Large language models display the identifiable victim effect at roughly twice the human baseline, strongly amplified by instruction tuning and chain-of-thought prompting but inverted by reasoning-specialized models.
Mixed citations
Journal of Optimization Theory and Applications 102(2), 345–371 (1999)
Mixed citation behavior. Most common role is background (65%).
90
Pith papers citing it
Background
65% of classified citations
citation-role summary
background 14
method 4
dataset 1
extension 1
citation-polarity summary
claims ledger
- background Rosner, R., Tucker, W.H., Vaiana, G.S.: Dynamics of the quiescent solar corona. Astrophys. J. 220, 643-645 (1978) https://doi.org/10.1086/155949 Rimmele, T.R., Warner, M., Keil, S.L., Goode, P.R., Knölker, M., Kuhn, J.R., Rosner, R.R., McMullin, J.P., Casini, R., Lin, H., al.: The Daniel K. Inouye Solar Telescope - Observatory Overview. Solar Phys. 295(12), 172 (2020) https://doi.org/10.1007/ 29 s11207-020-01736-7 Sasso, C., Andretta, V., Spadaro, D., Susino, R.: Solar low-lying cool loops and t
- background Recent work has also utilised large language models as recommenders [100, 136]. 1.2.1.3 Evaluation of Recommender Systems At a high level, recommender system evaluation protocols can be categorised into three: offline evaluation, user study, and online evaluation [54, 91, 248]. Offline eval- uation is done on a pre-collected dataset that contains the users' past interactions (e.g., ratings [94], listening history [196]), with the purpose of collecting quantitative data for predictive performance
co-cited works
representative citing papers
A Lean 4 formalization of Gröbner basis theory that uniformly handles arbitrary and infinite numbers of variables with verified division, Buchberger criterion, and reduced bases.
Develops a model-independent energy-based CutFEM formulation for finite-strain elasticity using automatic differentiation, with cut-independent stability analysis and convergence results for both smooth solutions and corner singularities.
SPIDeC methods achieve arbitrarily high-order accuracy for positive dynamical systems while unconditionally preserving positivity and equilibria via a multiplicative Volterra structure, and they are L-stable with asymptotic logarithmic contractivity under Gauss-Radau nodes.
A Gaussian process surrogate gate inserted between generative crystal models and property oracles matches or exceeds ungated fine-tuning while using roughly one-fifth the oracle calls for heat capacity and bulk modulus.
Compartmentalization of a single morphogen into intra- and extracellular fields with nonlinear coupling produces diffusion-driven instabilities enabling Turing patterns.
ReMax achieves the first sublinear regret bound for Gaussian rewards at M=2 by characterizing the optimal sampling distribution via an expected-improvement balance condition and separating saturation from underestimation effects.
A translation-invariant Banach space is constructed on which the non-classical maximal operator M^diamond is bounded but the sharp maximal operator M^sharp is not.
Construction of a tree-like fractal Dirichlet space separating weak and strong elliptic Harnack inequalities under a chosen self-similar measure.
GHGbench supplies a harmonized dataset and multi-task benchmark for company and building carbon emission prediction, with baselines showing large OOD gaps and benefits from multimodal embeddings.
dARL supplies a sound deductive refinement calculus with trace semantics for verifying and simplifying differential-algebraic programs, shown complete for index reduction certification.
An SMT-based active learning algorithm learns minimal nondeterministic weighted automata over arbitrary semirings, with partial correctness proofs, a sufficient termination condition, and experiments showing smaller models and fewer queries than baselines.
A complete linear inequality description and volume formula are derived for the convex hull of the graph of a monomial on a nonnegative box with at most one positive lower bound.
BRIDGE creates the first formal heterogeneous multi-dataset benchmark for IoT botnet detection with LODO evaluation, and TCH-Net achieves mean LODO F1 of 0.5577 while reaching F1 0.8296 on standard tests, outperforming twelve baselines.
WoodelfHD reduces Background SHAP preprocessing for decision trees from 3^D to 2^D complexity, enabling exact computation on depths up to 21 with reported speedups of 33x to 162x.
Refutational normalization reconciles efficiency and completeness for JSON Schema inclusion checking.
Metacognitive self- and co-regulation loops improve LLM agent performance in engineering design by mitigating fixation and enabling better exploration of design options.
