Skip to main content Skip to site alert

Secondary Navigation

  • Menu  Close 
  • Search
  • Info For Navigation

    • Industry
    • PhD Students
    • Graduate Studies
    • People
    • PhD Students
    • Master’s Students
Department of Applied Mathematics and Statistics
Department of Applied Mathematics and Statistics

Utility Navigation

  • Request Info
  • Apply
  • Give

Site Navigation

  • Academics
    • Degrees & Programs
    • Undergraduate Studies
    • Graduate Studies
  • Research
    • Computational and Applied Mathematics
    • Data Science
    • Discrete Mathematics
    • Financial Mathematics Research
    • Optimization and Operations Research
    • Probability and Stochastic Processes
    • Statistics and Machine Learning
    • Duncan Research Fund
    • Dissertations
  • People
    • Faculty
    • Affiliated Faculty
    • Staff
    • Postdoctoral Fellows
    • PhD Students
    • Master’s Students
  • News & Events
    • News
    • Events
    • Seminars and Endowed Lectures
  • Career Resources
    • Life Design Lab
    • Graduate Career Services
    • Employers and Recruitment
  • Alumni & Giving
  • About
    • History
    • Message from the Department Head
    • Employment Opportunities
Search

Info For Navigation

  • Industry
  • PhD Students
  • Graduate Studies
  • People
  • PhD Students
  • Master’s Students

You are here:

  1. Home
  2. News
  3. Ben Grimmer challenges assumptions for faster optimization

Ben Grimmer challenges assumptions for faster optimization

Quanta Magazine features Ben Grimmer's recent study on a traditional assumption in gradient descent, revealing that breaking the rule of small steps can lead to nearly three times faster results on optimization problems.

Published: August 11, 2023
Category:
  • Research
  • Department News

Share Options

  • Share toTwitter
  • Share toFacebook
  • Share toLinkedIn

A recent issue of Quanta Magazine featured a story spotlighting new research by Assistant Professor Benjamin Grimmer. Grimmer’s study suggests that a basic assumption about gradient descent, a long-relied on optimization method used in many machine learning programs, may be wrong.

Stay Connected

  • Facebook
  • Twitter
  • Instagram
  • YouTube
  • LinkedIn

Department of Applied Mathematics and Statistics

Address

3400 North Charles Street Baltimore, MD 21218
Get Directions

Contact

Email: [email protected]

Footer Navigation

  • Apply Now
  • Whiting School of Engineering
  • Johns Hopkins University

Legal Navigation

  • Privacy Statement
  • Accessibility
  • University Policies
  • Copyright Compliance Policy

2023 Johns Hopkins University. All rights reserved.

Site Menu

Site Navigation

  • Academics
    • Degrees & Programs
    • Undergraduate Studies
    • Graduate Studies
  • Research
    • Computational and Applied Mathematics
    • Data Science
    • Discrete Mathematics
    • Financial Mathematics Research
    • Optimization and Operations Research
    • Probability and Stochastic Processes
    • Statistics and Machine Learning
    • Duncan Research Fund
    • Dissertations
  • People
    • Faculty
    • Affiliated Faculty
    • Staff
    • Postdoctoral Fellows
    • PhD Students
    • Master’s Students
  • News & Events
    • News
    • Events
    • Seminars and Endowed Lectures
  • Career Resources
    • Life Design Lab
    • Graduate Career Services
    • Employers and Recruitment
  • Alumni & Giving
  • About
    • History
    • Message from the Department Head
    • Employment Opportunities

Utility Navigation

  • Request Info
  • Apply
  • Give

Secondary Navigation