https://developinglight.com/studies/isometry-invariant-valuations-on-spherical-polytopes-b037bc5c8f Developing Lighten 2026-08-30T05:47:34.530Z New theorem classifies invariant measurements on spherical shapes https://developinglight.com/studies/dualopsd-adaptive-privileged-teachers-for-on-policy-self-distillation-fa9c8fb0f8 Developing Lighten 2026-08-30T05:43:18.728Z Adaptive AI teacher linked to higher maths scores in larger models https://developinglight.com/studies/slasher-power-flexibility-for-cloud-datacenters-1e4163f70a Developing Lighten 2026-08-30T05:40:17.825Z Preprint models near-zero service cost at a 15% power-reduction goal https://developinglight.com/studies/atmospheric-newtonian-noise-in-torsion-balance-measurements-of-the-gravi-82b29dc4a8 Developing Lighten 2026-08-30T05:31:16.714Z Atmospheric noise may matter more as gravity measurements improve https://developinglight.com/studies/first-plasma-commissioning-and-operational-highlights-from-india-s-first-6443427afb Developing Lighten 2026-08-30T05:16:45.015Z India's New Spherical Tokamak Reaches First Plasma, but Current Fades https://developinglight.com/studies/electron-spin-resonance-driven-photogalvanic-effect-in-graphene-based-st-e7a9b69266 Developing Lighten 2026-08-30T05:07:28.650Z Graphene devices show electrical signals linked to spin resonance https://developinglight.com/studies/a-heterogeneously-integrated-coupled-cavity-frequency-beam-splitter-dc9e8c2ddc Developing Lighten 2026-08-30T05:05:28.054Z Chip Splits and Swaps Light Between Two Frequencies https://developinglight.com/studies/benchmark-brown-dwarfs-as-chemical-laboratories-linking-system-bulk-prop-18d0083223 Developing Lighten 2026-08-30T05:02:57.179Z JWST Models Point to Contrasting Clouds on Two Brown Dwarfs https://developinglight.com/studies/a-scalable-openlb-lammps-framework-for-fully-resolved-simulations-of-hin-97a2e9d6d9 Developing Lighten 2026-08-30T05:01:56.749Z Simulation reports different settling for cubes and spheres https://developinglight.com/studies/immittance-formulas-for-exact-blackbox-quantization-and-divergence-free-e04646bf49 Developing Lighten 2026-08-30T05:01:26.384Z New framework maps complex quantum circuits and flags model limits https://developinglight.com/studies/vulnerable-code-search-transferable-attack-for-code-language-models-11ac8cb9c6 Developing Lighten 2026-08-30T04:57:10.421Z AI Code Search Rankings Shift When Identifiers Are Renamed https://developinglight.com/studies/electrical-manipulation-of-oxygen-stoichiometry-in-multiterminal-yba-2-c-8d04c47c69 Developing Lighten 2026-08-30T04:39:23.440Z Electrical Pulses Are Linked to Selective YBCO Resistance Shifts https://developinglight.com/studies/sparse-operators-and-their-boundedness-on-morrey-type-spaces-an-exposito-1661fc5a65 Developing Lighten 2026-08-30T04:38:22.733Z Preprint outlines a route to weighted bounds on Morrey spaces https://developinglight.com/studies/beyond-local-surprise-grounded-dialogue-as-selective-belief-revision-und-f8c47accaa Developing Lighten 2026-08-30T04:37:07.017Z Grounded dialogue policies retrieved images similarly; DCP-B lagged https://developinglight.com/studies/dual-modes-in-kerr-spacetimes-and-the-whiting-transform-mode-stability-r-311612cbf3 Developing Lighten 2026-08-30T04:31:06.316Z Fourier proof rules out a class of Kerr quasinormal modes https://developinglight.com/studies/trace-integrity-for-llm-data-agents-a-vision-for-auditable-structured-re-967619ba79 Developing Lighten 2026-08-30T04:21:34.800Z LLM data agents can get answers right despite flawed reasoning https://developinglight.com/studies/partially-dynamic-all-pairs-maxflow-and-effective-resistance-via-stable-a6c4bb07c0 Developing Lighten 2026-08-30T04:21:04.131Z New Algorithm Reuses Sparse Graphs for Dynamic All-Pairs Queries https://developinglight.com/studies/rotated-semicircle-laws-for-permanental-roots-of-gaussian-random-matrice-83187ad43b Developing Lighten 2026-08-30T04:16:33.322Z Mathematicians find a rotated semicircle law in random polynomials https://developinglight.com/studies/accretion-disk-sizes-and-temperature-profiles-in-lensed-quasars-nir-micr-77a6d339db Developing Lighten 2026-08-30T04:13:47.588Z Quasar disk study finds size pattern at odds with thin-disk theory https://developinglight.com/studies/critical-norm-profiles-for-finite-prime-composition-operators-on-the-har-27c9a6795d Developing Lighten 2026-08-30T04:13:32.152Z Critical operator limit reflects prime-coordinate concentration https://developinglight.com/studies/adaptive-peer-clustering-with-hierarchical-random-linear-network-coding-84cf3599ca Developing Lighten 2026-08-30T04:12:16.689Z Adaptive wireless coding reports gains in changing networks https://developinglight.com/studies/experimental-characterization-of-additively-manufactured-metallic-alloys-5b96a421c7 Developing Lighten 2026-08-30T04:09:00.926Z 3D-Printed Steel Shows Higher Sputter Yields at Higher Ion Energy https://developinglight.com/studies/depth-control-of-room-temperature-quantum-emitters-in-gallium-nitride-0a61f71610 Developing Lighten 2026-08-30T04:03:45.010Z GaN study finds emitters at interface and a target depth https://developinglight.com/studies/rtlguard-a-lightweight-teacher-student-defense-for-poisoned-rtl-code-gen-9ce8e4214e Developing Lighten 2026-08-30T03:57:28.833Z AI Recovery Method Reports Fewer Hidden Attacks in Hardware Code https://developinglight.com/studies/the-value-of-human-expertise-0068df0def Developing Lighten 2026-08-30T03:46:42.393Z A Mathematical Model Puts a Value on Human Expertise https://developinglight.com/studies/robust-curvemoe-multi-norm-adversarial-defense-for-mixture-of-experts-mo-1fa7efa675 Developing Lighten 2026-08-30T03:45:56.706Z Sparse expert model reports gains against three attack norms https://developinglight.com/studies/polynomial-extensions-do-not-preserve-the-strong-finite-type-property-49e9ea69c8 Developing Lighten 2026-08-30T03:42:41.178Z New Preprint Finds Polynomial Rings Can Lose a Finiteness Property https://developinglight.com/studies/a-family-of-second-order-linear-unconditionally-stable-methods-for-the-c-d11e21ef9d Developing Lighten 2026-08-30T03:40:40.227Z Linear multiphase-flow method shows stability, but one cavity test fails https://developinglight.com/studies/numerical-direct-scattering-transform-for-dark-solitons-ba7a766863 Developing Lighten 2026-08-30T03:39:39.523Z Numerical method recovers scattering data from dark solitons https://developinglight.com/studies/nearly-optimal-strong-coresets-for-ell-p-subspace-approximation-6302da974a Developing Lighten 2026-08-30T03:36:23.983Z New analysis improves row-coreset bounds for subspace problems https://developinglight.com/studies/how-much-rank-does-lora-need-rank-error-bounds-for-transformer-attention-70a52b8aef Developing Lighten 2026-08-30T03:33:52.867Z LoRA theory ties adapter rank to attention error https://developinglight.com/studies/when-obstacles-bend-modeling-vegetation-deformation-in-the-context-of-fi-b1d67fca9d Developing Lighten 2026-08-30T03:33:37.549Z Robot study tests a way to measure how vegetation bends https://developinglight.com/studies/topology-of-fluctuation-bands-in-chiral-active-matter-8aca0c38cc Developing Lighten 2026-08-30T03:27:51.856Z Rotating Noise Linked to Topological Fluctuations in a Lattice https://developinglight.com/studies/linear-isoperimetric-filling-inequalities-in-hadamard-spaces-at-and-abov-267fb03c4f Developing Lighten 2026-08-30T03:27:06.398Z Preprint reports linear filling bounds at and above asymptotic rank https://developinglight.com/studies/giving-mechanical-engineers-intelligent-tools-a-project-based-ai-educati-d71c259525 Developing Lighten 2026-08-30T03:24:20.962Z AI course reports rising grades over three years, but cause is unclear https://developinglight.com/studies/one-policy-many-embodiments-unified-camera-centric-action-geometry-pre-t-d0b1bc6a85 Developing Lighten 2026-08-30T03:20:35.100Z A Single Robot Policy Works Across Robot Bodies, but Transfer Slips https://developinglight.com/studies/fine-tuning-whisper-for-automatic-speech-recognition-in-baniwa-a-prelimi-f6132c5a9a Developing Lighten 2026-08-30T03:16:04.380Z Whisper posts a 37.5% word error rate in Baniwa study https://developinglight.com/studies/asymptotically-optimal-purification-of-noisy-unitary-channels-in-any-dim-1232b42e4b Developing Lighten 