A new angular-momentum-preserving dissipative force model for point-mass N-body problems reduces central-configuration orbits to an equivalent dissipative two-body system whose phase-space topology is mapped via Poincaré compactification, and whose averaged Keplerian equations show no effect on peri
IdMAT and LIndA enable learning any regular language consistent with an incomplete inductive teacher by reducing uncertainties to incremental SAT solving for tasks like language separation and invariant synthesis.
Logic-based Weisfeiler-Leman variants enable graph-to-table conversion for classification that matches GNN and graph transformer accuracy while running 5-20x faster without GPUs.
Derives generalization bounds for quantum learning via quantum and classical Rényi divergences, with a new modified sandwich quantum Rényi divergence shown to outperform the Petz version analytically and numerically.
Order parameters for unitary ensembles solve the modified KP equation via the Volterra lattice while orthogonal ensembles yield a new integrable chain via the Pfaff lattice, with thermodynamic limits being hydrodynamic-type systems solved by the same semi-discrete dynamical chain.
GNTC satisfiability is 2ExpTime-complete and model checking is P^NP[O(log² n)]-complete via polynomial and exponential reductions to UNTC and 2-way alternating parity tree automata.
Multitask fine-tuning of an encoder-decoder model on prompted datasets produces zero-shot generalization that often beats models up to 16 times larger on standard benchmarks.
citing papers explorer
-
Narrative over Numbers: The Identifiable Victim Effect and its Amplification Under Alignment and Reasoning in Large Language Models
Large language models display the identifiable victim effect at roughly twice the human baseline, strongly amplified by instruction tuning and chain-of-thought prompting but inverted by reasoning-specialized models.
-
Formalizing Gr\"obner Basis Theory in Lean
A Lean 4 formalization of Gröbner basis theory that uniformly handles arbitrary and infinite numbers of variables with verified division, Buchberger criterion, and reduced bases.
-
A Unified CutFEM Formulation for Finite-Strain Elasticity: Energy Minimisation and Corner Singularities
Develops a model-independent energy-based CutFEM formulation for finite-strain elasticity using automatic differentiation, with cut-independent stability analysis and convergence results for both smooth solutions and corner singularities.
-
Stable Positive Integral Deferred Correction Methods for Positive Dynamical Systems
SPIDeC methods achieve arbitrarily high-order accuracy for positive dynamical systems while unconditionally preserving positivity and equilibria via a multiplicative Volterra structure, and they are L-stable with asymptotic logarithmic contractivity under Gauss-Radau nodes.
-
Surrogate-Gated Generation and Foundation-Model Embeddings for Bayesian Materials Design
A Gaussian process surrogate gate inserted between generative crystal models and property oracles matches or exceeds ungated fine-tuning while using roughly one-fifth the oracle calls for heat capacity and bulk modulus.
-
Single-morphogen Turing instability driven by nonlinear intracellular-extracellular coupling
Compartmentalization of a single morphogen into intra- and extracellular fields with nonlinear coupling produces diffusion-driven instabilities enabling Turing patterns.
-
Finite-Time Regret Analysis of Retry-Aware Bandits
ReMax achieves the first sublinear regret bound for Gaussian rewards at M=2 by characterizing the optimal sampling distribution via an expected-improvement balance condition and separating saturation from underestimation effects.
-
A Banach space that distinguishes two maximal operators
A translation-invariant Banach space is constructed on which the non-classical maximal operator M^diamond is bounded but the sharp maximal operator M^sharp is not.
-
A tree-like fractal Dirichlet space lying between strong and weak elliptic Harnack inequalities
Construction of a tree-like fractal Dirichlet space separating weak and strong elliptic Harnack inequalities under a chosen self-similar measure.
-
GHGbench: A Unified Multi-Entity, Multi-Task Benchmark for Carbon Emission Prediction
GHGbench supplies a harmonized dataset and multi-task benchmark for company and building carbon emission prediction, with baselines showing large OOD gaps and benefits from multimodal embeddings.
-
A Deductive Refinement Calculus for Differential-Algebraic Programs
dARL supplies a sound deductive refinement calculus with trace semantics for verifying and simplifying differential-algebraic programs, shown complete for index reduction certification.
