Open-source research · ML4SCI

Open research in machine learning, physics, and astronomy.

DeepLense is an open-source research group studying dark matter with strong gravitational lenses and developing machine-learning methods for scientific data.

01

Strong gravitational lensing. A foreground galaxy bends light from a distant source into arcs and rings. Dark-matter structure can perturb their shape.

AAAI 2026 Fall Symposium DeepLense AI Scientist presented at the AAAI Fall Symposium View the event Published and presented at The Astrophysical Journal Machine Learning: Science and Technology AAAI Fall Symposium CATS at ICML ML4PS at NeurIPS RPS at NeurIPS DiffuLM at NeurIPS Publication record
What is DeepLense?

Open research on dark matter and strong lensing.

Strong lenses provide a way to study matter that does not emit light. We build models that identify lenses, characterize their morphology, and test how much physical information can be recovered from images.

The work includes simulation, survey-data processing, representation learning, super-resolution, parameter inference, and experimental tools for scientific agents.

How the group works

Research

Two connected areas of research.

We develop methods for astrophysical data and study machine-learning problems that recur across the sciences.

A

AI for Science

Astrophysics and scientific computing.

  • Astrophysics
  • Strong Lensing
  • Scientific Simulation
  • PDEs
  • Autonomous Discovery
B

Core AI

Machine-learning methods.

  • Agents
  • Foundation Models
  • Generative Models
  • Representation Learning
  • Interpretability & Evaluation

Research workflow

From simulated lenses to physical inference.

The group works across the stages needed to connect controlled simulations with astronomical observations.

01

Simulate

Generate controlled lensing systems across candidate dark-matter models.

02

Detect

Identify rare lens candidates in large, noisy astronomical surveys.

03

Model

Train representations using known symmetries, geometry, and lensing physics.

04

Evaluate

Measure what the models recover and how reliably results transfer to survey data.

Different measurements require different representations.

A symmetry that improves classification can remove detail needed for reconstruction. A sharper image may not improve a physical estimate. We evaluate models against the scientific task they are intended to support.

News & project notes

Group updates.

Events and student writing

Organization

DeepLense is part of ML4SCI.

Machine Learning for Science (ML4SCI) is the nonprofit open-source organization that supports DeepLense and related research programs across scientific disciplines.

Explore DeepLense

Read the papers.
Use the software.