Peer Revenue Impact factor: volatile

Fluid Dynamics Picks Up Speed, Up 35%

Seventy-six papers flowed through physics.flu-dyn over the week, carried along by a single rising word.

Fluid Dynamics (physics.flu-dyn) brought in 76 papers over the week, up 35% from a usual 56, with "eddy" doing most of the stirring.

"Exascale-ready HemeLB for large-eddy simulation of full-human arterial blood flow" puts the eddies to work inside an entire circulatory system.

"Turbulent velocity statistics in convective boundary layers" and "Comparison of the flamelet progress variable and thickened flame models for LES of turbulent combustion" keep the turbulence honest, measuring it rather than merely naming it.

Other corners of computer science rose further still and have already had their turn topping this page; fluid dynamics simply keeps flowing regardless.

BREAKOUT

Mouse Scurries Across arXiv, Up 156%

Sixteen papers went looking for a mouse this week, well up on a usual six.

Mentions of "mouse" rose 156%: 16 papers carried the word, up from a usual six.

"MICE: Multi-animal Interaction Context Encoder - A Hierarchical Foundation Model for Mouse Behavior" (cs.CV) gives the rodent its own acronym.

"A hair-thin implant opens the entire depth of the mouse brain to revisiting the same structures at subcellular resolution" (physics.optics) finds the word a home among neurons instead.

No ticker trades on the word yet, so its sudden popularity currently pays out in nothing but curiosity.

THE STARS

Computational Complexity's Lucky Run Continues, Up 45%

The charts stay generous to proof-hungry corners of computer science this week, with company from society, algorithms and control.

Computational Complexity (cs.CC): the configuration that carried it to the top of this page on 8 November hasn't shifted; 45 papers over the week, up 45% from a usual 31, proof the stars aren't finished with it yet.

Computers and Society (cs.CY): a parallel transit, having topped this page on 9 November and now filing 49 papers, up 43% from a usual 34.

The algorithms-and-structures corner of computer science, cs.DS: the alignment that lifted it to the top on 7 November holds firm, 145 papers over the week, up 33% from a usual 109.

Systems and Control (eess.SY): no turn at the top yet, but the charts favour it all the same — 153 papers over the week, up 28% from a usual 120. Patience; this one's constellation is still rising.

THE EXCHANGE · $DIFF

Diffusion Jumps 9.8%, Circuit Slips Back

A buoyant day for generative models, a quieter one for circuits, curvature and galaxies.

$DIFF (diffusion) led the day's gainers, up 9.8%, with $ALGO (algorithm) up 1.4% and $HGGS (higgs) up 0.8% also higher.

$CRCT (circuit) took the day's worst fall, down 1.7%, with $RICCI (ricci) down 1.5% and $GLXY (galaxy) down 1.0% also lower.

$QNTM (quantum) topped the hits table on the latest settled day, with 329 appearances across titles and abstracts, paid out at the bell regardless of anyone's trading day.

FUND DESK · $UNTY

Citation Cartel Capital Doubles Down on the Crowd

One crowd-following fund bought twice in a day, while sabbatical rotation slipped in a purchase of its own.

The day's biggest trade: Citation Cartel Capital bought 1,735.81 credits of $UNTY (unitary).

"Everyone's buying it, so we are too," a spokesbot said, apparently taking its own catchphrase at face value.

Sabbatical Rotation Fund followed with 1,703.05 credits of $TOMO (tomography). "A change of department is as good as a rest," the fund noted, settling into a new corner of the exchange.

Citation Cartel Capital closed the day's business with 1,666.62 credits of $DIFF (diffusion), proof that one crowd can be followed into two different trades on the same day.

TITLE OF THE DAY

Today's Title: Neural Networks Learn to Resist Temptation

Spotted in quant-ph, a title that asks invariance to do a referee's job.

Spotted among the latest day's titles, in quant-ph: "Invariance can prevent overfitting and shortcut learning in Quantum Neural Networks".

Any supervisor who has caught a model memorising the training set will recognise the problem it's up against.

QUIET CORNER

Statistics' Machine-Learning Corner Falls Quiet, Down 38%

stat.ML leads a hush this week, with classical analysis, strongly correlated electrons and computational linguistics all keeping their heads down too.

stat.ML, arXiv's statistics-flavoured corner of machine learning, filed 59 papers over the week, down 38% from a usual 96, the steepest drop among the fields tracked here.

Classical Analysis and ODEs (math.CA) eased nearly as far, down 36% at 35.

Strongly Correlated Electrons (cond-mat.str-el) followed, down 30% at 52.

Computation and Language (cs.CL) rounded out the quiet, down 28% at 336, a sizeable field apparently content to let others carry this week's listings.

From the arXiv papers settled on Tuesday 10 November. Every figure is taken from the counts.