-
SMT-Based Active Learning of Weighted Automata
An SMT-based active learning algorithm learns minimal nondeterministic weighted automata over arbitrary semirings, with partial correctness proofs, a sufficient termination condition, and experiments showing smaller models and fewer queries than baselines.
-
On the convex hull of the graph of a simple monomial
A complete linear inequality description and volume formula are derived for the convex hull of the graph of a monomial on a nonnegative box with at most one positive lower bound.
-
BRIDGE and TCH-Net: Heterogeneous Benchmark and Multi-Branch Baseline for Cross-Domain IoT Botnet Detection
BRIDGE creates the first formal heterogeneous multi-dataset benchmark for IoT botnet detection with LODO evaluation, and TCH-Net achieves mean LODO F1 of 0.5577 while reaching F1 0.8296 on standard tests, outperforming twelve baselines.
-
WOODELF-HD: Efficient Background SHAP for High-Depth Decision Trees
WoodelfHD reduces Background SHAP preprocessing for decision trees from 3^D to 2^D complexity, enabling exact computation on depths up to 21 with reported speedups of 33x to 162x.
-
JSON Schema Inclusion through Refutational Normalization: Reconciling Efficiency and Completeness
Refutational normalization reconciles efficiency and completeness for JSON Schema inclusion checking.
-
Supervising Ralph Wiggum: Exploring a Metacognitive Co-Regulation Agentic AI Loop for Engineering Design
Metacognitive self- and co-regulation loops improve LLM agent performance in engineering design by mitigating fixation and enabling better exploration of design options.
-
An angular-momentum preserving dissipative model for the point-mass N -body problem
A new angular-momentum-preserving dissipative force model for point-mass N-body problems reduces central-configuration orbits to an equivalent dissipative two-body system whose phase-space topology is mapped via Poincaré compactification, and whose averaged Keplerian equations show no effect on peri
-
Automata Learning with an Incomplete but Inductive Teacher (Technical Report)
IdMAT and LIndA enable learning any regular language consistent with an incomplete inductive teacher by reducing uncertainties to incremental SAT solving for tasks like language separation and invariant synthesis.
-
Graph Learning via Logic-Based Weisfeiler-Leman Variants and Tabularization
Logic-based Weisfeiler-Leman variants enable graph-to-table conversion for classification that matches GNN and graph transformer accuracy while running 5-20x faster without GPUs.
-
Generalization Bounds for Quantum Learning via R\'enyi Divergences
Derives generalization bounds for quantum learning via quantum and classical Rényi divergences, with a new modified sandwich quantum Rényi divergence shown to outperform the Petz version analytically and numerically.
-
Random matrix ensembles and integrable differential identities
Order parameters for unitary ensembles solve the modified KP equation via the Volterra lattice while orthogonal ensembles yield a new integrable chain via the Pfaff lattice, with thermodynamic limits being hydrodynamic-type systems solved by the same semi-discrete dynamical chain.
-
Guarded Negation Transitive Closure Logic
GNTC satisfiability is 2ExpTime-complete and model checking is P^NP[O(log² n)]-complete via polynomial and exponential reductions to UNTC and 2-way alternating parity tree automata.
-
Multitask Prompted Training Enables Zero-Shot Task Generalization
Multitask fine-tuning of an encoder-decoder model on prompted datasets produces zero-shot generalization that often beats models up to 16 times larger on standard benchmarks.
-
Compositionality of Global Dynamics in Product and Skew-Product Systems
Proves an isomorphism between attractor lattices of product flows and coproducts of component lattices via Conley theory, with extensions to skew-products for dynamical decomposition.
-
A Stationary-Distribution Theory for Triplet-Based Plateau Search in Random Forest Ensemble-Size Selection
Derives the stationary distribution and asymptotic scaling O(ε^{-2}) for ensemble size in a Markov chain model of triplet-based plateau tuning for random forests.
-
Faster than Fast-LTS: Robust Regression and Outlier Detection with DC Programming
Proposes sBDCA with preconditioning for the LTS estimator, claiming up to 3.25 times faster runtime and up to 90% lower objective values than Fast-LTS on synthetic and real data.
