NYU Center for Data Science – Medium

NYU Center for Data Science

18.1K Followers

Why Win Rate Should Be the Guiding Principle in Preference Learning

CDS’ Lily H. Zhang and Rajesh Ranganath show why win rate should be the standard for evaluating models trained on preference data.

13h ago

Why Win Rate Should Be the Guiding Principle in Preference Learning

13h ago

Seeing the Unseeable: How AI Reveals Atomic-Level Dynamics in Nanoparticles

A new AI technique lets scientists watch atoms move in real time, revealing previously unseen surface dynamics.

2d ago

Seeing the Unseeable: How AI Reveals Atomic-Level Dynamics in Nanoparticles

2d ago

AI Isn’t Always Helpful — and Researchers Have Found a Way to Know When

CDS Faculty Fellow Umang Bhatt and visiting researcher Valerie Chen present MODISTE, a system that learns when a user benefits from AI…

Apr 18

AI Isn’t Always Helpful — and Researchers Have Found a Way to Know When

Apr 18

Training Transformers: Formal Languages as the Key to Efficient Learning

Pretraining transformers on formal languages improves learning efficiency and boosts generalization in syntax tasks.

Apr 17

structured strings of brackets

Apr 17

Neural Compression Advantages Diminish as Image Size Increases

Research from CDS’ Jingtong Su and colleagues: Neural network-based image compression loses its advantage as image size increases.

Apr 11

Neural Compression Advantages Diminish as Image Size Increases

Apr 11

What is a Goal? Advancing Machine Agency While Understanding Human Goal Creation

New CDS research models how people understand and formulate goals — knowledge that could improve AI alignment

Apr 10

What is a Goal? Advancing Machine Agency While Understanding Human Goal Creation

Apr 10

AI Model Discovers New Lung Cancer Subtype Associated with Poor Survival

CANVAS, an AI framework developed by CDS researchers and others, has identified a new lung cancer subtype associated with poor survival.

Apr 4

AI Model Discovers New Lung Cancer Subtype Associated with Poor Survival

Apr 4

Making the Invisible Visible: New Method Reveals How Language Models Encode Concepts

A new technique makes language model biases visible by translating representation-level interventions into natural language text.

Apr 2

Making the Invisible Visible: New Method Reveals How Language Models Encode Concepts

Apr 2

New CDS Admissions Ambassadors Bring a Range of Perspectives to the MS in Data Science Program

Six CDS Admissions Ambassadors share what it’s really like to be in the MS in Data Science program.

Mar 28

New CDS Admissions Ambassadors Bring a Range of Perspectives to the MS in Data Science Program

Mar 28

The ImageNet Moment for Physics Simulation: CDS Researcher Leads Creation of “The Well”

CDS Senior Research Scientist Shirley Ho has led the creation of The Well, a 15TB dataset designed to advance physics-based machine…

Mar 26

The ImageNet Moment for Physics Simulation: CDS Researcher Leads Creation of “The Well”

Mar 26

NYU Center for Data Science

NYU Center for Data Science

18.1K Followers

Official account of the Center for Data Science at NYU, home of the Undergraduate, Master’s, and Ph.D. programs in Data Science.

Following

Help
Status
About
Careers
Press
Blog
Privacy
Rules
Terms
Text to speech