Data Science Digest

2021-05-03 11:55:10 ​​Deep Learning for Audio with the Speech Commands Dataset

If you want to learn how to train a simple model on the Speech Commands audio dataset, this article by Peter Gao is for you. He explains how to choose a dataset and handle data, train, test, tune the model, and, most importantly, how to do error analysis (and analyze failure cases) to improve model performance over time.

https://bit.ly/2SbUIpR Open / Comment

2021-05-01 09:00:16 ​​NLP Profiler

A simple but useful NLP library created by @neomatrix369. It enables Data Science practitioners to easily profile datasets with one, two, or more text columns. The library is designed to return either high-level insights or low-level/granular statistical information about the text when given a dataset and a column name containing text data, in that column. Check out the library and let us know what you think.

https://bit.ly/3xEgSBp Open / Comment

2021-04-30 15:51:44 ​​Boosting Natural Language Processing with Wikipedia

In this hands-on tutorial, Nicola Melluso explains how you can take advantage of Wikipedia to improve your Natural Language Processing models. To illustrate how it works, he takes such NLP tasks as Named Entity Recognition and Topic Modeling, and then goes deep step by step, to explain how to collect and process data, build and train the models, etc.

https://bit.ly/3tfdiul Open / Comment

2021-04-29 12:24:31 ​​Token Labeling: Training a 85.4% Top-1 Accuracy Vision Transformer with 56M Parameters on ImageNet

In this paper, Zihang Jiang, Qibin Hou et al. explore vision transformers applied to ImageNet classification. They have developed new training techniques to demonstrate that by slightly tuning the structure of vision transformers and introducing token labeling, the models can achieve better results than the CNN counterparts and other transformer-based classification models.

Paper - https://bit.ly/32VtFRQ
Code - https://bit.ly/3eDxAbn Open / Comment

2021-04-28 17:18:01 ​​Data Science Digest — 28.04.21

The new issue of DataScience Digest is here! Hop to learn about the latest articles, tutorials, research papers, and books on Data Science, AI, ML, and Big Data. All sections are prioritized for your convenience. Enjoy!

https://bit.ly/3nrBYOT

Join @DataScienceDigest Open / Comment

2021-04-27 11:58:59 ​​VideoGPT: Video Generation using VQ-VAE and Transformers

In this research paper, Wilson Yan et al. present VideoGPT, a simple architecture for scaling likelihood-based generative modeling to natural videos. Despite its simplicity, it can generate samples competitive with advanced GAN models for video generation, as well as high fidelity natural images from UCF-101 and Tumbler GIF Dataset (TGIF).

Paper - https://bit.ly/3aHbpAa
Code - https://bit.ly/32NQQxw
Demo - https://bit.ly/3dUzuFF

@DataScienceDigest Open / Comment

2021-04-26 10:00:09 ​​Shedding Light on Fairness in AI with a New Data Set

Bias and fairness in AI are highly debatable topics. To address the problem, Facebook AI has created Casual Conversations, a new dataset consisting of 45,186 videos of participants having non scripted conversations, to help AI researchers identify and evaluate the fairness of their computer vision and audio models across subgroups of age, gender, apparent skin tone, and ambient lighting.

https://bit.ly/3tRt3bX

@DataScienceDigest Open / Comment

2021-04-25 12:05:20 ​​Zero-Shot Learning: Can You Classify an Object Without Seeing It Before?

Developing machine learning models that can perform predictive functions on data it has never seen before has become an important research area called zero-shot learning. We tend to be pretty great at recognizing things in the world we never saw before, and zero-shot learning offers a possible path toward mimicking this powerful human capability.

https://bit.ly/3xxMF7c

@DataScienceDigest Open / Comment

2021-04-22 10:43:37 ​​Interpretable Machine Learning: A Guide for Making Black Box Models Explainable

The book by Christoph Molnar goes deep to explain how to make supervised machine learning models more interpretable. You’ll start by exploring the concepts of interpretability to learn about simple, interpretable models such as decision trees, decision rules, and linear regression. Then, you’ll look into general model-agnostic methods for interpreting black box models like feature importance and accumulated local effects and explaining individual predictions with Shapley values and LIME. The book focuses on ML models for tabular data and less on computer vision and natural language processing tasks. Reading the book is recommended for machine learning practitioners, data scientists, statisticians, and anyone else interested in making machine learning models interpretable.

https://bit.ly/3sH8Ofq

@DataScienceDigest Open / Comment

2021-04-21 10:35:36 ​​Lviv Data Science Summer School

Hi folks,
I’m pleased to invite you all to enroll in the Lviv Data Science Summer School, to delve into advanced methods and tools of Data Science and Machine Learning, including such domains as CV, NLP, Healthcare, Social Network Analysis, and Urban Data Science. The courses are practice-oriented and are geared towards undergraduates, Ph.D. students, and young professionals (intermediate level). The studies begin July 19–30 and will be hosted online. Make sure to apply — Spots are running fast!

https://bit.ly/2Qc0QOx

@DataScienceDigest Open / Comment