Jonathan Davis

Data Science Consultant

[Jonathan Davis @Jonnynd]

For anyone wondering how @Google trained a model with 1.6 trillion parameters (GPT-3 has 175 billion), I've written an article in @TDataScience explaining the Switch Transformer architecture.
#MachineLearning #DataScience #DeepLearning #Transformers #AI https://t.co/hVP8HtlhGE

[Jonathan Davis @Jonnynd]

Ever wondered what happens when you set fp16 to True? My latest article looks at how mixed precision training reduces training time and memory requirements without affecting performance.
#DeepLearning #DataScience #nvidia
https://t.co/oMNxXu4son

[Jonathan Davis @Jonnynd]

A great new benchmark task for meta-reinforcement learning (learning to learn) by @DeepMind. Reminds us of the importance of benchmark tasks/datasets for ML!
#DataScience #DeepLearning #ReinforcementLearning
https://t.co/DaxGJavNnA

[Jonathan Davis @Jonnynd]

Disappointed by the inconsistency and lack of clarity surrounding statistics during Covid-19,
I've published in @TDataScience, looking at some statistical fallacies that can help us understand pandemics. #Statistics #COVID19 #data #Epidemiology
https://t.co/NOKM37KE88

[Tealfeed @tealfeed]

Presenting the top creators in Artificial Intelligence on CoFoundersTown!
#ai #ArtificialIntelligence

@moizsaifee @srinivasancj @OKharkovyna @nbiedrzy @IZehori @Jonnynd @Liberoliber @alexieidingli
@Ritika_Borkar @lopezyse

[Towards Data Science @TDataScience]

Understanding Mixed Precision Training by Jonathan Davis https://t.co/52lyhN13Ly

[Jonathan Davis @Jonnynd]

It's difficult to ensure our own biases don't affect data analysis, and although our assumptions and prior beliefs make it very difficult, biased analysis can limit insight and block change.
https://t.co/qmyXNktp7l
#DataScience #DataAnalytics #data #bias

[Jonathan Davis @Jonnynd]

People are selectively using data to reinforce whatever opinions they have on COVID, a great example of confirmation bias. We need to accept the uncertainty of what will happen next! #covid19 #data #confirmationbias #DataAnalytics
https://t.co/sQVd6R1B2F

[Jonathan Davis @Jonnynd]

My latest article explores how data scientists can use concepts in behavioural economics to understand how people think, improving the efficacy of machine learning models.
#MachineLearning #DataScience #cognitivebias #behaviouraleconomics #psychology
https://t.co/BbvnlnE6Mk

[Jonathan Davis @Jonnynd]

Following my article on AI and art, I've now looked at a different industry, AI in medicine.
#ai #DataScience #medicine
https://t.co/NYJ38nP88l

[Jonathan Davis @Jonnynd]

All these GPT-3 articles are great fun, but make sure to check the editor's note at the bottom. It's interesting to see that the Guardian treated it just like any human-written article during editing.
#datascience #machinelearning #deeplearning #nlp #GPT3
https://t.co/6g5ipvSgbs

[Jonathan Davis @Jonnynd]

AI projects are often over before they begin. This recent article I wrote lists important conversations that need to occur early in the planning process to improve the chances of success. #MachineLearning #ai #DataScience #projectmanagement
https://t.co/ctw7Ava9No

[Google DeepMind @GoogleDeepMind]

Exciting results from our partnership with @googlemaps! Using advanced Graph Neural Networks we were able to improve the accuracy of ETAs in major cities across the world by up to 50% https://t.co/ZOzy7AlBuX

[Hugging Face @huggingface]

Want speedy transformers models w/o a GPU?! 🧐
Starting with transformers v3.1.0 your models can now run at the speed of light on commodity CPUs thanks to ONNX Runtime quantization!🚀. Check out our 2nd blog post with ONNX Runtime on the subject! 🔥
https://t.co/OzPZ8y6XuW

[Jonathan Davis @Jonnynd]

Putting the art in ARTificial Intelligence... A recent article I wrote, looking at examples of AI being used to create different forms of art, from music to painting! #AI #DataScience #art #MachineLearning #DeepLearning #music
https://t.co/gdiH44azSx

[Composio @composioco]

Why All Data Scientists Should Understand Behavioral Economics -- https://t.co/oLWrhZIA3p --

Understanding behavioral economics can help data scientists create better, more effective machine learning models.

Continue reading on The Startup »

[Jonathan Davis @Jonnynd]

Another recent article I wrote. This one gives real examples of how I've worked daily to continually improve the quality and efficiency of my Python code. #Python #programming #coding #selfimprovement
https://t.co/Nr5Z4FDR1H

[Daniel Bigham @danielbigham]

GPT-3 is still surprising me:

[Jonathan Davis @Jonnynd]

Not only is code translation an interesting twist on natural language translation, but @facebookai have managed to develop a semi-supervised technique, without a complete training set! #DataScience #programming #NLP https://t.co/HcSlH2D68T

[Jonathan Davis @Jonnynd]

A nice, novel way to visualize classification results
#DataScience #MachineLearning
https://t.co/IDPtYCr8mF