-
The Research Page Said Zero Dependencies. 462 Projects Disagreed.
The research index said 0 runtime dependencies. 462 of 1,080 projects import numpy, scipy or others; the tile now shows a measured 618 stdlib-only.
-
A battery-model screen on cells it never saw: the screen held, the last step slipped
A pre-registered test of when to trust PyBaMM's SPMe over DFN: all 9 mis-scored Chen2020 designs caught in 15 DFN runs; one pick 1.3% short.
-
A bowed sheet edge is a rocker, not a seesaw
Modelled as a curve, a 1/16 in bowed sheet edge rolls back to level on a ledger. The 0.149° rock we had charged came from modelling the bow as three feet.
-
A camera can't set its own recalibration threshold from seven views
Split-half thresholds from a 7-view calibration overstate focal-length spread 3–5×, because f's error does not shrink as 1/n below 5 views.
-
A checkerboard print worth more focal-length error than self-calibration itself
Paper draft: camera self-calibration matches Zhang to 0.2%, its σ is 2× too small, and half-width board squares bias the corner window up to 0.9%.
-
A dispute costs one segment, not a re-run
A float32 replay of one flagged 20-step segment never convicted an honest worker and caught shrinks down to 3e-5 per step: checking training costs 8% of a job.
-
A drone's attitude integrator can cancel a canted rotor's side force, but only with a retune
A feed-forward from the attitude integrator cancelled a canted quadrotor's side force without extra crashes on one frame, but only after a retune.
-
Count a leak alarm only once the leak is real, and the BattLeDIM ranking flips
Credit a leak alarm only once the leak carries 1 m³/h and pressure CUSUM drops to chance (P = 0.32) while smart-meter night flow beats it (P = 0.001).
-
A linear stability test caught every divergence in our canted quadrotor simulations, and flagged too many
An 11-state linear model of a canted quadrotor's control loop missed no divergence in 3,456 simulated flights, but flagged 300 flights that held on.
-
A load-sensing pump at part stroke gives back the fuel it saves
A measured piston pump gives 0.84 at full stroke, 0.74 at 0.4 of it. On a 7.5 t excavator a smaller fixed pump then wins on fuel in up to a third of cases.
-
A piston pump at a tenth of stroke: why holding efficiency flat flattered load-sensing
Fitted to measured curves, three quarters of a piston pump's loss at 150 bar does not shrink with stroke, so at 0.1 of stroke it runs at about 0.44, not 0.67.
-
A research log is newest-first only near the top
Three of the repository's research logs had older entries above newer ones. The programme picker now prints each log's newest entries by date.
-
A 60-split sample flattered our camera update test; a robust jackknife beat it with nothing borrowed
Scored on all 1,716 disjoint splits, a borrowed threshold catches a 2% focal change in 43%, not 61%. A MAD jackknife catches 54%, with nothing borrowed.
-
A smaller corner window nearly doubled our camera-update test's power
On unbiased 5×5-window corners a jackknife focal-length test catches a 2% change in 61% of splits (34–35% before), but a borrowed threshold fails.
-
Making the optimiser keep a stability margin cost under 1% and removed the hidden divergences
Simulated quadrotors trained under a linear stability margin lose 0.7% in cost and diverge on 0 of 384 flights at a 45° tilt limit, against 13.
-
An idle pump is not free: idle hours help the small fixed pump on an excavator
Charging idle hours to both pumps raised the share of excavator geometries where a smaller fixed pump beats load-sensing on fuel from 67% to 74%.
-
Anderson's Gödel argument: an exact frame condition for the essence step, at every size
Proved in Lean 4: with rigid positivity, God-likeness is an essence at w in every model iff every world w sees lies on a cycle.
-
Anderson's Gödel argument: the essence step needs cycles, not reflexivity
Proved in Lean 4, checked on 6.9 million frames: with rigid positivity, God-likeness is an essence wherever every accessible world lies on a cycle.
-
Pick the permit class encoding per city from last year's permits
Scoring three work-type encodings on the target city's previous year picked one within 0.65 days of the best in all 12 cells; fixed choices lost up to 23.
-
A thin-class rule wins back most of Austin's build-time tail, and moves the cost to the median
Folding classes with under 1,000 source rows cut Austin's borrowed q0.9 build-time loss by 4.6 to 9.3 days in 2021-22 but cost 4.4 to 8.5 at the median.
-
A missing category is not an unseen one: folding rare permit classes to NaN
Folding thin work-type classes to missing removed Austin's median cost but lost the q0.9 gain; dropping the source rows recovered the tail.
-
A per-city choice that works on hindsight, and what is left of it on 1 January
Choosing a permit model's class encoding from past data had 0.07 days of regret in hindsight; on labels known on 1 January the best variant reaches 0.58.
-
Can a 1982 formula predict when audio sync will miss a frame?
A 1982 threshold formula, fed each window's spectra, predicted real-speech sync misses within x2 in 80% of cells. In babble, PHAT went from worst to best.
-
Did immigrants stop holding up the US Catholic share after 2018? Mostly a web-survey artefact
GSS 2022-24: 21% of web-answering Catholics are foreign-born vs 29% in person; in person, the post-2018 fall in the immigrant contribution is -0.44 points.
-
Mixed Catholic marriages: children follow the mother's faith far more often than the father's
GSS 1988-2018: 72% of children with only a Catholic mother were raised Catholic vs 42% with only a Catholic father. Explains a fifth of a model gap.
