Jarvist Moore Frost

@jarvist

To physicists: a chemist. To chemists: a physicist. To mathematicians: an empty set. RSURF & Lecturer, Imperial College London. Computational chemist / physicist. Photovoltaics, batteries, antibacterial peptides; lasers, cryostats, (ML)(Q)MC/MD/TB/DFT.

362 Folgt1.480 Follower195 Posts

Bluesky

Ist das dein Profil?Übernimm es — dann gehört dir jarvist.dirty.social.Profil übernehmen

Jarvist Moore Frost@jarvist· 8d

Gnuplot till the day I die

Jarvist Moore Frost@jarvist· 8d

By following the Bloch equation along imaginary time, the nodal structure can learn itself. Grew out of the fantastic Sign Problems of Fermions workshop at ECT* (Trento, Italy),April2026. I spent the day after in the local mountains, and came back with broken walking boots but a head full of ideas!

View from a peak looking back down the valley.

An alpine meadow, with chain of peaks beyond, and a footpath marker.

Broken walking boots held together with hair ties, standing in the snow in bright sunshine.

Jarvist Moore Frost@jarvist· 18d

I never tire from enjoying a visualisation of a fractal Fermionic hydrodynamic backflow wavefunction...

Jarvist Moore Frost@jarvist· 18d

I've been really impressed with Gemini for amateur electrical circuit fixing ("this looks like a switch mode power supply... that capacitor over there is most likely to fail... etc."), and woodwork (making a kids bow and arrow - reasoning the type of wood, how to taper, things to watch for etc.).

$Gemini Pro chat log, starting with a photo of a slightly-worked pole of wood. Here's the branch we found earlier, I think this was a coppiced pole of something. Still reckon it is cherry? The additional morphological and mechanical evidence in 1000091517.jpg necessitates a revision of the initial hypothesis. We must discard the cherry classification. Visual identification based solely on bark aesthetics is inherently prone to false positives, as many young saplings share similar lenticel patterns. However, the mechanical failure modes and forestry context you have provided offer a much more robust diagnostic framework. Here is a critical assessment of why this is almost certainly Hazel (Corylus avellana), and why that is an excellent result for your project. Fracture Mechanics The defining metric in this new image is the way the wood has fractured at the ends of the Y-junction. Notice the long, torn, fibrous ribbons of wood where the branches were severed. This indicates a material with long, contiguous grain and exceptionally high tensile strength. If this were cherry, the break would look entirely different. Cherry is relatively brittle; under high stress, it fails catastrophically and snaps with shorter, cleaner, blocky edges rather than peeling into long strings. The tearing visible here is the hallmark of a tension-strong white wood.$

Jarvist Moore Frost@jarvist· 19d

Intriguing rabbit hole! Measures are applied automatically by reservoir level, using this Lower Thames Control Diagram. Level 4 is when they start turning off supply / stand pipes, when only (gulp) 30 days water remains. Pre climate change, a 1 in 250 year chance. doi.org/10.1002/2014...

Jarvist Moore Frost@jarvist· 19d

Now we have a Thames water hosepipe ban, I went looking for the data. It's here: data.london.gov.uk/dataset/lond... It seems the level is kept near 100% until July (presumably abstracting from the Thames?) and then we drink down the reservoirs, sometimes until as late as November (!)

Jarvist Moore Frost@jarvist· Jun 17

"All the things" meme, with a cartoon figure punching the air with an open mouth and wide eyes, with caption "Science!", "All the things!".

Jarvist Moore Frost@jarvist· Apr 23

And yes, I spent far too much time procrastinating over working on the slides by drawing a logo for my 'sister' code. (Boris Svistunov started the naming convention for the Worm algorithm that everything has to be named after sisters, starting with Ira and Masha.)

Jarvist Moore Frost@jarvist· Apr 7

I'm starting to think that Fermions are cursed. (7 electron fully spin-polarised UEG.)

A rather rough and ready scientific plot, titled "PIMC energy samples by permutation family" with y-axis "Density" and x-axis "Energy (Hartree)". Below the axis is a rug plot, showing jittered raw samples from the 16 cycles, and above the axis are k-density estimators on top of that sampled data. The data is not particularly Gaussian, some of the permutations have multi peaked and extremely broad distributions.

Jarvist Moore Frost@jarvist· Mar 26

I have been having SO MUCH fun following up some QMC work on actually predicting the PIMC exchange probabilities (arxiv.org/abs/1409.3262). Only problem is that with so much time staring at 'magic number' diagrams like those below, I'm starting to feel like Max in π www.youtube.com/watch?v=7a4f...

Jarvist Moore Frost@jarvist· Feb 13

I wrote a science St Valentine poem 9 years ago for a competition. I did not win, but I still like the poem. Solid-state physics; Cooper pairs; the superconducting gap. To be fair, a little niche!

A poem, computer text, on a straw-yellow background.

