Evolutionary Powell's method is a discrete optimization algorithm I've found useful for hyperparameter tuning.

It makes weaker assumptions than Bayesian methods (and so is more robust), but stronger than random exploration (and so has better performance). It fills in the gap between then a bit.

Here's the full post on how Evolutionary Powell's method works:

We develop it as part of End-to-End Machine Learning Course 314:

The open source Ponderosa optimization package where it lives:

The line-by-line code walkthrough:

❇️ @AI_Python_EN

Usa Metabase


https://github.com/g-adamante/metabase_csv_export

❇️ @AI_Python_EN

Deep Speech, a good #Persian podcsts about #AI

We will talk about #ArtificialIntelligence, #MachineLearning and #DeepLearning news.

https://castbox.fm/channel/Deep-Speech-id2420707?country=us

❇️ @AI_Python_EN

Machine Learning Tutorial Suite - 90+ Free Tutorials

https://data-flair.training/blogs/machine-learning-tutorials-home/

NASA: Neural Articulated Shape Approximation.

Timothy Jeruzalski, Boyang Deng, Mohammad Norouzi, JP Lewis, Geoffrey Hinton, and Andrea Tagliasacchi

arxiv.org/abs/1912.03207

What is My Data Worth? – The Berkeley Artificial Intelligence

https://bair.berkeley.edu/blog/2019/12/16/data-worth/

The Akaike Information Criterion (AIC), Bayesian Information Criterion (BIC), and the Deviance Information Criterion (DIC) are perhaps the most widely-used information criteria (IC) in model building and selection. A fourth, Minimum Description Length (MDL), is closely related to the BIC. In a nutshell, they provide guidance as which alternative model provides the most "bang for buck," i.e., the best fit after penalizing for model complexity. Penalizing for complexity is important since, given candidate models of similar predictive or explanatory power, the simplest model is most likely to be the best choice. In line with Occam's razor, complex models sometimes perform poorly on data not used in the model building. There are several others, including AIC3, SABIC, and CAIC, and no clear consensus among authorities as far as I am aware as to which is "best" overall. IC will not necessarily agree on which model should be chosen. Cross-validation, Predicted Residual Error Sum of Squares (PRESS) statistic, a kind of cross-validation, and Mallows’ Cp are also used instead of IC. Information criteria are covered in varying levels in detail in most statistics textbooks and are the subject of numerous academic papers. I know of no single go-to source on this topic.

❇️ @AI_Python_EN

best to summarize the key trends I saw at NeurIPS 2019. The themes covered are:

1. Deconstructing the deep learning black box: many papers aim to understand the theories behind deep learning including convergence and generalization. It also analyzes exciting and contentious approach is neural tangent kernel.

2. New approaches: Bayesian deep learning & uncertainty estimation, graph neural networks, and convex optimization.
3. Neuroscience in machine learning: consciousness and attention And more.

https://huyenchip.com/2019/12/18/key-trends-neurips-2019.html

❇️ @AI_Python_EN

Best of arXiv.org for #AI, #MachineLearning, and #DeepLearning – November 2019

https://bit.ly/36OWsaD

❇️ @AI_Python_EN

Pandora/SiriusXM is currently looking for Summer 2020 interns on our Radio & Music Informatics team. We're looking for candidates with a strong background in recommender systems, machine learning, music generation, or natural language processing.

If you're looking to work on Audio Entertainment (including Music, Podcasts, News, etc) at scale we might have just what you're looking for. Our service spins music to over 100 million monthly active listeners and we have collected over 90 billion thumbs. On the content-based side we're working with millions of songs labeled by expert musicologists on up to 450 musical characteristics, and access to dozens of millions of music tracks from our catalog.

Selected candidates will be working with members of our research team, led by Òscar Celma. Interns are well compensated and our office is located in Oakland, California in the beautiful San Francisco Bay Area.

You can find more about our internship program and apply here:
http://www.pandora.com/careers/roadcrew
https://recruiting.adp.com/srccar/public/nghome.guid?c=1147611&d=ExternalCareerSite&prc=RMPOD3&r=5000541409506

Requirements:
- PhD candidate in CS/EE/Statistics/Math or other technical field

Pluses:
- Strong publication record in machine learning, recommender systems, natural language processing, and/or music information retrieval
- Ability to prototype and develop custom algorithms
- Knowledge of recommender systems
- Proficiency in Python/Scala/Java
- Experience in Spark/Hadoop/GCP

❇️ @AI_Python_EN

I am forming a new research group in Turkey and have multiple open positions for PhD students (you can keep living in Istanbul or in Ankara, Turkey and can work with me).

