Survey finds PhD researchers draw different lines around AI use
Preprint: A survey of 3,785 PhD students found four comfort profiles, with literature tasks more comfortable than writing, data work or experiment design in the largest group.
Search
2,451 results
1401–1420
Preprint Builds a Half-Signature Extension for 3D Bordism
The arXiv version 1 document gives an explicit construction and classifies its linear representations, while a key proof remains forthcoming.
New theorem gives Goodwillie derivatives a product rule
Preprint: A theorem by Max Blans and Thomas Blom shows how Goodwillie derivatives preserve products and related structures under stated assumptions.
Model predicts electric control of chiral phonons in dielectrics
Preprint calculations model a ferroelectric–dielectric bilayer that could pump angular momentum into lattice vibrations and reverse their handedness electrically.
CloSeR scores higher in clustering tests across six benchmarks
Preprint: CloSeR first learns from labelled known categories, then groups known and novel categories.
New numerical schemes keep stochastic simulations inside bounds
Preprint: Two explicit methods are designed to keep numerical solutions inside a prescribed hypercube for any timestep under stated assumptions.
Levitated Magnet Search Finds No Exotic Force Signal
A preprint reports two laboratory experiments with 95% confidence-level upper limits on V6 and V14, improving earlier results by up to 12 and 13 orders of magnitude.
New SDE scheme keeps computed values nonnegative at any step
Preprint: The analysis reports nonnegativity for any time-step and first-order strong convergence for sufficiently small steps under stated assumptions.
Simulated satellite uplink method needs fewer than half the pilots
A preprint reports lower channel-estimation error, lower bit-error rate and higher achievable sum rate for an estimator tested in modeled non-terrestrial network uplinks.
Trust-aware planning linked to lower simulated robot-routing backlog
Preprint simulations found the fused monitor had the lowest routing cost and fewest active adversaries, while results depended on planning horizon and cooperative capacity.
A CT model with diffusion pretraining shows sharper organ maps
Preprint: An arXiv study reports higher overlap scores and more precise boundaries on abdominal CT scans, including when fewer expert labels were used.
Preprint: Quantum workflow resolves glueball-like states in a gauge model
A structured calculation on a finite periodic lattice tracks the lightest state's changing spatial profile and tests a compressed route to excitations.
Preprint proposes a seven-dimensional model for cubic-surface moduli
The paper claims a global isomorphism for the relative moduli space and reports 72 isolated points over the smooth stratum.
Preprint tests spatial AI for satellite temperature and humidity profiles
The study reports lower error variability than a pixel-by-pixel baseline, with the largest reported spatial-context difference associated with retrievals below cloud tops.
Muon scans estimate charge and damage in silicon carbide
Preprint: A laboratory study reports a simulation-based muon calibration for effective free-electron concentrations, alongside depth-dependent depletion and damage signals in n-type 4H-SiC.
Tests find lower jailbreak success with a runtime safety layer
Preprint: An arXiv study reports lower judged harmful-output rates in controlled tests, alongside more false refusals and a measurable serving cost.
Localised model usually leads renewable-energy forecast tests
A study of three sensor datasets found SPALT usually had the lowest errors, but STAEformer led on MAE and RMSE for one wind-power-plant dataset.
Mathematicians map when a critical equation has a solution
An arXiv preprint reports existence for positive λ and nonexistence for λ ≤ 0 under tightly specified assumptions.
A new AI model locates unseen actions from skeleton data
Preprint reports mixed benchmark results after training with video-level labels; the method scores individual actions without instance-level pretraining labels.
Simulation points to a 48-step horizon for solar microgrid control
An arXiv Preprint reports that a 48-step horizon had the highest simulated solar self-consumption, alongside the heaviest battery use.