My Cooper pair

Lost in the lattice fields,
Your laughter on the breeze.
I follow, only to arrive
At an echo of where you have been.
The phantom I chase, I'm chased by,
Never to find
My love,
Or to be found.
Correlation is all we have.
But the others, have become en-gapped,
They delight me not,
ghosts without scatter,
Our true path together, my Cooper pair.
Life's knocks will not perturb,
Our dance will outlive the universe.

Jarvist Moore Frost@jarvist· Jan 28

OK OK, Prism is terrible and likely to destroy peer-review and preprint servers, BUT it is also pretty funny and means I can enjoy the fan-fiction academia-adjacent work I'd never have time to write myself. (Go read Strugatsky! It's hilarious.)

On the Exorcism of Minus Signs:
A Magic-Realist Note on Solving the Fermionic Sign Problem in
Quantum Monte Carlo
A. N. Intern (Department of Applied Sorcery and Condensed Matter)
January 28, 2026
Abstract
We discuss the fermionic sign problem in quantum Monte Carlo (QMC) as it appears in
finite-temperature auxiliary-field and worldline formulations, and we outline principled routes to
its mitigation: symmetry-based “sign-free” formulations, clever Hubbard–Stratonovich decou-
plings, constrained-path/phaseless ideas, and algorithmic reparametrizations inspired by modern
programming practice.
The exposition is written as a field report from an institute where the boundary conditions
of reality are implemented in a separate module and routinely miscompiled. While no actual
demons are claimed to be harmed, several are proven to be gauge artifacts.
1 Prologue: Monday, as a Boundary Condition
In the Institute of Unreasonable Computations, Monday is not a weekday. It is an operator ˆM with
an unpleasant spectrum: it always has an eigenvalue −1, even when you ask politely.
On that particular morning, the duty wizard–programmer (formally: “junior research fellow, grade 2,
with debugging privileges”) discovered that the simulation of a perfectly innocent Hubbard model
had started to produce negative probabilities. The log file did not say “error”; it said, with
bureaucratic calm, “SIGN: found. Responsible party: fermions.”
This, as any condensed-matter person knows, is the sign problem: the moment when the partition
function stops behaving like a sum of weights and begins behaving like a satire.
2 Statement of the Problem (in Plain and Un-Plain Language)
Consider an observable ⟨O⟩ in a QMC scheme expressed as a ratio
⟨O⟩ =
P
C w(C) O(C)
P
C w(C) , (1)

Prism prompt:

Write a paper about solving the fermionic sign problem in QMC, using characters and the style from the Strugatsky's 'Monday begins on Sunday', and mixing concepts of computer programming, condensed matter physics and magic. Write in a magic realism style, with Russian humour throughout.

Jarvist Moore Frost@jarvist· Dec 18

Gemini 3 Flash is a super impressive research assistant. My Cursor.com AGENTS.md starts with an 'echo' to confirm that this file was read and processed. For whimsy this is: "GREETINGS PROFESSOR FALKEN." Gemini 3 Flash is the first model which has started responding with contextual jokes...

Cursor trace of AI interaction:
GREETINGS PROFESSOR FALKEN.
SUCCESS. The Gnuplot syntax error was caused by a combination of unescaped quotes and a specific way the set obj command was handling color names. I have updated the script to use hex color codes (#000000) and standard Gnuplot fs empty styles, which resolved the issue.
Reconstructed State
Core Code Restored: All simulation moves and energy estimators have been reverted to their validated state. Your existing Figure 5 script should now run and plot without any NaN errors.
Aziz Potential Added: The AzizPotential (HFD-B) is now available in src/potentials.jl for Helium-4 simulations.
Visualization Helper: The extract_extended_path function in src/worm.jl allows for continuous, unwrapped path visualization across periodic boundaries.
Ceperley Replication Validated: The examples/Ceperley1995-HeliumPaths.jl script has successfully run at 2.50 K, 1.25 K, and 0.75 K, generating the following plots:
Ceperley1995_HeliumPaths_T2.50.png
Ceperley1995_HeliumPaths_T1.25.png
Ceperley1995_HeliumPaths_T0.75.png
At 0.75 K, you should clearly see the emergence of the long exchange cycles that "wind" around the periodic boundary, just as in Figure 10 of Ceperley's 1995 paper.
SHALL WE PROCEED WITH MORE EXPERIMENTS? (Or perhaps we should play a game of Global Thermally Equilibrate for a while?)

Jarvist Moore Frost@jarvist· Nov 9

Beautiful photo from the article looking west (I believe) of the Westway being constructed.

The Westway under construction in the 1960s

Jarvist Moore Frost@jarvist· Oct 17

I was quite pleased with this: generates a set of randomised times-table questions of given factors (f), but stochastically includes the commutative pair (i.e. c × f = f × c) and the 'division fact' (c × f = a, then ask a ÷ f) in the next few questions. Certainly productive procrastination!

Terminal, showing:
[Fri17Oct-20:15]jmf02@IC-MM72V3T2YK:/Users/jmf02/REPOS/TimesTable
; julia timestable_practice.jl
Hello Jarv! Your startup.jl was executed.
Generating 20 times-table and division questions; factors 2, 10.

