Lightnews — Scholar-powered news

Benjamin Pope @benjaminpope.bsky.social · 7h

not gonna lie

Benjamin Pope @benjaminpope.bsky.social · 8h

I think so. It might take more GPUs, and some clever optimisations to work in reduced precision to make it fast, but I think there is no fundamental reason you couldn’t do the phase retrieval and visibility extraction very fast with a nonlinear detector model

1

Reposted by Benjamin Pope

AstroMonnier @astromonnier.bsky.social · 8h

Amazing work solving a tricky calibration problem with JWST aperture masking data! Any hope to improve ground-based systematics with this general approach?
@benjaminpope.bsky.social

arxiv.org/abs/2510.09806

AMIGO: a Data-Driven Calibration of the JWST Interferometer

The James Webb Space Telescope (JWST) hosts a non-redundant Aperture Masking Interferometer (AMI) in its Near Infrared Imager and Slitless Spectrograph (NIRISS) instrument, providing the only dedicate...

arxiv.org

1 2 1

Reposted by Benjamin Pope

Chris Lintott @chrislintott.bsky.social · 10h

Cosmologists! We have a faculty job going here in Oxford - happy to answer questions. Brief version is that this remains a great & collaborative place to do science and the students are superb. 🔭 www.jobs.ac.uk/job/DOP628/a...

Associate Professor of Cosmology at University of Oxford

Recruiting now: Associate Professor of Cosmology on jobs.ac.uk. Click for details and explore more academic job opportunities on the top job board

www.jobs.ac.uk

4 29 38

Reposted by Benjamin Pope

Louis Desdoigts @gradientrider.bsky.social · 12h

Thanks @benjaminpope.bsky.social! None of this would have been possible without your guidance and unending dedication to your students 🙌🙌🎉

Benjamin Pope @benjaminpope.bsky.social · 16h

Thrilled to have two years' of work out, in a pair of papers led by @gradientrider.bsky.social and @maxecharles.bsky.social.

We've built a data-driven calibration of the James Webb Interferometer to near its fundamental limits for high-res imaging - explainer at @aunz.theconversation.com!

How we sharpened the James Webb telescope’s vision from a million kilometres away

The only Australian hardware on board the legendary telescope is starting to fulfil its duties.

theconversation.com

1 13

Benjamin Pope @benjaminpope.bsky.social · 13h

awesome!

1

Benjamin Pope @benjaminpope.bsky.social · 15h

how did we get this way as a culture

3

Benjamin Pope @benjaminpope.bsky.social · 15h

I should have done that to get it on the feeds shouldn't I

Benjamin Pope @benjaminpope.bsky.social · 15h

Congratulations!

1

Benjamin Pope @benjaminpope.bsky.social · 16h

Please don't hesitate to get in touch if you have any questions, if you want help using the code (docs still under construction!), if you've got ideas for improvements or problems to fix.

We're excited for the next chapter with AMI and applying these ideas to other instruments on Webb and Roman!

13

Benjamin Pope @benjaminpope.bsky.social · 16h

I especially want to call out @spiralstar.bsky.social for his vision and leadership, from designing AMI, proposing targets, supervising these projects, solving problems, and being a fantastic mentor to all of us. We were delighted to see him honoured with the Michelson Prize in 2024 for this.

Peter Tuthill at his desk, with Keck masks in the background

1 1 15

Benjamin Pope @benjaminpope.bsky.social · 16h

I'm immensely proud of the team for these papers. Louis has landed a dream postdoc in Leiden and Max is on the job market (so you know where to find a great Jax dev!). Shrish has been a hero from benchmarking the BFE to writing proposals, and Anand Sivaramakrishnan has guided us beginning to end.

1 13

Benjamin Pope @benjaminpope.bsky.social · 16h

This package dorito, built on top of amigo, is also available open-source! We hope you can use it in your JWST proposals and science.

GitHub - maxecharles/dorito: Differentiable Optics for Reconstructing Images from Telescope Observations

Differentiable Optics for Reconstructing Images from Telescope Observations - maxecharles/dorito

github.com

1 13

Benjamin Pope @benjaminpope.bsky.social · 16h

We can see Io rotate slightly over the hour-long timelapse, tracking volcanoes on its surface!

And check out the Conversation piece for a cool slider showing how NGC 1068 lines up perfectly with previous data from LBTI.

1 3 28

Benjamin Pope @benjaminpope.bsky.social · 16h

The raw data look like messy interferograms that are hard to understand by eye - and we enhance these data to restore beautiful images of the environment of the black hole in NGC 1068, Jupiter's moon Io, and a dusty binary system WR 137.

3x3 multipanel figure showing the AMI imaging results. Top to bottom: raw data, calibrator, and restored images, for the active galaxy NGC 1068, Jupiter's moon Io, and the colliding-wind binary WR 137.

1 4 23

Benjamin Pope @benjaminpope.bsky.social · 16h

In a companion paper, PhD student @maxecharles.bsky.social applies this to reconstructing complex images, not just fields of dots. We apply a regularised maximum likelihood method to restore clean, diffraction limited imaging by fitting directly to pixel level data.

Image reconstruction with the JWST Interferometer

Flying on board the James Webb Space Telescope (JWST) above Earth's turbulent atmosphere, the Aperture Masking Interferometer (AMI) on the NIRISS instrument is the highest-resolution infrared interfer...

arxiv.org

1 2 17

Benjamin Pope @benjaminpope.bsky.social · 16h

By forward-modelling through this, we can reverse out what the uncorrupted images look like, recovering known planets at the theoretical photon noise and diffraction limit of AMI. I told Louis it would Just Work™ and two years and tens of thousands of lines of code later, he made it work.

