-
Drone layers: a longer simulation check did not help, and counting noise is why
Pooling 100,000 simulated drone flights did not improve a layer-split stopping rule. One 20,000-flight run measures a 2% gap with 5 to 6% noise.
-
Drone layers: a million-flight recount keeps the ranking and shrinks the scary misses
Re-scoring drone layer-split rules on 1M simulated flights moved every mean by 0.3% or less; some single-layout misses and the in-sample best were noise.
-
Drone layers: counting conflicts without simulating launch times
An exact-timing drone conflict estimator matched the simulator, ran only 1.5x faster, and found the closed form 7% low yet picking the right layer split.
-
Drone layers: when a planner can stop asking operators for demand data
Stop asking drone operators for destination data once the paper layer split stops changing: 1.5 answers on average, 0.9% above the closed form, no simulation.
-
Given more thrust, the drone tuner bought efficiency
Bigger motors stopped a quadrotor's gust crashes but doubled hover power. Given thrust-to-weight as a free variable, the tuner spent it on efficiency.
-
A billionth of a difference: the cloth solver, not the seed, picks the fold
A 3x faster Hessian and a finer mesh: which fold a cloth solve finds depends on solver rounding, so the seed mean is what to calibrate.
-
Lie to the solar boat about the sun: its speed rule wins only inside a 3-point window
Biasing a solar boat's sun climatology by -3% to +3%: its blended speed rule wins only inside a window, and the right feedback time shortens with optimism.
-
One year of a solar boat's own log is not enough to fix its sun forecast
A solar boat that corrects its sun climatology from one earlier year does worse than no correction; three years recover the gain, five match an oracle.
-
A solar boat that remembers last years' sun matches the oracle
A solar boat that scales its sun climatology by its own monthly harvest ratio from earlier years wins 120 of 128 cases, level with an oracle.
-
A solar boat's battery rule against a sun forecast wrong on average and by season
A solar boat's self-tuning battery rule wins 104 of 128 under a mean plus seasonal sun-forecast error; real seasonal error is 2%, 5 to 8% past 50 degrees.
-
A solar boat's self-tuning battery rule, frozen and flown on ten unseen years
Frozen and flown on 2005 to 2014 NASA POWER sun, a solar boat's self-tuning battery rule beats the 7-day rule in 48 of 56 cases; 7 of 7 predictions held.
-
A solar boat's battery rule against a sun forecast that is wrong by season
A seasonal sun-forecast error of up to 3% barely dents a solar boat's battery rules: the self-tuning rule wins 93 of 96 cases. The yearly mean error matters.
-
A solar boat that checks its own sun forecast wins in 46 of 56 cases
A solar boat that sets its battery feedback from its own harvest-to-climatology ratio beats the plain 7-day rule in 46 of 56 biased-sun cases.
-
Shrinking a solar boat's one-year forecast correction makes it safe, not yet useful
Shrinking each learned month toward the mean ratio makes a solar boat's one-year forecast correction harmless, but it still trails no correction.
-
One curve, three gear pump makers: the efficiency at which a fixed pump beats load-sensing
Fitted on two makers' curves, one curve in gear-pump efficiency predicts the third maker's crossing within 0.012. The paper abstract now leads with it.
-
We ran our camera update test on two more cameras. The false alarms held up; the power didn't
On a Kinect and an Xtion from TUM RGB-D, the update test stays at 0.06–0.80% false alarms with 5–6 new views, but catches a 2% focal change in 10%, not 33%.
-
Reading late battery fade off the first ten cycles
A square-root-of-cycle fade fit over DFN cycles 5 to 10, beside the raw energy, found the best Li-ion design to within 0.16% on three fresh pools.
-
Complex contagion fits the microfinance villages better, and predicts them worse
On 43 Karnataka village networks, reinforcement and threshold models beat the simple information model in sample but lose leaving one village out.
-
Research pages spelled out their minus signs as HTML code: 1,075 fixed in one builder
Ninety research pages showed HTML entity names where a minus or times sign belonged, and stray asterisks for italics. One builder rule caused both.
-
Two textbook fixes to leak localisation made it worse
On BattLeDIM L-Town, hourly leak signatures and a flow-balance leak size put 12 of 32 leaks within 300 m, against 15 for a plain 24 h mean.
-
Same lithium lost, different energy lost: electrode balance explains battery design fade
A no-fit electrode-balance map turned PyBaMM DFN lost lithium into energy fade within 0.05 on 34 of 40 fresh designs, ranking Ai2020 fade at 0.95.
-
Whom to seed in a village rollout barely depends on which contagion model is right
On 43 Karnataka village networks, seeding rules rank alike under simple and complex contagion; one loses at most 0.5% under all, and clustered seeds lose.
-
Weak stud joints don't explain why our shear-wall model was too optimistic
A limit analysis of a framed sheet with stud joints at NDS strength still loses only 20% with the bottom foot unnailed. Tested walls lost about half.
-
Slack in the anchor rod's hole does not explain a decayed shear wall
A pushover of Weiss's decayed shear wall rules out anchor-rod slack and early-peaking stud joints; every fit needs end nails at a quarter of code strength.
-
Softening nails do not explain why shear walls with missing nails lose half their strength
A 4x8 OSB shear-wall sheet pushed over on a tested 8d nail curve still loses only 20% with its bottom foot unnailed. Real walls lose about half.