Students will receive TUBITAK’s scholarship to work on projects funded by TUBITAK and / or by EU. TUBITAK scholarship pays 4,500 TL per month which is equivalent to an amount around $800. Depending on the students’ performance, the amount that the students receive may increase over time. Also keep in mind that living costs for students in Turkey are cheaper than many big US cities and many other European cities.

I am looking for students with background in:

·      machine learning algorithm design and /or deep learning experience,

·      Python and C/C++ experience in programming,

·      Experience or interest in autonomous systems.

You should hold a M.Sc. (or an equivalent) degree and be excited to work on autonomous vehicles at various levels.

Where: Students can get into the PhD program at Istanbul Technical University at Istanbul Turkey or at Bilkent University at Ankara, Turkey. (You can work with me, if you get accepted into the PhD program at either university). Each institution has its own PhD acceptance requirement and interested students are encouraged to check those requirements.

Contact: If you are interested, then you are encouraged to send a CV and a brief statement listing your interests and relevant experiences. If you have published papers or a Github page, please include them in your email as well.

My web: http://cs.bilkent.edu.tr/~sedat/

Sedat Ozer, Ph.D.,
www.SedatOzer.com

❇️ @AI_Python_EN

Top 14 Machine Learning Research Papers Of 2019


https://bit.ly/2ExLslR

❇️ @AI_Python_EN

XGBoost: An Intuitive Explanation
Ashutosh Nayak :
https://towardsdatascience.com/xgboost-an-intuitive-explanation-88eb32a48eff
#MachineLearning #DataScience

MIT lecture series on deep learning in 2019
MIT lecture series on deep learning:Basics:
https://www.youtube.com/watch?v=O5xeyoRL95U&list=PLrAXtmErZgOeiKm4sgNOknGvNjby9efdf

MIT lecture series on deep learning: State of the Art:
https://www.youtube.com/watch?v=53YvP6gdD7U&list=PLrAXtmErZgOeiKm4sgNOknGvNjby9efdf

MIT lecture series on deep learning: Introduction to Deep RL:
https://www.youtube.com/watch?v=zR11FLZ-O9M&list=PLrAXtmErZgOeiKm4sgNOknGvNjby9efdf

Find The Most Updated and Free Artificial Intelligence, Machine Learning, Data Science, Deep Learning, Mathematics, Python Programming Resources

https://www.marktechpost.com/free-resources/

GL2vec: Graph Embedding Enriched by Line Graphs with Edge Features

Code: https://github.com/benedekrozemberczki/karateclub

Paper: https://link.springer.com/chapter/10.1007/978-3-030-36718-3_1

https://karateclub.readthedocs.io

Named Entity Recognition (NER) from social media posts is a challenging task. User-generated content which forms the nature of social media, is noisy and contains grammatical and linguistic errors. This noisy content makes it much harder for tasks such as named entity recognition. However some applications like automatic journalism or information retrieval from social media, require more information about entities mentioned in groups of social media posts. Conventional methods applied to structured and well typed documents provide acceptable results while compared to new user generated media, these methods are not satisfactory. One valuable piece of information about an entity is the related image to the text. Combining this multimodal data reduces ambiguity and provides wider information about the entities mentioned. In order to address this issue, we propose a novel deep learning approach utilizing multimodal deep learning. Our solution is able to provide more accurate results on named entity recognition task. Experimental results, namely the precision, recall and F1 score metrics show the superiority of our work compared to other state-of-the-art NER solutions.

https://arxiv.org/abs/2001.06888

❇️ @AI_Python_EN

Decision trees are extremely fast when it comes to classify unknown records. Watch this video to know how Decision Tree algorithm works, in an easy way - http://bit.ly/2Ggsb9l

#DataScience #MachineLearning #AI #ML #ReinforcementLearning #Analytics #CloudComputing #Python #DeepLearning #BigData #Hadoop

a new NLU benchmark for testing the ability of models to break down a question into the required steps for computing its answer. https://allenai.github.io/Break/ A work by Tomer Wolfson, accepted to TACL 2020.