2026-08-30T03:14:03.801Z Quantum protocol reaches limit for cleaning noisy unitary channels https://developinglight.com/studies/neural-network-and-reduced-order-modeling-workflows-for-ai-driven-cfd-fa-8d828e7ac2 Developing Lighten 2026-08-30T03:11:33.287Z Jet-Flow Study Finds AI Model Choice Matters by Task https://developinglight.com/studies/three-omitted-values-and-non-blaschke-point-divisors-in-half-planes-9cd76b3039 Developing Lighten 2026-08-30T03:10:47.813Z Preprint builds meromorphic exception to half-plane value rule https://developinglight.com/studies/strong-cp-and-the-pmns-phase-in-dirac-left-right-symmetry-7373c2977e Developing Lighten 2026-08-30T03:09:17.265Z A neutrino model links CP violation to strong-CP constraints https://developinglight.com/studies/lab-earthquakes-confirm-the-theory-of-frictional-slip-pulses-4cc5995252 Developing Lighten 2026-08-30T03:05:31.619Z Lab tests trace how earthquake slip pulses grow and fade https://developinglight.com/studies/group-shared-low-rank-approximation-for-mobile-efficient-pointwise-convo-ee4f7fe9b9 Developing Lighten 2026-08-30T03:03:30.826Z Preprint reports smaller models in large-kernel CNN tests https://developinglight.com/studies/all-once-extended-3d-tqfts-are-reshetikhin-turaev-theories-6aed5d7ddf Developing Lighten 2026-08-30T03:02:30.351Z Geometry puts 3D theories in a Reshetikhin–Turaev framework https://developinglight.com/studies/virtoos-a-ros-2-unity-virtualization-toolkit-for-fleet-management-of-aut-93564c4355 Developing Lighten 2026-08-30T03:01:29.618Z Robot Fleet Toolkit Demonstrated in Virtual and Mixed Reality https://developinglight.com/studies/gating-before-commitment-anticipating-intent-divergence-to-prevent-post-0e96d4897c Developing Lighten 2026-08-30T02:57:43.275Z Autonomous-driving gate shows mixed results in replay tests https://developinglight.com/studies/from-fleet-to-lab-revisiting-the-security-and-complexity-of-industrial-r-420f779c2f Developing Lighten 2026-08-30T02:54:27.391Z Analysis flags vulnerability in Sigries Rowhammer defense https://developinglight.com/studies/incommensurate-spin-fluctuations-in-one-dimensional-kondo-metal-ceco2ga8-95329b7d7e Developing Lighten 2026-08-30T02:53:26.998Z Neutron and ARPES data point to chain-like magnetism in CeCo2Ga8 https://developinglight.com/studies/epistemic-networks-collective-misperception-and-the-manipulation-of-soci-690d3c8127 Developing Lighten 2026-08-30T02:50:11.441Z Model finds expressive bias can persist after signals settle https://developinglight.com/studies/agentic-autoresearch-for-cell-edge-power-control-radically-redefining-th-82b57036dd Developing Lighten 2026-08-30T02:49:56.016Z Preprint: AI Agent Nearly Matches Wireless Benchmark in Simulation https://developinglight.com/studies/optical-extinction-and-near-field-properties-of-plasmonic-dimers-role-of-94777a3870 Developing Lighten 2026-08-30T02:47:25.122Z Silver-wire models reveal shape-dependent hotspots and resonances https://developinglight.com/studies/parity-anomaly-as-modular-commutator-with-massless-dirac-fermion-9701f84fcd Developing Lighten 2026-08-30T02:37:09.066Z A Single Dirac Cone Shows a Half-Quantized Entanglement Signal https://developinglight.com/studies/integrated-hardware-annealing-based-on-langevin-dynamics-for-ising-machi-0cb2d26c37 Developing Lighten 2026-08-30T02:36:53.728Z Preprint reports higher best-known-solution reach in Ising machines https://developinglight.com/studies/negative-effective-divisors-and-bridgeland-stability-of-line-bundles-on-f6527539da Developing Lighten 2026-08-30T02:32:07.753Z Negative divisors can detect line-bundle instability on surfaces https://developinglight.com/studies/yang-lee-criticality-as-a-dissipative-dynamical-phase-transition-quantum-8e69552d27 Developing Lighten 2026-08-30T02:30:06.990Z Quantum Model Links Yang-Lee Criticality to Damping Transition https://developinglight.com/studies/finding-and-using-interpretable-latents-in-a-neutrino-foundation-model-w-59756c50c2 Developing Lighten 2026-08-30T02:28:51.492Z Neutrino AI model maps hidden signals and ranks events by error https://developinglight.com/studies/massive-ghost-confinement-dipole-equation-and-multipole-states-in-quadra-1c21823acf Developing Lighten 2026-08-30T02:24:20.537Z Quantum gravity proposal targets a massive ghost for confinement https://developinglight.com/studies/localic-esakia-duality-via-conic-frames-2f2720c38f Developing Lighten 2026-08-30T02:23:50.107Z Constructive Framework Recasts Esakia Duality https://developinglight.com/studies/the-multivariable-strong-monodromy-conjecture-for-plane-curves-80a4592082 Developing Lighten 2026-08-30T02:22:04.688Z Preprint links plane-curve poles to Bernstein–Sato zeros https://developinglight.com/studies/planetary-prediction-engine-autonomous-geospatial-prediction-via-intelli-9c72fa8434 Developing Lighten 2026-08-30T02:20:48.962Z AI system posts higher scores in varied geospatial tests https://developinglight.com/studies/plansightrag-a-visual-first-multimodal-rag-for-automating-question-answe-fac04f128d Developing Lighten 2026-08-30T02:15:47.609Z AI Finds Civil Plan Pages, But Rules Remain Hard https://developinglight.com/studies/optimal-differentiability-of-isotropic-positive-definite-functions-on-ev-8ef38544c9 Developing Lighten 2026-08-30T02:12:47.044Z Math Study Sets Sharp Smoothness Limit on Even-Dimensional Spheres https://developinglight.com/studies/a-visual-dependence-aware-framework-for-multimodal-unsupervised-continua-01d61a5607 Developing Lighten 2026-08-30T02:12:01.666Z AI method posts higher accuracy while learning new visual tasks https://developinglight.com/studies/myomechanix-biomechanically-grounded-compositional-skilled-activity-unde-3e3160778c Developing Lighten 2026-08-30T02:08:00.931Z New AI dataset pairs fitness video with muscle signals and errors https://developinglight.com/studies/the-isotopic-and-elemental-abundances-of-planet-host-star-trappist-1-a4c8d1cc79 Developing Lighten 2026-08-30T02:01:15.079Z TRAPPIST-1 Shows Solar-Like Chemistry, With a Carbon Twist https://developinglight.com/studies/torsion-balances-as-operational-probes-of-semiclassical-gravity-matched-dafe736a82 Developing Lighten 2026-08-30T01:59:59.650Z New framework turns torsion-balance noise into gravity-test limits https://developinglight.com/studies/viscochiral-transport-chiral-selection-of-hydrodynamic-vortices-by-berry-697b8f6215 Developing Lighten 2026-08-30T01:58:44.238Z Model links Hall-viscosity gradient to unequal electronic vortices https://developinglight.com/studies/probing-the-properties-of-neutral-clouds-in-the-vicinity-of-the-local-bu-155ab057e7 Developing Lighten 2026-08-30T01:55:58.292Z Dust maps suggest some hydrogen clouds lie beyond the Local Bubble https://developinglight.com/studies/refvideo-6m-a-reliable-reference-based-dataset-for-instructional-video-e-8c8570faf5 Developing Lighten 2026-08-30T01:51:57.482Z New six-million-pair dataset targets reference-based video editing https://developinglight.com/studies/multicomponent-magnetic-domain-walls-in-rhombohedral-graphene-c63be76f31 Developing Lighten 2026-08-30T01:48:56.920Z Graphene Preprint Models Two Distinct Magnetic Domain-Wall Textures https://developinglight.com/studies/graded-face-lifts-and-free-extreme-points-of-free-spectrahedra-50f6cde874 Developing Lighten 2026-08-30T01:45:10.694Z Free square contains infinitely many non-unitarily equivalent free extreme points https://developinglight.com/studies/zero-wam-in-context-world-action-modeling-from-human-videos-for-open-end-adf983b547 Developing Lighten 2026-08-30T01:43:54.781Z Robot Policy Scores Higher on Unseen Tasks Using Video Prompts https://developinglight.com/studies/from-natural-language-requirements-to-graphical-user-interfaces-automate-9ffbf92c57 Developing Lighten 2026-08-30T01:38:53.269Z AI GUI prototype reports higher component counts in 19-person study https://developinglight.com/studies/cross-lingual-biography-enrichment-via-claim-extraction-and-alignment-8d142b8432 Developing Lighten 2026-08-30T01:35:07.587Z Outputs using selected claims had more support, fewer hallucinations https://developinglight.com/studies/equilibrium-in-closed-constant-function-market-maker-economies-f0b0f79a11 Developing Lighten 2026-08-30T01:27:51.645Z Preprint finds trading order can shape outcomes in a fee-free market