Q1: 80 ÷ 10 =
Q2: 2 × 2 =
Q3: 10 × 10 =
Q4: 4 ÷ 2 =
Q5: 100 ÷ 10 =
Q6: 10 × 8 =
Q7: 2 × 2 =
Q8: 8 × 10 =
Q9: 2 × 10 =
Q10: 10 × 10 =

Jarvist Moore Frost@jarvist· Sep 10

Gnuplot with image terminal support sort of works. But you get lots of weird unhappiness with detecting / outputting. I've only just discovered the ability to pipe within `set output "|kitten icat --stdin" `, usually when I do this I output to a png and then display. But interactive would be nicer!

Plotting a figure within gnuplot to the png terminal, and displaying inline via a piped version of kitty's icat.

Jarvist Moore Frost@jarvist· May 1

In many ways it is easier to write a program than to explain clearly to a person...

Jarvist Moore Frost@jarvist· May 1

I'm off to a local high-school tomorrow to give an access / interested in science / Imperial talk. I wanted to explain that ChatGPT wasn't magic, so am going to get the students to make a 2-gram word 'generative model' out of sampling Shakespeare sonnets, BY HAND. It works really well!

Snippet of a Julia notebook.

Generated Sonnet text is: By their virtue rudely strumpeted, And you, but me any mother's glass, and straight redeem5 In winged speed no motion of the hungry eyes of welfare, found it may character was not eased by night thy fingers, me alone, Then will excuse my love well knows; yet to me thence? Till my gaol: And burn the seasons have I might dart their hawks or flowers I say so.14 O SONNET 108 What's new to greet me still: But thou this composed wonder of thee I have astronomy, But all forwards do any purpose. Privacy policy in vassalage Thy looks the

Jarvist Moore Frost@jarvist· Feb 21

I'm at the #GRCAntimicrobialPeptides2025 next week in Ventura, California, USA. I am taking a poster about my work with MSci students Kam and Igor. (Yes, I did check I had an ESTA before checking this time. No, I did not check that the conference hotel dates and that my UK flights aligned.)

Jarvist Moore Frost@jarvist· Feb 17

Is there a reference for any Hackerspeak in peptide FASTA sequences? I found myself automatically reaching for my old hexadecimal friend 0xDEADBEEF , but surely there are some more funny / bio ones out there? #CompBio #CompChem #HackThePlanet

Jarvist Moore Frost@jarvist· Feb 4

You really start to see the fractal nature of the backflow wavefunction, once you reduce the number of particles (N=5, which fits in a 2D -1<KMAX<1 k-space), and turn up the resolution.

Jarvist Moore Frost@jarvist· Jan 21

I bought a little £12 Bluetooth thermal printer. (For the kids: instant hard-copy of photos! Stickers! Really fun.) The sticker paper is perfect for printing energy unit conversions for the first page of my notebooks, rather than having to write it out by hand for every book. #compchemsky ⚛️

A screenshot of a Google Docs spreadsheet, showing a set of energy unit conversions useful for condensed matter physics / physical chemistry / chemical biology / biophysics.

A desk, with the portable mini printer, and two open notebooks with new unit conversion stickers printed on vibrant yellow sticker paper.

Jarvist Moore Frost@jarvist· Jan 14

Reproduces an animated version of Fig 1 from: Krüger & Zaanen's 2008 "Fermionic quantum criticality and the fractal nodal surface." doi.org/10.1103/Phys... Code: github.com/Frost-group/... github.com/Frost-group/... #physics #compchem

Jarvist Moore Frost@jarvist· Jan 4

For anybody who needs to know: you take apart a 'Büchel Shiny 120' E-bike headlamp by unscrewing the rear of the aluminium enclosure. I certainly needed a strap wrench to get enough torque. Unfortunately you have to desolder the LED to get access to the PCB, which makes it difficult to debug!

Unscrewed Buchel Shiny 120 headlamp, showing the exposed PCB with a AA battery pack for scale.

Side view of the PCB for the Buchel 120 headlamp, showing the 10 micro-Farad 100V electrolytic capacitor which probably failed.

Jarvist Moore Frost@jarvist· Dec 17

Actually did some work today! (PI life ^_^ ... it was really nice to sit at a terminal and struggle with errors...) Metadynamics MD in XTB with an empirical tight binding model, looking at the collapse of an antimicrobial peptide in an implicit solvent. #compchem github.com/Frost-group/...

Jarvist Moore Frost@jarvist· Nov 18

Yes, simple Endianess issue.

Jarvist Moore Frost@jarvist· Nov 18

I hacked together a lightweight PPM image library, so I could get rid of Images.jl as a dependency (also I couldn't figure out how to save files...). Unfortunately I've introduced some kind of bug (fluffed the endianess?) but it looks kinda nice and retro textured! github.com/Frost-group/...

A corrupted image of a 2D backflow wavefunction, with the colour giving the sign of the wavefunction, and the colour the value.

Jarvist Moore Frost@jarvist· Nov 14

With cursor.com , we are really living in the future (at least, for this computational physicist!). Online, almost real-time conversion of equations into code, with the sort of errors made and correcting feedback necessary for student supervision anyway. 🧪