Learn the Universe — an ML X Cosmology Workshop

The Center for Computational Astrophysics (CCA) is located at 162 5th Avenue. The entrance to the Flatiron Institute is on 21st Street. 

&npsp;

Please NOTE that anyone requesting travel and/or hotel funding will be contacted separately.

Hotel accommodations will be at the James New York – NoMad Hotel located at 22 E 29th Street between Madison and 5th Avenues: http://www.jameshotels.com/new-york-nomad/.

Check-in at the hotel is at 2:00 pm. Check-out is at 11:00 am.

Please email [email protected] or [email protected] for the reservation link.

Travel arrangements must be made by Monday, August 16th.

Tuesday:
10:00am-10:15am Introduction- Welcome & Outline of the Workshop
 
10:15am-11:15am structured section: Simulations Based Inference
Chair: Francisco Villaescusa-Navarro

1. Roger de Belsunce: Unbiased inference from large-scale CMB data using likelihood-approximation schemes
2. ChangHoon Hahn: Higher-Order LSS with SBI
3. Ben Horowitz: hyphy – Conditional Posterior Surrogate Modeling of Hydrodynamical Physics
4. Boryana Hadzhiyska: TBD
5. Digvijay Wadekar : Symbolic regression for cluster mass estimation

 
11:45am-12:15pm Coffee
 
12:15pm-1:00 pm unstructured section: Problems we can’t solve with current tools that we may have a shot with ML
Chair: Simone Ferraro and Colin Hill
 
 
1:00pm-2:00pm: Lunch
 
2:00pm-3:45pm: structured section: Extracting Non-Gaussian Information
Chair: Andrina Nicola

1. Sihao Cheng: The scattering transform in cosmology, or, a CNN without Training
2. Kate Storey-Fisher: Emulation of Summary Statistics for Cosmology from Galaxy Surveys
3. Agnès Ferté: Categorizing cosmological models with unsupervised learning
4. Colin Hill: Non-Gaussian Information in CMB Secondary Anisotropies
5. Luisa Lucie-Smith: Deep learning insights into dark matter halo formation

 
3:45pm-4:15pm: Coffee
 
4:15pm-5:00pm: unstructured section: Have non-Gaussian statistics been made redundant?
Chair: Zoltan Haiman
 
RECEPTION
 
 
 
Wednesday:
10:00am-11:45am structured section: Simulation X ML
Chair: Sultan Hassan

1. Francisco Villaescusa-Navarro: The Cosmology and Astrophysics with Machine Learning Simulations (CAMELS) project
2. Siamak Ravanbakhsh: Deep Networks for Spherical Data
3. Neerav Kaushal: Mapping from Fast Simulations to Full N-body Simulations
4. Yin Li: Cosmological Forward Modeling with Adjoint Method
5. Leander Thiele: DeepSets applied to Clusters: Machine learning the Lagrangian way

 
11:45am-12:15pm Coffee
 
12:15pm-1:00pm: unstructured section: Pitch your Astro challenge in 5 mins OR Pitch your method in 5 mins
(participants can prepare a slide or two for this session)
Chair: ChangHoon Han and Miles Cranmer
 
 
1:00pm-2:00pm: Lunch
 
2:00pm-3:45pm: structured section: Robust ML for science
Chair: Yin Li

1. Biwei Dai: Translation and Rotation Equivariant Normalizing Flow (TRENF) for Optimal Cosmological Analysis
2. David Yallup : Principled Bayesian Neural Networks
3. Miles Cranmer: Histogram Pooling Operators for Interpretable Deep Learning in Cosmology

 
3:45pm-4:15pm: Coffee
 
4:15pm-5:00pm: unstructured section: ML 4 Science – Promises and Problems
Chair: Viviana Acquaviva
 
DINNER
 
 
 
Thursday:
10:00am-11:00am unstructured section: The next Decade in Cosmology
Chair: David Spergel
 
 
11:00am-11:30am Coffee
 
11:30am-1:00pm: structured section: Cosmological Applications
Chair: Alice Pisani

1. Andrina Nicola : Forecasting cosmological and astrophysical constraints from electron-matter cross-correlations
2. Christina Kreisch: Precision Cosmology from Voids in the Machine Learning Era
3. Ira Wolfson: The Fault In Our Spectrum: A ‘no go’ on small field models analytics
4. Adrian Bayer: The Look-Elsewhere Effect
5. Max Lee: Gradient based inference in cosmology using MADLens
6. Alina Sabyr: Cosmological Constraints from Weak Lensing Peaks: Can Halo Models Accurately Predict Peak Counts?

 
 
1:00pm-2:00pm: Lunch
 
2:00pm-3:45pm: structured section: Simons Collaboration “Learning the Universe”
Chair: Shirley Ho

1.  Greg Bryan, Introducing the Simons Collaboration “Learning the Universe”
2. Rachel Somerville: New methods to model galaxy formation so that we can Learn the Universe
3. Ana Maria Delgado: Modeling Galaxy-Halo connection with Machine Learning
4. Sultan Hassan: HIFlow: Fast Emulator of HI maps using Normalizing Flow.

 
 
3:45-4:15 pm: Coffee
 
4:15pm-5:00pm: unstructured section: What qualifies as interpretability? Do we need it?
Chair: Julia Kempe and Vanessa Bohm
 
 
SOCIAL