstraction and Reasoning Corpus".

Paper: https://arxiv.org/abs/1911.01547
ARC: https://github.com/fchollet/ARC

#AI #machinelearning #deeplearning

❇️ @AI_Python_EN

Kick-start your Python Career with 56 amazing Python Open source Projects
#python #programming #technology #project

https://data-flair.training/blogs/python-open-source-projects/

❇️ @AI_Python_EN

Wherefore Multivariate Regression?

Multivariate analysis (MVA), in a regression setting, typically implies that a single dependent variable (outcome) is modeled as a function of two or more independent variables (predictors).

There are situations, though, in which we have two or more dependent variables we wish to model simultaneously, multivariate regression being one example. I tend to approach this through a structural equation modeling (SEM) framework but there are several alternatives.

Why not run one #regression for each outcome? There are several reasons, and the excerpt below from Multilevel Analysis: An Introduction to Basic and Advanced Multilevel Modeling (Snijders and Bosker) is a particularly succinct explanation in the context of multilevel models.

"Why analyze multiple dependent variables simultaneously? It is possible to analyze all m dependent variables separately. There are several reasons why it may be sensible to analyze the data jointly, that is, as multivariate data.

1. Conclusions can be drawn about the correlations between the dependent variables – notably, the extent to which the unexplained correlations depend on the individual and on the group level. Such conclusions follow from the partitioning of the covariances between the dependent variables over the levels of analysis.

2. The tests of specific effects for single dependent variables are more powerful in the multivariate analysis. This will be visible in the form of smaller standard errors. The additional power is negligible if the dependent variables are only weakly correlated, but may be considerable if the dependent variables are strongly correlated while at the same time the data are very incomplete, that is, the average number of measurements available per individual is considerably less than m.

3. Testing whether the effect of an explanatory variable on dependent variable Y1 is larger than its effect on Y2, when the data on Y1 and Y2 were observed (totally or partially) on the same individuals, is possible only by means of a multivariate analysis.

4. If one wishes to carry out a single test of the joint effect of an explanatory variable on several dependent variables, then a multivariate analysis is also required. Such a single test can be useful, for example, to avoid the danger of capitalization on chance which is inherent in carrying out a separate test for each dependent variable.

A multivariate analysis is more complicated than separate analyses for each dependent variable. Therefore, when one wishes to analyze several dependent variables, the greater complexity of the multivariate analysis will have to be balanced against the reasons listed above. Often it is advisable to start by analyzing the data for each dependent variable separately."

Source: Multilevel Analysis: An Introduction to Basic and Advanced Multilevel Modeling, Tom Snijders

❇️ @AI_Python_EN

How to deliver on Machine Learning projects

A guide to the ML Engineering Loop.

By Emmanuel Ameisen and Adam Coates:
https://blog.insightdatascience.com/how-to-deliver-on-machine-learning-projects-c8d82ce642b0

#ArtificialIntelligence #BigData #DataScience #DeepLearning #MachineLearning

❇️ @AI_Python_EN

Data Science Life Cycle


#artificialintellegence #machinelearning

❇️ @AI_Python_EN

Introduction to Autoencoders - Unsupervised Deep Learning Models (Cont'd) | Coursera

https://bit.ly/2Nw5CCh

❇️ @AI_Python_EN

Can We Learn the Language of Proteins?

#DataScience #MachineLearning #ArtificialIntelligence

http://bit.ly/33jCof6

❇️ @AI_Python_EN

What if you can generate a whole new Image just by giving its textual description?

Learn for Shibsankar Das in his hack session here:
http://bit.ly/DHS2019_66

He’ll be talking about “Generating Synthetic Images from Textual Description using GANs” in which he’ll implement GANs from scratch, formulate business use-cases

❇️ @AI_Python_EN

A neural network that transforms a design mock-up into a static website

https://github.com/emilwallner/Screenshot-to-code

#ArtificialIntelligence #DeepLearning #MachineLearning

❇️ @AI_Python_EN

Optimizing Millions of Hyperparameters by Implicit Differentiation Lorraine et al.:

https://arxiv.org/abs/1911.02590
#ArtificialIntelligence #MachineLearning

❇️ #AI_Python_EN

Story Realization: Expanding Plot Events into Sentences Ammanabrolu et al.:

https://arxiv.org/abs/1909.03480

#ArtificialIntelligence #DeepLearning #MachineLearning

❇️ @AI_Python_EN

News classification using classic Machine Learning tools (TF-IDF) and modern NLP approach based on transfer learning (ULMFIT) deployed on GCP
Github:
https://github.com/imadelh/NLP-news-classification

Blog:
https://imadelhanafi.com/posts/text_classification_ulmfit/

#DeepLearning #MachineLearning #NLP

❇️ @AI_Python_EN

Research Guide: Advanced Loss Functions for Machine Learning Models

http://bit.ly/36HBefu

#DataScience #MachineLearning #ArtificialIntelligence

❇️ @AI_Python_EN

Intro to optimization in deep learning: Momentum, RMSProp and Adam

https://bit.ly/2zwBLV0

❇️ @AI_Python_EN

A list of the biggest machine learning datasets from across the web

https://bit.ly/2TYGdVD

❇️ @AI_Python_EN

Self-training with Noisy Student improves ImageNet classification Xie et al.:

https://arxiv.org/abs/1911.04252

#ArtificialIntelligence #DeepLearning #MachineLearning

❇️ @AI_PythonEN

Productionizing #NLP Models

https://bit.ly/2OkdRAD

❇️ @AI_Python_EN