https://developinglight.com/studies/automatic-model-card-generation-using-an-llm-b7f2f9ff55 Developing Lighten 2026-08-30T01:23:05.830Z AI can tidy model cards, but generated errors persist https://developinglight.com/studies/interpretable-physics-informed-retrieval-augmented-generation-language-m-7bde63ac50 Developing Lighten 2026-08-30T01:19:20.167Z AI retrieval system scores higher in crystal route planning https://developinglight.com/studies/geoformer-geometry-aware-transformer-and-its-application-to-5d-first-arr-427b530ed7 Developing Lighten 2026-08-30T01:19:04.518Z GeoFormer posts lowest error in four-site seismic picking test https://developinglight.com/studies/bridge-damage-detection-from-low-light-uav-imagery-via-degradation-aware-a312fbaf5a Developing Lighten 2026-08-30T01:17:03.967Z Study Reports Higher Bridge-Damage Scores After Image Restoration https://developinglight.com/studies/semantic-reconstruction-and-3-d-detection-via-learned-multi-pair-fusion-bdc5e21e8d Developing Lighten 2026-08-30T01:15:18.450Z RF Simulation Finds Learned Labels Hold Together in Heavy Noise https://developinglight.com/studies/pointing-vla-typed-spatial-grounding-interfaces-for-vision-language-acti-ecc33934db Developing Lighten 2026-08-30T01:13:47.929Z Robot tests favor geometric targets over text-generated coordinates https://developinglight.com/studies/first-principles-spin-lattice-coupling-from-downfolded-electron-phonon-i-a1ba3ab11c Developing Lighten 2026-08-30T01:11:17.337Z New method maps how atomic motion changes magnetic exchange https://developinglight.com/studies/the-photon-gas-in-classical-mechanics-a-statistical-mechanical-treatment-e1bf89514a Developing Lighten 2026-08-30T01:11:01.675Z Classical Theory Reproduces Radiation Relations, Not Blackbody Spectrum https://developinglight.com/studies/dynamic-inversion-an-incrementally-evolving-methodology-for-flight-contr-01c9b88030 Developing Lighten 2026-08-30T01:09:01.213Z Flight-control review finds trade-offs among NDI architectures https://developinglight.com/studies/is-next-chunk-reasoning-rl-really-better-than-sft-revisiting-training-st-6129523940 Developing Lighten 2026-08-30T01:05:00.453Z Mixed training scores higher than next-chunk reasoning in AI tests https://developinglight.com/studies/tracking-molecular-hydrogen-formation-from-ionized-water-in-real-time-40b1f87c43 Developing Lighten 2026-08-30T01:00:14.252Z Laser study maps how ionized D2O forms a D-D bond https://developinglight.com/studies/df-moe-generalizable-deepfake-detection-via-multimodal-sparse-mixture-of-cef74a1deb Developing Lighten 2026-08-30T00:59:13.535Z DF-MoE tops MAVOS-DD tests, but wider scores vary https://developinglight.com/studies/classification-testing-a-new-framework-for-drawing-qualitative-conclusio-04cc0d4a86 Developing Lighten 2026-08-30T00:56:12.982Z New statistical framework adds an inconclusive option to research https://developinglight.com/studies/test-time-adaptation-for-ecg-classification-via-sqi-gated-self-training-5c37462266 Developing Lighten 2026-08-30T00:55:42.544Z ECG adaptation method reports gains across three test settings https://developinglight.com/studies/a-comprehensive-analysis-of-arabic-natural-language-processing-research-b4d4d2aaaf Developing Lighten 2026-08-30T00:51:56.975Z Arabic NLP research surges, but dialect coverage remains uneven https://developinglight.com/studies/mediskill-evo-process-constrained-self-evolution-for-evidence-grounded-c-7ea0d8b7ee Developing Lighten 2026-08-30T00:51:11.276Z Preprint reports higher benchmark scores for a clinical AI agent https://developinglight.com/studies/cross-domain-multi-task-data-to-text-generation-without-in-domain-traini-02e35969cd Developing Lighten 2026-08-30T00:45:10.225Z Small AI Models Transfer Across Domains Without Local Training Text https://developinglight.com/studies/trust-but-verify-rigorously-profiling-best-effort-high-performance-compu-6f7c8c7c1e Developing Lighten 2026-08-30T00:42:09.614Z Best-effort computing runs faster in digital-evolution tests https://developinglight.com/studies/cat-gs-balanced-multimodal-learning-via-calibrated-gating-and-fusion-sur-7c9796793b Developing Lighten 2026-08-30T00:39:54.189Z Multimodal training controller leads on two benchmarks but trails on VGGSound https://developinglight.com/studies/momadv2-reliable-temporal-memory-for-end-to-end-autonomous-driving-b2bd07ef0a Developing Lighten 2026-08-30T00:38:08.416Z MomADv2 Reports Fewer Collisions in Six-Second Planning Tests https://developinglight.com/studies/an-elementary-note-on-three-consecutive-powerful-numbers-1d76477069 Developing Lighten 2026-08-30T00:34:52.455Z A preprint claims a narrow ban on three consecutive powerful numbers https://developinglight.com/studies/resilience-matters-for-embodied-agents-system-new-metrics-systematic-eva-284b8a3891 Developing Lighten 2026-08-30T00:30:21.523Z Simulated Tasks Reveal a 4.8-Fold Gap in Agent Recovery Costs https://developinglight.com/studies/temporal-property-driven-design-space-exploration-with-reinforcement-lea-85f08ee9ec Developing Lighten 2026-08-30T00:25:50.924Z Reinforcement learning finds mine-pump designs with fewer simulations https://developinglight.com/studies/repairing-locally-misspecified-gmm-an-empirical-bayes-approach-694eb784ac Developing Lighten 2026-08-30T00:22:19.981Z In one IV simulation, empirical Bayes had the lowest GMM error https://developinglight.com/studies/ade-agentic-data-evolution-framework-for-human-centered-objectives-1abaf6c62c Developing Lighten 2026-08-30T00:21:04.143Z AI tutoring answers win more benchmark comparisons after refinement https://developinglight.com/studies/borell-brascamp-lieb-inequality-with-finitely-many-output-functions-b8c63bcbd3 Developing Lighten 2026-08-30T00:18:18.250Z Math preprint presents BBL inequality with separate outputs https://developinglight.com/studies/partab-partition-aware-reasoning-with-structured-evidence-for-scalable-t-65a206a672 Developing Lighten 2026-08-30T00:17:47.500Z PARTAB reports benchmark gains with less table context https://developinglight.com/studies/mnemodyn-learning-resting-state-dynamics-from-40k-fmri-sequences-5af61c5d83 Developing Lighten 2026-08-30T00:10:31.406Z AI model rebuilds resting-state brain scans across datasets https://developinglight.com/studies/generalization-memorization-and-overfitting-for-diffusion-models-trained-841a6e0a63 Developing Lighten 2026-08-30T00:07:45.475Z Diffusion models may overfit long before they memorize data https://developinglight.com/studies/luce-relightable-gaussians-for-3d-asset-generation-e1fd1b0cd5 Developing Lighten 2026-08-30T00:06:14.871Z AI System Turns One Image Into Relightable 3D Assets https://developinglight.com/studies/free-bakry-emery-calculus-364f24a958 Developing Lighten 2026-08-30T00:03:59.259Z Preprint links curvature to inequalities in free Langevin dynamics https://developinglight.com/studies/sensorless-damage-safe-grasping-38f93a6088 Developing Lighten 2026-08-30T00:03:42.960Z Sensorless Gripper Reports Lower Damage on Soft TPU Cubes https://developinglight.com/studies/the-empire-long-divided-must-unite-architectural-convergence-in-three-ll-50e8995873 Developing Lighten 2026-08-29T23:56:26.820Z Three Coding-Agent Systems Converge, but a Trust Gap Remains https://developinglight.com/studies/hybrid-semantic-tool-discovery-for-enterprise-mcp-gateway-architecture-a-ab9510ea5a Developing Lighten 2026-08-29T23:56:11.171Z PayPal preprint reports SCOUT's tool-schema context comparison https://developinglight.com/studies/ragsentinel-certifiable-geometric-consensus-for-robust-retrieval-augment-294d5bdb3d Developing Lighten 2026-08-29T23:50:55.503Z RAG tests report lower attack success with a new filter https://developinglight.com/studies/physics-informed-wifi-sensing-for-robust-3d-human-pose-estimation-in-mob-b4bd55c7f9 Developing Lighten 2026-08-29T23:45:24.734Z WiFi system reports lower errors estimating 3D human pose https://developinglight.com/studies/incorporating-cognitive-load-and-knowledge-transfer-for-multi-domain-kno-7782855ca5 Developing Lighten 2026-08-29T23:41:23.851Z AI model tops tests across four multi-subject learning datasets https://developinglight.com/studies/partitions-of-canonical-bases-3fccb64c09 Developing Lighten 2026-08-29T23:40:23.120Z Three mathematical views of canonical bases are shown to agree https://developinglight.com/studies/tagr-temporally-adaptive-generative-recommendation-for-industrial-live-s-ba815568af