Metric Learning Tips & Tricks

In this article, the author presents ways of overcoming the limitations of classification, such as the number of training samples, production integration, and scaling. Specifically, he’ll explain how to train an object matching model with no labeled data and use it in production, to ensure metric learning is more scalable and flexible.

https://bit.ly/3cznb05

Subscribe to our weekly newsletter — https://bit.ly/3gqPUp5

Tinkering with the Mobile Apps Dataset

In this article, the author demonstrates how you can use an open-source dataset featuring mobile apps data to build your own models. The article includes such steps as choosing a dataset, exploratory data analysis, feature engineering, and predicting with a model. The dataset and the models are available for re-use.

https://bit.ly/3xiMQT0

Building Scalable Machine Learning Pipelines for Multimodal Health Data on AWS

Machine learning is used extensively in the healthcare and life sciences industries. Among many approaches and methods to increase the accuracy and efficiency of ML models, Multimodal ML stands out as one of the most promising. In this article, you’ll learn how to build a scalable, cloud architecture for Multimodal ML on health data.

https://amzn.to/3wmXM1D

PyCon US 2021 [Conference Materials]

This playlist features all keynotes, talks, and other materials from PyCon US 2021, a virtual conference for the community using and developing the open-source Python programming language. Over 80 videos in total!

https://bit.ly/2TsBJIi

Session-based Recommender Systems

In this extensive research report by Cloudera Fast Forward, you’ll learn all the ins and outs of designing, building, and managing AI/ML-powered recommender systems. The authors will demonstrate how to use specific algorithms and datasets to arrive at conclusions about the do’s and don’ts of building such systems (e.g. while using word2vec).

https://bit.ly/3pS8URC

Dynamically Generating DAGs in Airflow

In this guide, the Astronomer team looks into specific methods of dynamically generating DAGs in Airflow, from single-file methods to multiple-file methods. Every method is accompanied by code and examples. The team also presents DAG Factory, an open source Python library for dynamically generating Airflow DAGs from YAML files.

https://bit.ly/3xt8RhF

​​Data Science Digest — 17.06.21

The new issue of DataScienceDigest is here! Facebook AI migrates its systems to PyTorch, metric learning tips & tricks, session-based recommender systems, AndroidEnv, materials from PyCon US 2021, and more…

https://bit.ly/3vshrMs

Join 👉@DataScienceDigest

Mathematics for Machine Learning

«Mathematics for Machine Learning» by Marc Peter Deisenroth, A. Aldo Faisal, and Cheng Soon Ong brings the mathematical foundations of basic ML concepts to all those who struggle with the mathematical knowledge required to read an ML textbook. This book is intended to be a guidebook to the vast mathematical literature that forms the foundations of modern machine learning.

https://bit.ly/3gDHfkG

Conversational AI

In this video, Merve Noyan, Google Developer Expert on Machine Learning, gives an overview of the conversational AI niche. The talk is hosted by Alexey Grigorev, the founder of DataTalks.Club.

https://bit.ly/35ESW4d

Deep Learning Do It Yourself!

The website is a collection of more than 20 modules on learning deep learning. As a student, you can walk through the modules at your own pace and interact with others. You can also contribute to the materials by adding new modules yourself.

https://bit.ly/3vMub0l

#DataScienceDigest #DataScience #MachineLearning #ArtificialIntelligence #AI #ML #deeplearning

Subscribe to our weekly newsletter — https://bit.ly/3cS4G7y

Comparing Test Sets with Item Response Theory

In this paper, Clara Vania et al. use the Item Response Theory to evaluate 29 datasets using predictions from 18 pretrained Transformer models on individual test examples. Quoref, HellaSwag, and MC-TACO are best suited for distinguishing among state-of-the-art models, while SNLI, MNLI, and CommitmentBank seem to be saturated for current strong models.

https://bit.ly/3xHDkJj

How Airbnb Standardized Metric Computation at Scale

The engineering team of Airbnb reveals the design principles of Minerva compute infrastructure. Minerva is a single source of truth metric platform that standardizes the way business metrics are created, computed, served, and consumed. The article features the link to the first post on Minerva. Check it out, too!

https://bit.ly/3zLDv8g

AI Can Now Emulate Text Style in Images in One Shot — Using Just a Single Word

In this article, the engineering team of Facebook AI presents TextStyleBrush, an AI research project that can copy the style of text in a photo using just a single word. With this AI model, you can edit and replace text in images. The team hopes to spur dialogue and research into detecting potential misuse of this type of technology, so make sure to contribute.

https://bit.ly/3wU3t7I

​​Data Science Digest — 24.06.21

The new issue of DataScienceDigest is here! The impact of NLP and the growing budgets to drive AI transformations. How Airbnb standardized metric computation at scale. Cross-Validation, MASA-SR, AgileGAN, EfficientNetV2, and more…

https://bit.ly/3qnuy0u

Join 👉@DataScienceDigest

Ingestion and Historization in the Data Lake

In this video, Alexey Grigorev, the founder of DataTalks.Club, hosts Illia Todor, Data Engineer, to talk about ingestion and historization of data in the data lake.

https://bit.ly/3zWKaMP

MLOps Toys
The platform is a collection of MLOps projects by category, including data versioning, training orchestration, feature store, experiment tracking, model serving, model monitoring, and explainability.

https://bit.ly/3xV97GF

DynamicViT: Efficient Vision Transformers with Dynamic Token Sparsification

In this research, Yongming Rao et al. propose a dynamic token sparsification framework to prune redundant tokens progressively and dynamically based on the input. A lightweight prediction module can estimate the importance score of each token given the current features. The module is added to different layers to prune redundant tokens hierarchically.

Web Page — https://bit.ly/3dgESlq
Paper — https://bit.ly/3xWf5XJ
Code — https://bit.ly/3jlUtUK