Thresholded and composite version of Figure 16 from the paper, showing detection maps for known companions B and C to the star HD 206893. In the full figure, there is a lot more scruff, where low-probability regions are shown so the shape of this sensitivity pattern can be clearly understood.

$Contrast curves for the HD 206893 data. The brighter companion at 200 mas is clearly recovered and the fainter companion at 100 mas is marginally recovered. The empirical curves (lines with shaded regions) closely match the theoretical photon noise limits (dotted lines), and the curve is flat right up to the diffraction limit just inside of 100 mas.$

1 14

Benjamin Pope @benjaminpope.bsky.social · 16h

Now we are able to correct the mask metrology, infer the wavefront, and infer the mask is not *quite* in focus. We also learn the pixel sensitivities and a neural network for how pixels interact with their neighbours - this 'effective detector model' is probably the most impactful part of the paper.

Multi-panel figure showing the calibrated metrology of the mask, and the inferred effects of telescope jitter, aberrations, mask metrology, and Fresnel defocus on the interferograms.

Multipanel figure showing the detector model - intra-pixel response (the map of a single pixel's sensitivity shape), inter-pixel response (the gain of each pixel separately) and nonlinearity, along with a model for how charge bleeds from bright pixels into their neighbours (the brighter-fatter effect).

1 15

Benjamin Pope @benjaminpope.bsky.social · 16h

In a pair of papers in 2023, Steph Sallum (arxiv.org/abs/2310.11499) and our team member Shrish Ray (arxiv.org/abs/2310.11508) showed that AMI seriously underperformed in terms of resolution and sensitivity to faint objects, mainly because of the 'brighter-fatter effect', an electronic systematic.

The JWST Early Release Science Program for Direct Observations of Exoplanetary Systems III: Aperture Masking Interferometric Observations of the star HIP 65426 at 3.8 um

We present aperture masking interferometry (AMI) observations of the star HIP 65426 at $3.8\,\rm{μm}$ as a part of the JWST Direct Imaging Early Release Science (ERS) program obtained using the Near I...

arxiv.org

1 16

Benjamin Pope @benjaminpope.bsky.social · 16h

This is available as an open-source pipeline amigo, and built on our team's Jax differentiable optics code ∂Lux (github.com/louisdesdoig...). Check it out (and use it for your JWST proposals!)

GitHub - LouisDesdoigts/amigo: AMI-GO brrrrrrr

AMI-GO brrrrrrr. Contribute to LouisDesdoigts/amigo development by creating an account on GitHub.

github.com

1 14

Benjamin Pope @benjaminpope.bsky.social · 16h

In the first paper Louis Desdoigts (ex PhD in my group, now Leiden postdoc) learns a model for the entire AMI system's optical physics together with a neural network 'effective detector model' the sensor, which suffers from a serious 'brighter-fatter effect' that blurs images at the pixel level.

AMIGO: a Data-Driven Calibration of the JWST Interferometer

The James Webb Space Telescope (JWST) hosts a non-redundant Aperture Masking Interferometer (AMI) in its Near Infrared Imager and Slitless Spectrograph (NIRISS) instrument, providing the only dedicate...

arxiv.org

1 1 15

Benjamin Pope @benjaminpope.bsky.social · 16h

The Aperture Masking Interferometer (AMI) is a tiny piece of metal in Webb's NIRISS camera, filtering light through a pattern of holes where no pairs are have the same direction and spacing. This means that we can unambiguously tie changes in the telescope images to the optics ('phase retrieval').

The AMI mask, with seven hexagonal holes in a circular piece of metal superimposed with the JWST mirror shape.

1 16

Benjamin Pope @benjaminpope.bsky.social · 16h

Thrilled to have two years' of work out, in a pair of papers led by @gradientrider.bsky.social and @maxecharles.bsky.social.

We've built a data-driven calibration of the James Webb Interferometer to near its fundamental limits for high-res imaging - explainer at @aunz.theconversation.com!

How we sharpened the James Webb telescope’s vision from a million kilometres away

The only Australian hardware on board the legendary telescope is starting to fulfil its duties.

theconversation.com

1 33 120

Reposted by Benjamin Pope

Dr. Jessie Christiansen @aussiastronomer.bsky.social · 1d

Less than a year later, once again feeling anxious for my remaining fabulous JPL colleagues. 😬

Dr. Jessie Christiansen @aussiastronomer.bsky.social · Nov 12

Feeling anxious for my many fabulous JPL colleagues tomorrow with another round of layoffs.

www.jpl.nasa.gov/news/jpl-wor...

JPL Workforce Update

Workforce statement and memo to employees

www.jpl.nasa.gov

6 4 55

Reposted by Benjamin Pope

Daniel Knowles @dlknowles.bsky.social · 1d

Honestly almost nobody criticising the use of the "passive voice" has any idea what it means. "The man died after an interaction with federal agents" is active voice. "The man was murdered by federal agents" is passive voice.

jamelle @jamellebouie.net · 1d

underrated part of this is that stoller very clearly has no idea what "the passive voice" is

Ghostin this Hell @hoovybaby.bsky.social · 1d

no quarter given lmao

6 11 97