Top Twitter Topics by Data Scientists #15

Trending this week: Speech-To-Text with Wav2Vec 2.0; Algorithmic bias in data problem; Use AWS SageMaker Clarify to spot ML bias; Bitcoin Sell-Off.

Every week we analyze the most discussed topics on Twitter by Data Science & AI influencers.

The following topics, URLs, resources, and tweets have been automatically extracted using a topic modeling technique based on Sentence BERT, which we have enhanced to fit our use case.

Want to know more about the methodology used? Jump into this article for more details, and find the codes in this Github repository!


This week, Data Science and AI influencers on Twitter have talked about:

  • Machine & Deep Learning Applications
  • Bias In Machine Learning
  • Bitcoin Sell-Off

The following sections provide all the details for each topic.

Machine & Deep Learning Applications

This week, the influencers have shared articles and papers related to machine learning and deep learning applications.

KDnuggets has shared an article explaining how to perform Speech-to-Text with Wav2Vec 2.0, an open-source framework recently introduced by Facebook AI for self-supervised learning of representations from raw audio data. 

You can learn more about the Wav2Vec 2.0 model on Facebook AI’s blog. Also, the latest version of Hugging Face transformers is version 4.30 and it comes with Wav2Vec 2.0. This is the first Automatic Speech recognition speech model included in the Transformers.

On his side, Dr. Ganapath Pulipaka shared a research paper titled “Dynamical machine learning volumetric reconstruction of objects’ interiors from limited angular views.” This paper presents an approach that accurately reconstructs volume interiors using a Recurrent Neural Network (RNN) architecture with a novel Separable-Convolution Gated Recurrent Unit (SC-GRU) as the fundamental building block.

Finally, Kirk Borne shared a 15 mins video explaining the main ideas of Adjoint Differentiation and Backpropagation, with application in Machine Learning and Finance.

Bias In Machine Learning

This week, the influencers have shared resources related to bias in machine learning.

Sandra Wachter, associate professor and senior research fellow in law and ethics at Oxford Internet Institute, shared an article published on Amazon Science latest news titled “How a paper by three Oxford academics influenced AWS bias and explainability software”. This article explains how the Amazon SageMaker Clarify tool uses one metric in particular — conditional demographic disparity (CDD) — that was inspired by a research paper titled “Why Fairness Cannot Be Automated: Bridging the Gap Between EU Non-Discrimination Law and AI”, co-authored by Sandra Wachter.

She also shared a post talking about the lack of a universal AI ethics framework that has led to a haphazard approach. This article talks about the fear that audits might not just fail to catch bias, but legitimize harmful technologies. It explains that many techniques for detecting prejudice in algorithms are in fact bias preserving because they assume the status quo in society is fair.

Hugo Larochelle has shared a post titled Moving beyond “algorithmic bias is a data problem”. This article explains that algorithms are not impartial, and how do our model design choices could contribute to algorithmic bias. This post mitigates the precarious assumption that bias can be fully addressed in the data pipeline, it suggests that in a world where our datasets are far from perfect, overall harm is a product of both the data and our model design choices.

Bitcoin Sell-Off

Bitcoin plummets as much as 15% just days after hitting

a record high. This created a buzz in the data science community

Jean-Baptiste Lefevre shared an infographic exhibiting how Bitcoin beats giants like Apple, Microsoft, Amazon, and Google to be the fastest asset to reach a $1 trillion market cap.

Carlos E. Perez commented on a thread discussing if black market transactions are in BTC, there will always be some price pressure in. The black market is $10T, current BTC cap is $1T. He reasoned that how the equilibrium could be a $500k figure for Bitcoin.

He also raised questions on the rationale behind a tweet that claimed that bitcoin could reach 69K on 4/20.

Colin McGuire shared an article explaining that Dogecoin spikes 400% in a week, stoking fears of a cryptocurrency bubble. Like many cryptocurrencies, it has a tendency for volatile swings in price, leading to strong speculations that this is a bubble.

Finally, Yves Mulkers shared an article on how the NFT market has exploded, driving up the price of digital artworks to fantastical levels.

But these artworks rely on blockchain technology, which also forms the basis of cryptocurrencies like Bitcoin, and produces enormous greenhouse-gas emissions. Hence raising questions on the environmental impact of the whole NFT market.