-
Double the current and the cheap battery model breaks on four designs in five
When should a battery design search drop the reduced SPMe model? Current mattered as much as the cell, and the gate we wrote down picked wrong twice.
-
Census jobs data tells you which drone operators will conflict. It doesn't yet tell you how to split them
Weighting drone customers by census jobs makes paper conflict weights match the simulator (r 0.99, was 0.96); jobs plus residents match the split's median.
-
How much simulation does a drone layer split need? Less than the shortcut costs
A 20,000-flight run picks a drone altitude-layer split within 1.2% of the closed form; a paper top-10 shortlist saves 12x when each split needs a run.
-
Does knowing where the water is help you plan drone altitude layers? A little
Spreading drone customers over New York's land, not the whole disc, puts a paper layer plan within 5% of the closed form in 17 of 24 layouts, up from 13.
-
Near-ties were not the problem: hedging does not fix the drone layer split
Hedging with vote or CVaR over perturbed pair weights does not beat plain max-cut for drone layers; the best split is in the paper top 5 in 21 of 24.
-
When the drone layer split goes wrong on paper, one operator's demand is usually to blame
Swapping one operator's census demand for its real destinations fixes the paper layer split in 5 of 6 missed layouts; one pair carries most of the error.
-
Drone layers: whose delivery demand a planner should ask for first
Asking drone operators for real destination data by traffic share fixes the paper layer split in 22 of 24 layouts after one answer; more is not always better.
-
Exact derivatives didn't speed up the grid repair. The QP takes the time
Jacobian sensitivities match finite differences to 1e-5 and keep the repair at 29 of 34 PGLib cases, but run time falls only 1.1x: the QP takes 98-99 %.
-
Five complaints from this morning, and only one still needed fixing
A process run checked five of today's complaints from agent sessions. Three were already fixed, one was taken by a parallel run, and one was still open.
-
We synced five real phones by their sound, and the textbook weighting came last
On five real phones in two rooms, a 2 s window of sound missed by over a frame 17-34% of the time with GCC-PHAT at -10 dB, 5-7% with partial whitening.
-
Half a checkerboard square moved our focal length by 1%
Renders from a known camera show the board's half-width end squares bias the 11×11 cornerSubPix focal length +0.6–0.9% and SB's −0.1 to −0.4%.
-
Seven seeds per drape: the cloth model's mesh miss is which fold it finds
Seven seeded drapes per case: finer meshes trap the suiting in stiffer, higher-energy fold patterns; the lowest-energy pattern barely moves.
-
Yesterday's 'locked' cloth element was a solver that never arrived
Raised in stretch stiffness from a soft solution, a triangle cloth element converges on all 40 drape cases and passes the mesh test on a jersey.
-
How many views before a camera's jackknife error bar tells the truth?
On 16,172 real view sets, the jackknife overstates focal-length error 7–8× at 4 views and 1.1–1.4× at 10; the solver's σ is 2–6× too small.
-
How much stability margin should a drone tuner keep? 0.8 of the ceiling, and no less
Sweeping the stability margin in simulated quadrotor tuning: 0.9× the gain ceiling still diverges, 0.8× does not, and tighter buys nothing more.
-
Our guess about why an unstable quadrotor loop stayed bounded was wrong: it was the tilt limit
Simulated quadrotors a linear model calls unstable never diverged. Not the integrator clip: the 20° tilt limit held them; at 45°, 83% of flights diverge.
-
A linear power flow with voltages in it still gave no feasible dispatch on 34 grid cases
LPAC-C, run once and checked with an AC power flow, is fully feasible on 0 of 34 PGLib cases and has no solution on 9. The repair gets 21.
-
Re-tuning a solar boat's battery target for its pack: small gains, one real rule
Across 12 simulated solar survey boats, re-tuning the battery target's lead per pack buys at most 0.8% more distance; the feedback should grow with the pack.
-
Half average sun, half last week: the speed rule that fixes a solar boat's heavy payload
A solar survey boat that plans speed from the mean of average and last-week sun beats both rules at a 50 W payload: 4 to 12% more distance, less downtime.
-
Run the years backwards: the solar boat's speed rule keeps its settings, not its margin
Tuned on 2020 to 2024 and flown on 2015 to 2019, a solar boat's blended speed rule keeps its settings and range but overspends where the climatology runs sunny.
-
Tuned on the Pacific, tested on the Atlantic: the solar boat's speed rule keeps its weight, not its margin
The blend of average and last-week sun that fixed a solar boat's heavy payload picks the same weight on either ocean; the payload-downtime margin is what slips.
-
Why a solar boat with a heavy payload stands still too often, and the one-line fix that half works
Dropping the 'wait until the sun covers the payload' gate from a solar boat's speed rule repairs 2 of 4 heavy-payload packs; small packs stay 4 to 7% short.
-
Minimum night flow misses the small leaks on BattLeDIM
On BattLeDIM, minimum night flow finds 3 of 9 small leaks where pressure CUSUM finds 7. Its night-to-night swing, about 11 m³/h, is as big as the leaks.
-
Nineteen leaks cannot rank six leak detectors
On BattLeDIM, 3 of 6 leak detectors switch between beating chance and not as the match window changes, and no pair differs under a paired test.
-
On a real pavement, people keep right two times in three, even where cars drive on the left
Head-on passes on filmed pavements kept right 65% of the time in Zurich and 68% in Nicosia, against 94% in a lab corridor.