Developing Lighten 2026-08-29T23:39:52.666Z Live-ad system reports 16.1% revenue lift in randomized test https://developinglight.com/studies/jepa-x-cross-predictive-physics-grounding-for-forecastable-latent-dynami-5cb1afb03a Developing Lighten 2026-08-29T23:36:22.047Z Robot-vision model shows more predictable dynamics with physics data https://developinglight.com/studies/algorithmic-impact-reveals-the-hidden-social-choice-structure-of-alignme-8ed0366568 Developing Lighten 2026-08-29T23:34:36.152Z AI alignment study finds welfare costs as harm limits tighten https://developinglight.com/studies/quadrilateral-and-hexahedral-immersed-finite-element-methods-for-ellipti-cc72e32917 Developing Lighten 2026-08-29T23:32:35.389Z Finite-element proposal expands mesh options for interface problems https://developinglight.com/studies/don-t-just-listen-try-planning-graph-based-retrieval-generation-agent-fo-2acf13f2e0 Developing Lighten 2026-08-29T23:32:19.729Z AI agent scores 65% on meeting inference questions, versus 25% https://developinglight.com/studies/ponderpounce-a-pretrained-mllm-as-an-episode-context-engine-for-robot-co-aee4a41612 Developing Lighten 2026-08-29T23:29:19.237Z Larger AI memory model linked to better robot control in tests https://developinglight.com/studies/compression-trinity-exploring-sparsity-quantization-and-low-rank-approxi-d3da1c91da Developing Lighten 2026-08-29T23:21:18.106Z Three-way compression reports speed and accuracy gains https://developinglight.com/studies/a-mesh-free-multiresolution-deep-energy-method-with-phase-field-modeling-59cdf1c323 Developing Lighten 2026-08-29T23:16:02.168Z Mesh-free model reproduces complex crack patterns in tests https://developinglight.com/studies/robust-code-rl-via-faulty-code-driven-test-case-synthesis-and-dense-rewa-dc5aeda608 Developing Lighten 2026-08-29T23:13:16.668Z Fault-driven tests linked to higher scores for a code model https://developinglight.com/studies/decoupling-candidate-dual-agn-from-chance-superpositions-in-the-gothic-s-c8a8739759 Developing Lighten 2026-08-29T23:11:16.111Z AI screens 46,061 rejected galaxies for paired nuclei https://developinglight.com/studies/from-state-to-action-ooda-tool-for-reliable-multi-turn-tool-use-76418aba71 Developing Lighten 2026-08-29T23:09:15.539Z Staged AI Agent Tops Multi-Turn Tool Tests Across Model Sizes https://developinglight.com/studies/strive-multi-agent-structured-temporal-reasoning-with-integrated-verific-dd14177100 Developing Lighten 2026-08-29T23:08:30.139Z STRIVE scores higher at tracking change in chest X-ray reports https://developinglight.com/studies/task-adaptive-rubrics-for-gui-reward-modeling-32037800a7 Developing Lighten 2026-08-29T23:07:59.686Z Task-adaptive rubrics outperform other GUI reward verifiers in tests https://developinglight.com/studies/an-ab-initio-framework-for-simulating-ultrafast-nonlinear-cavity-quantum-c89e49e8bd Developing Lighten 2026-08-29T23:00:28.688Z Computer model simulates polariton signals in pyrazine https://developinglight.com/studies/gradient-extrapolation-based-distributed-mirror-descent-algorithm-for-mu-50104d8a3b Developing Lighten 2026-08-29T22:57:57.859Z Preprint reports quicker equilibrium searches in group games https://developinglight.com/studies/end-point-non-regularity-of-the-dirichlet-problem-on-lipschitz-domains-a-d6d705a203 Developing Lighten 2026-08-29T22:57:42.422Z Maths preprint finds endpoint regularity failure on one domain https://developinglight.com/studies/robust-training-and-rigorous-error-analysis-of-physics-informed-neural-n-ced98e40b1 Developing Lighten 2026-08-29T22:54:11.874Z Weak-form neural network reports lower errors in p-Laplace tests https://developinglight.com/studies/opal-jl-a-comprehensive-composable-framework-for-data-assimilation-in-ju-c755ff02d6 Developing Lighten 2026-08-29T22:53:41.505Z Opal.jl brings data-assimilation methods into one workflow https://developinglight.com/studies/task-disentangled-low-rank-adaptation-for-versatile-audio-visual-multi-m-a305c3dc30 Developing Lighten 2026-08-29T22:52:55.808Z Audio-video model reports broad benchmark gains, with one caveat https://developinglight.com/studies/metarag-belief-action-aligned-policy-optimization-for-agentic-rag-756285b86e Developing Lighten 2026-08-29T22:49:39.918Z AI method posts higher answer accuracy across seven QA tests https://developinglight.com/studies/validation-of-hrv-studio-a-transparent-and-quality-control-aware-platfor-eb6c712452 Developing Lighten 2026-08-29T22:45:54.239Z HRV software agrees closely on many measures, but VLF varies https://developinglight.com/studies/fifty-years-of-wave-equations-with-time-dependent-propagation-speeds-a-v-e43a769ab2 Developing Lighten 2026-08-29T22:45:38.577Z A constructed wave equation stays well posed beyond half-order regularity https://developinglight.com/studies/real-world-knowledge-guided-change-data-synthesis-for-remote-sensing-833c1edd18 Developing Lighten 2026-08-29T22:41:53.029Z Preprint reports gains from AI-guided remote-sensing data https://developinglight.com/studies/contrastive-branch-policy-optimization-9139bdd153 Developing Lighten 2026-08-29T22:39:37.581Z Branch-Based AI Method Scores Higher on Math and Search Tests https://developinglight.com/studies/senseshift-continuous-sentiment-controlled-text-generation-via-encoder-b-8edef89094 Developing Lighten 2026-08-29T22:38:51.892Z AI system tracks sentence-level sentiment targets in benchmark tests https://developinglight.com/studies/qismc-a-model-checker-for-qiskit-program-debugging-3a130831f5 Developing Lighten 2026-08-29T22:31:20.326Z Quantum checker reaches 100-qubit tests but hits limits https://developinglight.com/studies/data-driven-dimension-reduction-for-industrial-load-modeling-using-inver-027af76bff Developing Lighten 2026-08-29T22:27:49.693Z Study reports lower errors for data-driven industrial-load models https://developinglight.com/studies/a-line-search-free-coordinate-proximal-predictor-corrector-method-for-mo-a1abf26649 Developing Lighten 2026-08-29T22:24:18.847Z Coordinate solver leads tests of larger absolute-value systems https://developinglight.com/studies/on-the-index-of-borel-subalgebras-of-lie-superalgebras-21400db67f Developing Lighten 2026-08-29T22:24:03.407Z Preprint Finds Zero-Index Cases in Lie Superalgebra Modules https://developinglight.com/studies/bridging-adversarial-and-collaborative-learning-for-ai-generated-image-q-ddedc3fec7 Developing Lighten 2026-08-29T22:21:17.614Z AI image grader weighs visual quality against prompt fit https://developinglight.com/studies/interstellar-aromatic-infrared-bands-ir-absorption-spectroscopy-of-the-l-0b6508c356 Developing Lighten 2026-08-29T22:17:31.992Z Naphthalene’s Triplet State Shows a Distinct Infrared Signature https://developinglight.com/studies/nearly-optimal-amplitude-estimation-at-any-depth-68b9740ea7 Developing Lighten 2026-08-29T22:16:46.332Z Quantum method targets uniform accuracy for amplitude estimates across depths https://developinglight.com/studies/implicit-q-learning-bootstrapped-ant-colony-optimization-for-maritime-mo-6e597d0a10 Developing Lighten 2026-08-29T22:16:00.978Z Adaptive satellite scheduler leads in simulated maritime tests https://developinglight.com/studies/beyond-static-interpretability-anticipating-post-sft-mechanisms-from-pre-767bb2bd6d Developing Lighten 2026-08-29T22:12:45.281Z Language-model method predicts target-task mechanisms before fine-tuning https://developinglight.com/studies/when-stale-constraints-go-unchecked-budgeted-verification-failures-in-in-0f4fc0dbd9 Developing Lighten 2026-08-29T22:10:59.827Z Targeted Checks Cut AI Errors From Outdated Memory https://developinglight.com/studies/scalable-and-versatile-identification-for-hierarchical-structural-causal-5f2df87ae2 Developing Lighten 2026-08-29T22:10:14.141Z New model reports 36.80 class-level contrast in STAR data https://developinglight.com/studies/iapo-influence-aware-policy-optimization-for-credit-assignment-in-multi-5b77a557ae Developing Lighten 2026-08-29T22:09:28.456Z IAPO Scores Higher Than Matched GRPO on Several Qwen3-4B/8B Agent Tests https://developinglight.com/studies/it-depends-incorporating-correlations-for-joint-aleatoric-and-epistemic-d262effca4 Developing Lighten 2026-08-29T22:06:57.633Z Correlated uncertainty model tops held-out