Feature Engineering

topic on 2 shows · 5 statements across 5 episodes

No Priors the MAD Podcast

5 statements about Feature Engineering, every show

NO PRIORS Prediction Not checkable as stated
Guu: Fundamental AI Advances May Make Agent Workflow Engineering Obsolete
“But it's quite possible that fundamental advances could come along and this kind of workflow hacking or workflow engineering could go the same way as feature engineering or specialized architectures.”
Kelvin Guu May 4, 2023 ▶ 21:28 No Priors Ep. 15 | With Kelvin Guu, Staff Research Scientist, Google Brain
MAD Insight
Alex Dean: Schematized data creation accelerates feature engineering and model training
“Because you're deliberately creating this data, because you're creating it in a very schematized way, it's much quicker to get to the feature Engineering, it's much quicker to get to the model training.”
Alex Dean Oct 24, 2022 ▶ 4:51 Behavioral Data Creation for AI | Snowplow Co-Founder & CEO Alex Dean
MAD Assertion Not checkable as stated
Mason: Deep learning completely removes the need for domain feature engineering
“In deep learning, you're not doing feature engineering anymore. Nobody cares about what you know about the domain.”
Hilary Mason Dec 8, 2016 ▶ 20:20 A Process for Discovery // Hilary Mason, Fast Forward Labs [FirstMark's Data Driven]
MAD Insight
LeCun: Deep learning automates feature engineering by integrating representation with classification
“And deep learning basically opens the possibility that you, to automate the process of feature engineering. So, it essentially learns to, It integrates the process of producing good representations with the process of learning a classifier in one fell swoop.”
Yann LeCun Dec 18, 2014 ▶ 17:47 Yann Lecun, Facebook // Artificial Intelligence // Data Driven #32 (Hosted by FirstMark Capital)
MAD Assertion Supported
Kaggle competition winners rely primarily on feature engineering and algorithm ensembles
“If you look at, you know, who wins most of the Kaggle competitions, it tends to be, you know, combinations of ensembles of different algorithms, but it's really the feature engineering.”
Mike Abbott Oct 16, 2014 ▶ 5:03 Mike Abbott, KPCB // Data Driven #30 // Oct 2014 (Hosted by FirstMark Capital)

← every entity, every show

Made with StarZero

Turn any episode into a week of clips.

This entire site, thousands of episodes across every show transcribed, diarized, checked and made playable, runs on the StarZero media pipeline. Drop in your own episode and the podcast clipper finds the moments worth sharing, cuts them, captions them, and reframes them for every feed.