Domino Data Lab

JUNE 13, 2019

Recently, Chhavi Yadav (NYU) and Leon Bottou (Facebook AI Research and NYU) indicated in their paper, “ Cold Case: The Lost MNIST Digits ”, how they reconstructed the MNIST (Modified National Institute of Standards and Technology) dataset and added 50,000 samples to the test set for a total of 60,000 samples. Did they overfit the test set?

Testing 83

Testing Data Science Experimentation Metadata 83

Data Science and Beyond

APRIL 5, 2021

The Australian bushfires of 2019-20 provided me with extra motivation to help nudge Automattic to do more in the fight against climate change. I summarised this work in a post on the company’s blog , and discussed it in an interview with PublishPress. This aligns well with my long-standing interest in causal inference.

Experimentation 61

Experimentation Testing Publishing Measurement 61

The Unofficial Google Data Science Blog

APRIL 23, 2024

This blog post discusses such a comprehensive approach that is used at Youtube. To find optimal values of two parameters experimentally, the obvious strategy would be to experiment with and update them in separate, sequential stages. And we can keep repeating this approach, relying on intuition and luck.

Experimentation 52

Experimentation Optimization Uncertainty Metrics 52

DataKitchen

APRIL 13, 2021

Read the complete blog below for a more detailed description of the vendors and their capabilities. Testing and Data Observability. It orchestrates complex pipelines, toolchains, and tests across teams, locations, and data centers. Testing and Data Observability. Production Monitoring and Development Testing.

Testing 300

Testing Machine Learning Consulting Data Quality 300

O'Reilly on Data

JULY 25, 2023

So far I’ve read a gazillion blog posts about people’s experiences with these AI coding assistance tools. For instance, if I’m reading a paper from 2019, a popular song from that year could start playing. Swift Papers felt like a well-scoped project to test how well AI handles a realistic yet manageable real-world programming task.

Consulting 246

Consulting Interactive Software IT 246

Domino Data Lab

MAY 8, 2019

Finale Doshi-Velez, Been Kim (2017-02-28) ; see also the Domino blog article about TCAV. They also require advanced skills in statistics, experimental design, causal inference, and so on – more than most data science teams will have. Other good related papers include: “ Towards A Rigorous Science of Interpretable Machine Learning ”.

Machine Learning 105

Machine Learning Data Science Modeling Visualization 105

Perceptual Edge

AUGUST 26, 2019

A serious approach would begin with a thorough understanding of data visualization, which is not Pangilinan’s area of expertise, and would then proceed scientifically by designing and running experimental studies to test its usefulness. Her case is hollow.

Visualization 52

Visualization Cost-Benefit Interactive Experimentation 52