likelihood scores https://developinglight.com/studies/x-multi-vlm-based-imaging-factor-disentanglement-for-factor-aware-image-a8a10221d0 Developing Lighten 2026-08-29T22:06:41.965Z AI image system scores higher on unseen imaging combinations https://developinglight.com/studies/array-agnostic-ambisonics-encoding-via-diffusion-posterior-sampling-c14f18f7a5 Developing Lighten 2026-08-29T22:00:40.763Z ADEPS Tops Signal-Distortion Tests Across Mic Layouts https://developinglight.com/studies/direct-linearization-cauchy-matrix-and-sato-grassmannian-5f609a53c9 Developing Lighten 2026-08-29T21:59:24.985Z Preprint links nonlinear matrix equations to a wider KP framework https://developinglight.com/studies/is-discrete-difficulty-sufficient-leveraging-continuous-difficulty-for-e-de0f733d70 Developing Lighten 2026-08-29T21:53:53.746Z Flexible Self-Consistency reports lower token use in language-model benchmarks https://developinglight.com/studies/taming-foundation-model-with-invariance-oriented-pre-training-for-broad-78c2745374 Developing Lighten 2026-08-29T21:48:52.990Z INCEPT posts strong results across varied EEG benchmark tasks https://developinglight.com/studies/bandit-submodular-maximization-under-matroid-constraints-learning-compre-a5c001bbb8 Developing Lighten 2026-08-29T21:46:07.314Z Preprint reports sublinear regret with feasible queries only https://developinglight.com/studies/when-similarity-is-interaction-driven-quantum-kernels-for-regime-sensiti-2d5e0aa905 Developing Lighten 2026-08-29T21:43:51.794Z Quantum kernel tops synthetic tests, but trails on one fraud benchmark https://developinglight.com/studies/ultrafast-tuning-of-the-focusing-efficiency-of-a-nonlinear-atomically-th-4d841d8276 Developing Lighten 2026-08-29T21:43:36.448Z Atom-thin lens shows 30% modulation in focused light https://developinglight.com/studies/everest-endogenous-vision-language-reinforcement-reasoning-exploration-f-cc40183150 Developing Lighten 2026-08-29T21:40:35.861Z AI mapping system leads urban segmentation benchmark https://developinglight.com/studies/thermal-diffuse-scattering-in-tem-complex-absorptive-potentials-compared-69596eee24 Developing Lighten 2026-08-29T21:40:20.195Z TEM Models Diverge in Heavy Crystals, but Aperture Helps https://developinglight.com/studies/inertial-primal-dual-dynamics-methods-featuring-implicit-hessian-driven-bf53f0246a Developing Lighten 2026-08-29T21:36:33.895Z Optimization method reports fast convergence in theory and tests https://developinglight.com/studies/visualizing-flat-band-spatial-renormalization-in-rhombohedral-graphene-s-9e93e88f4e Developing Lighten 2026-08-29T21:36:03.444Z Graphene Flat Bands Shift Across Moiré Sites Beyond Roughly 10 Nanometres https://developinglight.com/studies/the-isaacs-navarro-wolf-conjecture-b04b6d881f Developing Lighten 2026-08-29T21:31:17.371Z AI-Assisted Preprint Claims Proof of Finite-Group Conjecture https://developinglight.com/studies/skillforge-evolving-verifiable-skills-for-reinforcement-learning-agents-fde71a5ada Developing Lighten 2026-08-29T21:28:16.436Z SkillForge reports higher benchmark scores for language-model agents https://developinglight.com/studies/multiple-pattern-formation-in-quorum-sensing-of-density-enhanced-motilit-5cb4928f1c Developing Lighten 2026-08-29T21:24:45.601Z Simulation Associates Local Sensing With Three Particle Patterns https://developinglight.com/studies/splay-induced-charging-of-ferroelectric-droplets-in-highly-non-uniform-e-5049090d85 Developing Lighten 2026-08-29T11:07:00.702Z Uneven electric fields linked to motion in some liquid-crystal droplets https://developinglight.com/studies/learning-late-guiding-early-timestep-decoupled-semantic-guidance-for-fai-025664a93b Developing Lighten 2026-08-29T10:58:44.639Z Preprint Reports More Balanced Synthetic Faces With One-Step Guidance https://developinglight.com/studies/unveiling-spectral-mechanisms-in-training-free-llm-text-detection-4306e2c84c Developing Lighten 2026-08-29T10:55:44.032Z AI Text Detectors Need Different Signals for Different Writing https://developinglight.com/studies/strategybench-evaluating-explicit-strategy-induction-in-large-language-m-2da1f1e51f Developing Lighten 2026-08-29T08:45:15.007Z Combined strategy setup led the main AI benchmark comparison https://developinglight.com/studies/interpretable-fundus-image-classification-via-ring-based-retinal-vascula-fe060bd2b8 Developing Lighten 2026-08-29T08:43:14.383Z Retinal vessel rings match an AI model, but lag it elsewhere https://developinglight.com/studies/valid-test-for-multi-arm-trials-with-generalized-linear-models-under-cov-6a9770b13d Developing Lighten 2026-08-29T08:37:13.528Z New test targets false alarms in multi-arm clinical trials https://developinglight.com/studies/shsp-structure-aware-hierarchical-solution-prediction-for-mixed-integer-f5b004bf7f Developing Lighten 2026-08-29T08:35:27.715Z Preprint reports 54% smaller optimization gaps for SHSP https://developinglight.com/studies/wave-reversing-the-guidance-hierarchy-for-coarse-to-fine-guided-depth-su-5544381f45 Developing Lighten 2026-08-29T08:32:27.058Z WAVE reports lower depth errors at 32× in benchmark tests https://developinglight.com/studies/insightsr-refining-symbolic-regression-search-spaces-via-parallel-semant-0b42e9eee3 Developing Lighten 2026-08-29T08:28:41.333Z AI-guided equation search reports strong benchmark results https://developinglight.com/studies/metis-typed-runtime-mediation-for-tool-using-software-agents-8435b63e08 Developing Lighten 2026-08-29T08:17:40.167Z Agent-runtime study finds faster runs but exposes safety limits https://developinglight.com/studies/not-all-attention-heads-contribute-to-critical-visual-token-selection-he-0f65a7d011 Developing Lighten 2026-08-29T08:12:09.042Z AI Study Finds Some Vision-Model Heads Matter More Than Others https://developinglight.com/studies/recovering-the-effective-impact-parameters-in-integral-field-absorption-366f4be7c0 Developing Lighten 2026-08-29T07:57:37.294Z Arc-tomography simulation favors flux-weighted positions over spaxel centres https://developinglight.com/studies/framing-war-across-languages-power-agency-and-sentiment-in-wikipedia-s-m-60b474aeda Developing Lighten 2026-08-29T07:52:36.259Z Wikipedia war narratives differ sharply across languages https://developinglight.com/studies/pasta-noisy-node-classification-with-partial-label-learning-d2774713f6 Developing Lighten 2026-08-29T07:44:19.809Z Graph-learning preprint reports higher accuracy on noisy labels https://developinglight.com/studies/electrostatic-control-of-li-density-and-transport-rate-in-double-gated-v-3047253de6 Developing Lighten 2026-08-29T07:39:18.659Z Preprint reports separate control of lithium density and flow https://developinglight.com/studies/absence-of-lateral-domain-wall-mobility-in-zn1-xmgxo-thin-films-91244b971c Developing Lighten 2026-08-29T07:32:17.041Z Preprint finds no detectable sideways motion in thin-film domains https://developinglight.com/studies/highly-stable-glasses-of-cis-decalin-and-cis-trans-decalin-mixtures-51f8498e43 Developing Lighten 2026-08-29T07:27:16.164Z Vapor-deposited decalin glasses show unusual stability https://developinglight.com/studies/comparison-of-mechanical-and-molecular-measures-of-mobility-during-const-038453e968 Developing Lighten 2026-08-29T07:22:45.359Z Two mobility measures in a polymer glass diverged before yield https://developinglight.com/studies/trustdabench-benchmarking-reliability-and-robustness-of-llms-for-structu-96f4ae980d Developing Lighten 2026-08-29T07:14:59.271Z LLMs often answer table questions without enough evidence https://developinglight.com/studies/classification-of-metal-insulator-transitions-relating-characteristic-pr-ae3cca8541 Developing Lighten 2026-08-29T06:59:12.465Z Pressure study proposes three types of metal-insulator transition https://developinglight.com/studies/literarybigfive-author-personalized-text-generation-in-a-unified-interpr-6d8cceea32 Developing Lighten 2026-08-29T06:55:56.537Z Preprint Reports Stronger Literary Style Matching Across Four Books https://developinglight.com/studies/anomalous-stabilization-of-excitons-by-metallic-proximity-f9a8409c22 Developing Lighten 2026-08-29T06:52:40.951Z Exciton binding energy rose before palladium-doped films turned metallic https://developinglight.com/studies/environmental-control-extends-beyond-quantum-dephasing-in-exciton-energy-0806103d78 