Skip to content

Berkeley AI Research Blog company blog

How to Escape Saddle Points Efficiently

by C.K. Wolfe bair.berkeley.edu 11 min read

The BAIR Blog

Read on bair.berkeley.edu → Discuss on HN (38) Archived copy

Points: 144
HN comments: 38
Submissions: 1
Cited by: 0 commenters

AI & Machine Learning 2017

On Hacker News

Submitted	Title on HN	Points	Comments
Sep 1, 2017	How to Escape Saddle Points Efficiently by mfagio	144	38

More from Berkeley AI Research Blog

All 6 posts →

Does GPT-2 Know Your Phone Number?

2020321 points155 comments AI & Machine Learning details

Reinforcement learning is supervised learning on optimized data

2020241 points18 comments AI & Machine Learning details

Learning Diverse Skills via Maximum Entropy Deep Reinforcement Learning

2017139 points15 comments AI & Machine Learning details

Sample Efficient Evolutionary Algorithm for Analog Circuit Design

2019114 points25 comments AI & Machine Learning details

End-to-End Deep Reinforcement Learning Without Reward Engineering

2019112 points4 comments AI & Machine Learning details

Related posts from the same era

Google Duplex: An AI System for Accomplishing Real World Tasks Over the Phone

Google AI Blog20181,875 points750 commentscited by 10 details

How Technology Hijacks People’s Minds

Tristan Harris20161,612 points427 commentscited by 7 details

Announcing SyntaxNet: The World’s Most Accurate Natural Language Parser

Google Research Blog20161,083 points235 comments details

AI winter is well on its way

Filip Piekniewski2018993 points492 commentscited by 3 details

AlphaGo Zero: Learning from scratch

DeepMind Blog2017923 points311 commentscited by 5 details

DeepMind and Blizzard to release StarCraft II as an AI research environment

DeepMind Blog2016917 points329 commentscited by 4 details