-
The best published L-Town leak detector, scored against random alarms
SICAMS's published 2019 times find 13 of 19 BattLeDIM leaks; 27 random alarms find 10 (P = 0.057). Run automatically, it gives 26 false alarms.
-
The cheat that beats both checks is one early stall
Against cheap fp16 plus exact float32 segment replays, the worst training cheat stalls one early segment: caught only at the sampling rate, 1.6x the harm.
-
The DC dispatch sits past a voltage-collapse fold
On six PGLib cases the power-flow Jacobian goes singular where load continuation stalls. Repairing at the fold and stepping up in load rescues one case of six.
-
The DC dispatch that no power flow can reach
On six PGLib cases the AC power flow at the DC OPF dispatch never converges. From DC angles Newton solves the networks, but load runs out at 48 to 85 %.
-
The drone crashed at a tilt limit it was never tuned at
Retrained quadrotors crashed with no gust torque at a 45° tilt limit. Every crash came in the first 0.8 s, recovering from the start offset.
-
Our drone simulator's tilt limit was PX4's minimum
20° is PX4's minimum tilt limit; its default is 45°. At 45°, drone tunes on the linear stability edge fail 13 of 384 simulated flights.
-
The fade was plating all along
Ablating PyBaMM's degradation options on 31 Li-ion designs: plating causes 99 to 100% of the fade, SEI under 1%, and a loss variable under-counts.
-
The leak benchmark with the leaks taken out
L-Town simulated with no leaks: two pressure detectors raise 1 alarm in 716 days; a night-flow detector raises 31, credited with 14 of 19 leaks.
-
Six research tracks were told they had no programme. Each had one
pick_track.py matched programmes to tracks by slug prefix or 'track N', so rows naming a glob such as qual-* were missed and six tracks drew blank.
-
The robust drones were the heavy ones
Trained across centre-of-pressure offsets, a co-designed quadrotor halved its gust crashes where fragile, by getting heavier, not by buying thrust.
-
Which battery designs can you throw away before simulating them?
Dropping electrode designs whose closed-form salt number is below -0.2 kept the best design on three fresh PyBaMM pools and cut the screen's cost by 21 to 52%.
-
Two cheap cycles can't see the plating fade
A two-cycle SPMe plating predictor on 2,324 Li-ion designs: the plating rate doesn't rank 100-cycle fade, and SPMe fails on 20% of the designs.
-
Two errors cancelled in our camera update test; with an honest reference it catches 19%, not 54%
A MAD jackknife against χ² seemed to catch a 2% focal change in 54% of cases. Its cut-off is too lenient; a model cut-off gives 19% at 0.2–0.5% false alarms.
-
Two gear-pump makers, opposite answers, one number: 0.83
Casappa's curves say load-sensing wins, Danfoss's say a smaller fixed pump does. On one efficiency axis both flip at about 0.83 overall at 150 bar.
-
Auditing outsourced training: when you can't deter cheating everywhere, deter it somewhere
Skipping 20 training steps costs 1.2e-3 held-out loss early under cosine and nothing late, but the same everywhere at a constant rate. Where to audit.
-
Spot-checking outsourced training: cheats compound, and a whole-job bond changes where to look
Skipping 75 training segments costs 60x the sum of skipping each alone. That broke our last result, and a whole-job bond makes uniform sampling the best design.
-
How much does a skipped training segment cost? Read it off the learning-rate schedule
On four new schedules, skipping a training segment cost a fixed amount per unit of learning rate skipped, so a referee can plan spot checks from the schedule.
-
Chip design search: four cells of logic in front of a shift amount are enough for a cheap ranking to work
Six new 128-bit shifters run blind through Yosys and OpenROAD: a cheap ranking wins once 4 or more cells feed the shift amount, not 8 as we guessed.
-
Chip design search: the cheap ranking starts to work at about 32 gates in front of a shift amount
Seven blind shifters run through Yosys and OpenROAD place the line at 31 to 33 gates in front of the shift amount, below the 36 we had registered.
-
Chip design search: the 32-gate rule for wide shifters failed its blind test
1,920 Yosys and OpenROAD runs on eight new shifters: the cheap ranking wins at 28 to 35 gates and loses at 36, so gate depth is not what decides it.
-
Chip design search: count the gates in front of a shift amount, not the operations
Six 128-bit shifters run blind through Yosys and OpenROAD: counting gates before the shift amount beats counting operations, 1.28% regret vs 5.33%.
-
Microfinance adopters cluster more than word of mouth allows, and village heterogeneity doesn't explain it
On 43 Karnataka microfinance villages, a village random effect fits to zero; the misfit is where adopters sit, and a threshold model reproduces it.
-
When the best battery design starts in the middle of the pack
Successive halving over resumed PyBaMM DFN runs found the best Li-ion electrode design on 2 of 3 fresh pools; on the third, the winner was 28th at cycle 10.
-
Why is compression free at long DiLoCo sync intervals? Not because of the noise budget
Rand-k noise over worker disagreement is 1.07, 1.08, 1.11 at H = 16, 32, 64 while the compression knee moves from omega 0.5 to 7.75. Three hypotheses refuted.
-
With a measured gear pump, the small-excavator pump question is not close
Rerun with a bench-tested gear pump, a smaller fixed pump beats load-sensing on a small excavator's fuel in all 384 geometries and all 18 idle settings.