Developing Lighten 2026-08-29T06:40:54.601Z Molecular Antenna Transfers Energy Fastest at Moderate Temperatures https://developinglight.com/studies/quantifying-the-biophysical-properties-of-red-blood-cells-in-gaucher-dis-c20ed974bb Developing Lighten 2026-08-29T06:37:08.989Z Gaucher disease models suggest red cells may struggle in tight spaces https://developinglight.com/studies/an-end-to-end-trained-vision-language-model-for-native-language-prostate-6c84f59a17 Developing Lighten 2026-08-29T06:32:23.184Z German AI model spots prostate cancer better than it grades it https://developinglight.com/studies/depth-resolved-evolution-of-buried-polar-topologies-in-a-pbtio3-srtio3-s-0c0a030149 Developing Lighten 2026-08-29T06:28:52.577Z Microscopy finds hidden depth changes inside polar superlattice https://developinglight.com/studies/the-achilles-tendon-enthesis-rebuilds-its-mineralization-front-on-reload-0607fe45fd Developing Lighten 2026-08-29T06:11:35.965Z Tendon boundary sharpened after reloading, but mineral imprint remained https://developinglight.com/studies/the-power-law-in-human-mobility-is-a-mixture-artifact-evidence-from-a-pa-053306412f Developing Lighten 2026-08-29T06:08:35.319Z A pooled-data effect may explain human travel’s power law https://developinglight.com/studies/mapping-metallic-d-wave-altermagnetism-across-the-tinisi-structural-fami-2f4fa40357 Developing Lighten 2026-08-29T05:51:33.322Z Computational screen finds 12 new metallic altermagnets https://developinglight.com/studies/ggss-geodesic-gated-spherical-steering-for-inference-time-debiasing-of-g-d95e8ccfb2 Developing Lighten 2026-08-29T04:19:54.589Z New method shows lower bias scores in four vision-language models https://developinglight.com/studies/expanding-solutions-to-the-compressible-navier-stokes-equations-with-deg-f5fddff4d4 Developing Lighten 2026-08-29T04:15:23.846Z Small Expanding Flow Disturbances Stay Stable in New Preprint https://developinglight.com/studies/poscosea-a-computationally-efficient-sensitivity-analysis-for-bayesian-m-c0b929fe7d Developing Lighten 2026-08-29T04:10:53.105Z Covariance method matches GLMM influence checks, weaker N-mixture agreement https://developinglight.com/studies/distance-is-not-enough-forget-retain-alignment-gap-predicts-llm-relearni-fdcd04c4a6 Developing Lighten 2026-08-29T04:07:22.544Z AI unlearning study finds update patterns matter more than size https://developinglight.com/studies/gloss-geometric-local-self-similarity-learning-for-faithful-reference-gu-25b83a2382 Developing Lighten 2026-08-29T03:52:36.124Z AI texture tool scores best on three of four reported measures https://developinglight.com/studies/exact-hausdorff-measures-and-fine-geometry-of-collisions-of-independent-93f40f01a0 Developing Lighten 2026-08-29T03:49:50.528Z New framework maps the fractal geometry of random collisions https://developinglight.com/studies/smart-mllm-guided-temporal-alignment-for-unifying-sign-language-recognit-cbe20e74d0 Developing Lighten 2026-08-29T03:46:04.853Z AI system combines sign-language recognition and sign spotting https://developinglight.com/studies/reliablerag-combating-misinformation-in-retrieval-augmented-generation-v-0a3c17a22d Developing Lighten 2026-08-29T03:43:49.390Z AI method tests a finer way to screen misinformation in answers https://developinglight.com/studies/transretrieval-scaling-up-transformer-based-retrieval-for-industrial-rec-2dc60fce89 Developing Lighten 2026-08-29T03:31:47.924Z Recommendation model reports higher recall with less computing power https://developinglight.com/studies/reflection-steering-disentangling-reflection-from-reasoning-in-activatio-cbb3e90bd0 Developing Lighten 2026-08-29T03:29:17.343Z AI Study Reports Fewer Reasoning Tokens With Math Accuracy Intact https://developinglight.com/studies/vizanchor-decoding-manipulation-intent-from-tampering-visualizations-via-c608fa6424 Developing Lighten 2026-08-29T03:23:31.252Z AI system traces how charts were altered and why https://developinglight.com/studies/quantum-vibronic-dynamics-shape-catalytically-relevant-au-ligand-interfa-255c228da6 Developing Lighten 2026-08-29T03:20:00.673Z Gold cluster study tracks a staged sequence of vibrations https://developinglight.com/studies/algaas-nanowires-as-a-universal-platform-for-gaas-ingaas-and-inas-quantu-ca18f980ed Developing Lighten 2026-08-29T03:17:14.807Z Quantum-dot light varies across AlGaAs nanowire designs https://developinglight.com/studies/intrinsic-radial-landau-rainbow-in-triaxial-strained-photonic-crystal-99b672b43f Developing Lighten 2026-08-29T03:13:59.235Z Simulation maps resonance frequency to radial position https://developinglight.com/studies/farca-fact-aligned-reliability-aware-credit-assignment-for-reinforcement-c5e1c65af9 Developing Lighten 2026-08-29T03:10:28.536Z AI training method reports stronger factuality scores in tests https://developinglight.com/studies/method-mind-and-morality-how-people-make-sense-of-artificial-intelligenc-643092c938 Developing Lighten 2026-08-29T02:57:27.159Z Study maps how people make sense of artificial intelligence https://developinglight.com/studies/data-leakage-inflates-generalizability-of-power-outage-prediction-models-7e0ad99c27 Developing Lighten 2026-08-29T02:53:56.624Z Power-outage models falter on unseen states and events https://developinglight.com/studies/the-optimal-redshift-for-dark-energy-ii-application-to-cosmological-data-f61a4187cd Developing Lighten 2026-08-29T02:46:40.717Z Cosmology preprint reports possible dark-energy crossing https://developinglight.com/studies/metalens-stereoscopic-optical-palpation-for-imaging-cancer-mechanics-cdcd473c42 Developing Lighten 2026-08-29T02:42:24.780Z Metalens scanner maps mechanical differences in cancer tissue samples https://developinglight.com/studies/plastic-relaxation-without-dislocations-in-ga-2-o-3-heteroepitaxy-a-stru-4c48033ff3 Developing Lighten 2026-08-29T02:38:24.091Z Model suggests how beta-gallium oxide films relax without dislocations https://developinglight.com/studies/global-geometry-of-the-genotype-phenotype-map-illuminates-a-trade-off-be-2985110da3 Developing Lighten 2026-08-29T02:33:53.477Z Evolutionary model links reliable expression to slower adaptation https://developinglight.com/studies/flare-verifying-milp-reformulations-with-llm-based-theorem-proving-e6303dabf9 Developing Lighten 2026-08-29T01:29:33.180Z AI system checks optimization reformulations with proof certificates https://developinglight.com/studies/a-topological-view-of-algebraic-structures-in-c-p-x-dd346bbc78 Developing Lighten 2026-08-29T01:21:47.037Z Theorems link function rings to properties of abstract spaces https://developinglight.com/studies/two-branch-detector-response-for-dirac-infall-into-a-schwarzschild-mog-b-570a527f9d Developing Lighten 2026-08-29T01:18:01.007Z Black-hole detector model finds thermal behaviour only in a narrow limit https://developinglight.com/studies/crossmambatuning-synergistic-spatial-and-cross-layer-adaptation-for-mach-4075017c5d Developing Lighten 2026-08-29T01:15:15.449Z Small adapters show promise in machine-vision compression https://developinglight.com/studies/grip-granular-reward-guided-parameter-interpolation-for-efficient-reason-61fc84be89 Developing Lighten 2026-08-29T01:11:14.420Z GRIP model blend reports similar accuracy with fewer tokens https://developinglight.com/studies/super-odometry-2-0-resilient-odometry-via-hierarchical-adaptation-f1b8854f78 Developing Lighten 2026-08-29T01:08:13.385Z Adaptive odometry reports low drift in degraded sensor tests https://developinglight.com/studies/vista-visually-inferred-spatial-contact-attention-for-contact-rich-manip-d0175332c2 Developing Lighten 2026-08-29T01:03:57.474Z Preprint Reports Stronger Robot Results With Visual Contact Cues https://developinglight.com/studies/cramer-control-via-request-aware-masking-for-editing-recommenders-4b72b1ab1b Developing Lighten 2026-08-29T00:53:40.787Z AI layer posts higher offline scores for user-requested recommendations https://developinglight.com/studies/narcissus-program-synthesis-using-context-aware-llm-approximations-1a8316681b Developing Lighten 2026-08-29T00:30:22.929Z One-shot AI guidance outperforms static help in program searches https://developinglight.com/studies/deep-learning-segmentation-of-diffusion-weighted-mri-acute-ischaemic-str-05f70c6d1c