DiscoverThe Data Exchange with Ben LoricaExhaustion of High-Quality Data Could Slow Down AI Progress in Coming Decades
Exhaustion of High-Quality Data Could Slow Down AI Progress in Coming Decades

Exhaustion of High-Quality Data Could Slow Down AI Progress in Coming Decades

Update: 2023-03-161
Share

Description

Pablo Villalobos is a Staff Researcher at  Epoch, and lead author of the recent paper “Will we run out of data? An analysis of the limits of scaling datasets in Machine Learning”.  We discuss the key findings in this paper, as well as a related study Pablo conducted on scaling laws. 

Subscribe to the Gradient Flow Newsletter:  https://gradientflow.substack.com/

Subscribe: Apple • Spotify • Stitcher • Google • AntennaPod • Podcast Addict • Amazon •  RSS.

Detailed show notes can be found on The Data Exchange web site.

Comments 
In Channel
Building and Deploying Foundation Models for Enterprises

Building and Deploying Foundation Models for Enterprises

2023-06-0134:09

Building Robust AI Infrastructure for Critical Solutions

Building Robust AI Infrastructure for Critical Solutions

2023-05-2531:40

Machine Learning for High-Risk Applications

Machine Learning for High-Risk Applications

2023-05-1846:25

Boosting Perception With Synthetic Data

Boosting Perception With Synthetic Data

2023-05-1135:34

Revolutionizing B2B: Unleashing the Power of AI and Data

Revolutionizing B2B: Unleashing the Power of AI and Data

2023-05-0443:08

AI Metadata

AI Metadata

2023-04-2731:30

The 2023 AI Index

The 2023 AI Index

2023-04-2043:38

Custom Foundation Models

Custom Foundation Models

2023-04-1338:06

Uncovering and Highlighting AI Trends

Uncovering and Highlighting AI Trends

2023-04-0649:02

How Data and AI Happened

How Data and AI Happened

2023-03-3048:48

Blazing fast bulk data transfers between any cloud

Blazing fast bulk data transfers between any cloud

2023-03-2331:29

Exhaustion of High-Quality Data Could Slow Down AI Progress in Coming Decades

Exhaustion of High-Quality Data Could Slow Down AI Progress in Coming Decades

2023-03-1633:09

Generating high-fidelity and privacy-preserving synthetic data

Generating high-fidelity and privacy-preserving synthetic data

2023-03-0935:47

How technology is disrupting the venture capital industry

How technology is disrupting the venture capital industry

2023-03-0236:19

Running Machine Learning Workloads On Any Cloud

Running Machine Learning Workloads On Any Cloud

2023-02-2337:11

2023 Trends in Data Engineering and Infrastructure

2023 Trends in Data Engineering and Infrastructure

2023-02-1645:47

Preparing for the Implementation of the EU AI Act and Other AI Regulations

Preparing for the Implementation of the EU AI Act and Other AI Regulations

2023-02-0936:38

The Open Source Stack Unleashing a Game-Changing AI Hardware Shift

The Open Source Stack Unleashing a Game-Changing AI Hardware Shift

2023-02-0241:55

Data Science and AI in Context

Data Science and AI in Context

2023-01-2647:40

Evaluating Language Models

Evaluating Language Models

2023-01-1945:36

loading
Google Play
Download from Google Play
Castbox
Download from App Store
usUnited States
00:00
00:00
x

0.5x

0.8x

1.0x

1.25x

1.5x

2.0x

3.0x

Sleep Timer

Off

End of Episode

5 Minutes

10 Minutes

15 Minutes

30 Minutes

45 Minutes

60 Minutes

120 Minutes

Exhaustion of High-Quality Data Could Slow Down AI Progress in Coming Decades

Exhaustion of High-Quality Data Could Slow Down AI Progress in Coming Decades

Ben Lorica