Developing Lighten 2026-08-29T00:26:37.087Z A simple AI model segmented stroke lesions on MRI in seconds https://developinglight.com/studies/psrl-efficient-training-for-agentic-ai-via-training-time-prefix-sharing-0d3797ea90 Developing Lighten 2026-08-29T00:22:36.278Z AI training system reports faster updates on agent workloads https://developinglight.com/studies/dynamic-polyhedral-logic-bc048d42c6 Developing Lighten 2026-08-29T00:18:50.396Z New Logic Links Reachability, Shape and Reversible Motion https://developinglight.com/studies/the-limits-of-automatic-evaluation-of-creativity-in-large-language-model-c166d79118 Developing Lighten 2026-08-29T00:16:19.686Z Study finds most AI creativity metrics diverge from human ratings https://developinglight.com/studies/mimonet-multi-scale-input-and-multi-scale-output-network-for-salient-obj-3af96dfea3 Developing Lighten 2026-08-29T00:13:33.668Z MIMONet reports better scores in most salient-object comparisons https://developinglight.com/studies/propensity-straight-through-gradients-for-discrete-stochastic-systems-72149cd245 Developing Lighten 2026-08-28T23:50:01.175Z Preprint Method Keeps Stochastic Simulations Exact Forward https://developinglight.com/studies/watching-a-gan-transistor-switch-real-time-nanoscale-strain-and-heat-dyn-e31fbedb18 Developing Lighten 2026-08-28T21:17:43.203Z X-ray study finds opposite strain in a switching GaN transistor https://developinglight.com/studies/dark-matters-of-principle-principles-in-the-dark-matter-paradigm-2779633e27 Developing Lighten 2026-08-28T21:15:42.619Z Sterile Neutrinos Lead a Principle-Based Dark Matter Ranking https://developinglight.com/studies/srpo-self-reflective-policy-optimization-for-long-horizon-reasoning-8d501abc00 Developing Lighten 2026-08-28T21:09:11.437Z AI method reports benchmark gains over GRPO with far less compute https://developinglight.com/studies/wall-crossing-for-equivariant-dt4-invariants-7833648e39 Developing Lighten 2026-08-28T21:04:10.380Z Mathematicians extend wall-crossing formula to Calabi–Yau four theories https://developinglight.com/studies/exact-mutual-information-for-free-massless-fermions-in-any-dimension-d16f378a35 Developing Lighten 2026-08-28T20:59:54.567Z Study derives exact formulas for how massless fermions share information https://developinglight.com/studies/diffractive-optical-element-for-super-gaussian-beam-shaping-on-intersate-bf871e183b Developing Lighten 2026-08-28T20:56:23.558Z Simulated Beam Shaper Shows Limits at Higher Super-Gaussian Orders https://developinglight.com/studies/bluebird-6-is-fainter-than-block-1-satellites-9349abb394 Developing Lighten 2026-08-28T20:52:37.762Z BlueBird 6 Satellite Fainter Than Earlier Spacecraft https://developinglight.com/studies/autonomous-mathematical-discovery-in-an-open-world-multi-agent-environme-7c2aba6933 Developing Lighten 2026-08-28T20:46:21.744Z AI agents report five math results judged novel in test https://developinglight.com/studies/driftad-visually-guided-text-drift-for-few-shot-industrial-anomaly-detec-4a89c48599 Developing Lighten 2026-08-28T20:41:50.887Z Few-shot AI method reports strong scores on industrial defect tests https://developinglight.com/studies/gap-prompt-gated-adaptive-prompting-for-efficient-continual-learning-0b8ce5d058 Developing Lighten 2026-08-28T20:37:50.083Z GAP-Prompt reports higher accuracy on three image benchmarks https://developinglight.com/studies/velocity-coupled-representation-refinement-for-satellite-orbit-predictio-b25b47a0a2 Developing Lighten 2026-08-28T20:33:34.043Z OrbitNet leads Starlink tests with 3.34-metre average error https://developinglight.com/studies/generating-intervention-hypotheses-using-explainable-explanations-on-gra-7a1436861c Developing Lighten 2026-08-28T20:27:32.861Z Graph method reports model-based coverage for all 51 at-risk individuals https://developinglight.com/studies/the-dialect-tax-dialectal-biases-persist-throughout-the-language-modelin-36e09a6cb1 Developing Lighten 2026-08-28T20:20:46.621Z Study Finds AAVE Reasoning Accuracy Gap in Language Models https://developinglight.com/studies/what-reaches-expert-review-representation-structural-screening-and-candi-5ce84a8902 Developing Lighten 2026-08-28T20:13:15.626Z AI screening settings linked to different test items for review https://developinglight.com/studies/expconcad-experience-guided-text-to-cad-generation-from-shape-descriptio-1334b1c2ff Developing Lighten 2026-08-28T19:58:58.766Z Experience-guided AI scores higher on hard CAD prompts https://developinglight.com/studies/infant-care-video-dataset-for-classification-of-interventions-using-tran-4501346fae Developing Lighten 2026-08-28T19:55:57.717Z NICU video models outscore a frame-by-frame system in manikin test https://developinglight.com/studies/shardmeter-sharded-and-geo-distributed-training-without-the-guesswork-eeeeceb6d9 Developing Lighten 2026-08-28T19:51:41.954Z Preprint model forecasts speed and cost of distributed AI training https://developinglight.com/studies/botscan-an-adaptive-active-probing-approach-for-identifying-live-iot-bot-84efcce4cd Developing Lighten 2026-08-28T19:45:25.827Z New botnet probe finds more live C2 servers at the same budget https://developinglight.com/studies/wide-field-opto-iontronic-iscat-mapping-of-interfacial-charging-and-elec-8ac72532aa Developing Lighten 2026-08-28T19:42:39.984Z Optical maps distinguish connected and isolated electrode regions https://developinglight.com/studies/markets-not-planners-decentralized-orchestration-of-llm-agents-with-priv-a3469ccfb3 Developing Lighten 2026-08-28T19:40:54.445Z AI agent market posts its best benchmark scores with subcontracting https://developinglight.com/studies/handling-mild-outliers-and-unobserved-values-in-compositional-datasets-u-9ca0c66904 Developing Lighten 2026-08-28T19:30:53.349Z Preprint tests model for missing data and mild outliers https://developinglight.com/studies/stieltjes-polynomial-interpolation-f93f6a1178 Developing Lighten 2026-08-28T19:27:07.692Z Preprint sets conditions for unique Stieltjes interpolation https://developinglight.com/studies/a-counterexample-to-nevanlinna-s-century-old-half-plane-problem-15d02c619e Developing Lighten 2026-08-28T19:24:37.116Z Half-plane example omits 0, 1 and infinity but is not in the Nevanlinna class https://developinglight.com/studies/real-time-inverse-solutions-via-neural-matrix-operators-47e9af17d4 Developing Lighten 2026-08-28T19:12:50.685Z NEMO shows major speedups in simulated inverse problems https://developinglight.com/studies/mote-mixture-of-task-experts-for-multi-task-video-understanding-077916e3cd Developing Lighten 2026-08-28T19:08:04.892Z AI video model routes tasks to experts in benchmark tests https://developinglight.com/studies/evolution-families-and-variation-of-constants-formula-for-abstract-funct-bc718817aa Developing Lighten 2026-08-28T19:06:04.074Z Study sets conditions for equations with infinite delay to have solutions https://developinglight.com/studies/the-paradox-of-strategic-altruism-26b797ca22 Developing Lighten 2026-08-28T19:02:33.331Z Preprint finds strategic altruism narrows implementable goals https://developinglight.com/studies/atmospheric-retrieval-of-combined-jwst-hst-and-ground-based-observations-25db5fc319 Developing Lighten 2026-08-28T18:56:47.359Z Analysis Favors Bare Rock on GJ 1132 b, but a Thin Atmosphere Is Possible https://developinglight.com/studies/moe-based-feature-adapter-for-prompt-free-binary-coronary-artery-segment-19b6c4e32c Developing Lighten 2026-08-28T18:53:46.446Z AI model posts top scores in coronary artery segmentation tests https://developinglight.com/studies/fedv-kgqa-multi-hop-question-answering-over-vertically-partitioned-knowl-06edfc2a39 Developing Lighten 2026-08-28T18:51:15.888Z AI system ranks multi-hop answers across split knowledge graphs https://developinglight.com/studies/browserforge-scaling-web-episode-via-parallel-browser-sandboxes-9e713ba415 Developing Lighten 2026-08-28T18:19:27.905Z BrowserForge training data linked to higher web-agent scores https://developinglight.com/studies/parameterized-complexity-of-l-p-lipschitz-constants-for-input-convex-neu-d95bfe90fe Developing Lighten 2026-08-28T18:15:57.310Z Exact Norm Maximization and ICNN Lipschitz Problems Stay Hard https://developinglight.com/studies/theory-as-data-compression-e479b1bb68 Developing Lighten 2026-08-28T18:11:41.547Z Economic model rankings differ under a data-compression test https://developinglight.com/studies/visualizing-and-quantifying-atomic-contributions-to-raman-intensities-go-893abe0be6 Developing Lighten 2026-08-28T18:07:55.636Z Raman intensity rankings switch between two peaks in three molecules https://developinglight.com/studies/simverity-when-does-simulated-agent-success-survive-physical-deployment-867c3c3075 Developing Lighten 2026-08-28T17:58:09.587Z Smart-Home Simulator Passes Did Not Match Every Physical Check https://developinglight.com/studies/nvexplain-explaining-time-series-forecasting-with-latent-trajectory-anal-e06338f2b7 Developing Lighten 2026-08-28T17:42:53.228Z Forecast-explanation method scores well, but results vary by dataset https://developinglight.com/studies/capra-scalable-healpix-native-intensity-reconstruction-for-high-resoluti-af2406a23e Developing Lighten 2026-08-28T17:39:22.521Z Capra recovers major IBEX sky patterns in a new preprint https://developinglight.com/studies/romnet-a-hybrid-reduced-order-modeling-and-machine-learning-approach-to-dc0b9e6a57 Developing Lighten 2026-08-28T17:35:36.639Z ROMNet tops neural baselines in wave-speed reconstruction tests https://developinglight.com/studies/complete-mappings-of-semigroups-ea2ac3e049 Developing Lighten 2026-08-28T17:33:05.987Z Mathematicians classify when finite semigroups have complete mappings https://developinglight.com/studies/barycentric-weak-inner-product-gromov-wasserstein-4ab617cc6d Developing Lighten 2026-08-28T17:30:20.442Z PBMC tests report higher ARI and NMI in weak-transport runs https://developinglight.com/studies/exploring-electron-transport-in-carbon-wire-systems-using-a-wave-functio-2eeb6c956a Developing Lighten 2026-08-28T17:26:34.670Z Carbon-wire simulations show model choices change electron transport https://developinglight.com/studies/extending-ground-constraint-lidar-imu-calibration-to-tilted-surfaces-in-28f4b24027 Developing Lighten 2026-08-28T17:19:48.690Z New sensor-calibration method handles tilted ground better https://developinglight.com/studies/entropy-stable-moving-wall-boundary-conditions-for-the-ale-formulation-o-1f85b54e5d Developing Lighten 2026-08-28T17:14:17.484Z Study reports conditional entropy stability in selected moving-wall tests https://developinglight.com/studies/snap-kg-streaming-node-assignment-via-projection-for-knowledge-graph-ent-64d4acb1e1 Developing Lighten 2026-08-28T16:59:00.559Z AI method assigns new knowledge-graph entities without retraining https://developinglight.com/studies/lightweight-machine-learning-driven-monocular-sidewalk-path-extraction-f-3372240861 Developing Lighten 2026-08-28T16:44:13.480Z Camera-image planner shows lower error and latency than BEV in scooter test https://developinglight.com/studies/separating-quantum-indistinguishability-obfuscation-from-falsifiable-ass-d2549f1d6e Developing Lighten 2026-08-28T16:38:12.272Z Preprint finds conditional barrier to quantum security proofs https://developinglight.com/studies/fast-rates-in-bayesian-online-learning-with-approximate-posteriors-99875812f1 Developing Lighten 2026-08-28T16:32:56.469Z Approximate Bayesian updates can preserve fast prediction rates https://developinglight.com/studies/fairness-aware-test-time-prompt-tuning-671286234e Developing Lighten 2026-08-28T16:24:55.456Z Fairness-Aware CLIP Tuning Reports Better Subgroup Scores https://developinglight.com/studies/evaluation-of-the-exradin-a30-parallel-plate-ion-chamber-as-a-reference-5172f82e5f Developing Lighten 2026-08-28T16:18:54.566Z Tests find Exradin A30 response varies in ultra-high-dose-rate electron beams https://developinglight.com/studies/d3er-supporting-multi-modal-recommendation-via-disentangle-and-distillat-69cefefd49 Developing Lighten 2026-08-28T15:49:21.660Z Preprint reports stronger results from a multimodal recommender https://developinglight.com/studies/a-constitutive-markov-physics-informed-neural-operator-mpno-for-autoregr-16f77b8f15 Developing Lighten 2026-08-28T15:46:20.989Z Preprint Reports MPNO Staying Bounded in Long Physics Rollouts https://developinglight.com/studies/learning-of-deep-neural-network-regression-estimates-using-gradient-desc-d0ede563b4 Developing Lighten 2026-08-28T15:40:49.891Z Pruned network nears optimal regression rates in theory https://developinglight.com/studies/hamiltonian-spectral-temporal-dissipative-dynamics-for-sequential-recomm-49d2c08f9a Developing Lighten 2026-08-28T15:33:18.845Z New sequential recommender tops most benchmark tests https://developinglight.com/studies/from-estimate-to-proof-certified-ground-state-energy-bounds-for-singular-7f008a37da Developing Lighten 2026-08-28T15:30:18.259Z Preprint reports certified energy range for hydrogen molecular ion https://developinglight.com/studies/beam-search-self-consistency-and-the-limits-of-inference-time-scaling-fo-137d5bc846 Developing Lighten 2026-08-28T15:16:31.241Z More inference compute rarely helped small AI models catch larger ones https://developinglight.com/studies/tost-a-tree-of-thought-socratic-teaching-framework-for-multi-path-guidan-99c58ccad5 Developing Lighten 2026-08-28T15:12:30.426Z AI maths tutor reports stronger guidance across several benchmarks https://developinglight.com/studies/waveop-litefm-lightweight-neural-operator-flow-matching-for-satellite-to-a7cae3211b Developing Lighten 2026-08-28T14:49:57.734Z Lightweight AI model uses less compute for satellite rain retrieval https://developinglight.com/studies/steer-the-sampling-not-the-kernel-grid-geometry-guided-sampling-operator-a7b9f5e9b3 Developing Lighten 2026-08-28T14:46:27.105Z GeoSample Reports Best Mean Metric Combination Among Tested Operators https://developinglight.com/studies/from-passive-response-to-proactive-correction-enhancing-llm-robustness-a-27536d5eef Developing Lighten 2026-08-28T14:25:25.064Z DEDUCE scores higher on benchmarks with misleading inputs https://developinglight.com/studies/lutseg-a-longitudinal-multi-expert-dataset-for-ulcer-tissue-segmentation-5ddafc63f8 Developing Lighten 2026-08-28T14:21:39.096Z Ulcer-image AI beats UniMatch-V2 in six of seven tests https://developinglight.com/studies/answer-is-cheap-show-me-the-evidence-augmenting-automated-vulnerability-6a214f93ce Developing Lighten 2026-08-28T13:39:34.489Z AI vulnerability tool leads benchmark, but evidence has limits https://developinglight.com/studies/livingrag-augmenting-graph-rag-with-experience-31d6f94716 Developing Lighten 2026-08-28T13:14:47.521Z AI system reuses reasoning with higher accuracy and lower estimated cost https://developinglight.com/studies/simulating-an-imaging-atmospheric-radio-telescope-to-observe-cosmic-gamm-d0d61c97f0 Developing Lighten 2026-08-28T13:06:16.316Z Preprint simulates radio telescope for gamma rays and neutrinos https://developinglight.com/studies/lost-but-not-erased-finding-traces-of-a-forgotten-language-in-neural-spe-784a2a3599 Developing Lighten 2026-08-28T13:03:45.285Z Preprint finds traces of a lost language inside speech models https://developinglight.com/studies/prefix-sliding-for-efficient-test-time-scaling-6527db8dc5 Developing Lighten 2026-08-28T12:59:44.472Z Preprint reports 3x faster AI inference in long-reasoning tests. https://developinglight.com/studies/fast-generative-grasping-via-lie-group-constrained-meanflow-0bb5194f33 Developing Lighten 2026-08-28T12:56:13.713Z Robot grasping model leads simulation tests with five evaluations https://developinglight.com/studies/swarmworld-stigmergic-technological-evolution-in-societies-of-language-m-c804c356b6 Developing Lighten 2026-08-28T12:37:11.507Z AI simulations show stronger portfolios, but solo search has the top artifact https://developinglight.com/studies/pipeline-native-transformers-co-designing-model-architecture-and-cpu-inf-bd04360668 Developing Lighten 2026-08-28T12:31:55.557Z Preprint reports a calculated twofold reduction in CPU decode bandwidth https://developinglight.com/studies/hypoforge-a-self-improving-multi-agent-framework-for-automated-hypothesi-66fa9393a7 Developing Lighten 2026-08-28T12:12:22.501Z AI system HypoForge tops reported generation and test-pass scores https://developinglight.com/studies/symmetry-origins-of-the-field-free-superconducting-diode-effect-in-the-k-439f2d8916 Developing Lighten 2026-08-28T12:09:20.936Z Field-free diode response seen in altered CsV3Sb5 devices