-
A Neural Surrogate Approach for Simulating Natural Convection Problems
A Fourier neural operator trained on Boussinesq-compressible simulation pairs corrects Boussinesq predictions for natural convection, achieving SSIM near unity and MSE reductions of one to three orders of magnitude.
-
Feasibility-driven QAOA with penalty scheduling
Introduces Λ-lr-QAOA and piecewise-ramp QAOA that promote penalty schedules to variational parameters and use a feasibility-driven loss on budget-constrained MWIS satellite planning instances.
-
Curvature-Adaptive Consistency Flow Matching: Autonomous Trajectory Optimization via Reinforcement Learning
CACFM applies RL to adaptively select critical regions in probability flow ODE trajectories for consistency distillation, yielding SOTA few-step results on FLUX and SDXL.
-
Letting Homogeneity Entropy Select S-Pairs in Buchberger's Algorithm
Homogeneity Entropy S-pair selection outperforms classical Degree/Normal/Sugar heuristics on random polynomial systems but underperforms on the PHCpack real-world benchmark set.
-
An Octahedral Fibrous Constitutive Model for Heart Valve Mechanics and Function
Develops and calibrates an anisotropic hyperelastic constitutive model based on octahedral fiber networks for mitral valve leaflets to examine effects of anisotropy and reorientation on closure and mitral regurgitation.
-
Hybrid Adaptive Kalman Filtering for Data-Efficient Joint Tracking and Classification
Self-supervised hybrid adaptive Kalman filter learns structured corrections for data-efficient joint tracking and classification.
-
Convergence Analysis of Evolution Strategies for Mixed-Integer Optimization
Drift analysis on a mixed-integer benchmark shows (1+1)-LB-ES risks premature convergence with large numbers of integer variables while (1+1)-LUB-ES achieves linear convergence after integers are fixed under suitable bounds.
-
An $L^{\#}$ Based Algorithm for Active Learning of Minimal Separating Automata
An L#-inspired active learning algorithm learns minimal separating DFAs for disjoint languages when one exists and outperforms prior methods on random and industrial benchmarks.
-
Weierstrass semigroups at totally ramified places of degree one on Kummer extensions
Explicit descriptions of Weierstrass semigroups and gaps at totally ramified places on Kummer extensions, plus symmetry conditions and minimal generators, with applications to GGS and BM curves.
-
Scenario generation of intraday electricity price paths for optimal trading in continuous markets
A kernel-based regression model plus scenario generation from forecast errors and a new Support Vector Sorting step produces ensemble price trajectories that improve both statistical accuracy and trading profits over benchmarks on German intraday continuous market data.
-
Trapping Attacker in Dilemma: Examining Internal Correlations and External Influences of Trigger for Defending GNN Backdoors
PRAETORIAN reduces GNN backdoor attack success rate to 0.55% with 0.62% clean accuracy drop by targeting the need for many or highly influential trigger nodes.
-
A Scaling Limit of Random Walks in the Rational Adeles
Adele-valued random walks converge weakly to an adelic Lévy process in the J1 Skorokhod topology after appropriate scaling.
-
Trident: Improving Malware Detection with LLMs and Behavioral Features
Trident combines static decision trees, LLM-generated behavioral rules from sandbox reports, and direct LLM analysis via majority voting to outperform static methods while resisting concept drift without retraining.
-
Do Larger Models Really Win in Drug Discovery? A Benchmark Assessment of Model Scaling in AI-Driven Molecular Property and Activity Prediction
Benchmark across 78 endpoint-split entries finds classical ML winning 47.4% of best performances over pretrained models, GNNs, and LLMs, with performance depending on model-task-split fit rather than scale.
-
Skew-orthogonal polynomials for a quartic Freud weight: two classes of quasi-orthogonal polynomials
Skew-orthogonal polynomials for a quartic Freud weight are written as linear combinations of orthogonal polynomials via new recursive coefficient relations; even and odd degrees separately become quasi-orthogonal families for two semi-classical Laguerre weights, with the first closed recursions that
-
Left-invariant harmonic spinors on three-dimensional Lie groups
Equivalent conditions on the structure equations of 3D Lie algebras determine which left-invariant pseudo-Riemannian metrics admit left-invariant harmonic spinors.
-
Simultaneous Fragment Docking for Geometrically Linkable Pose Pairs
Q-SFD, a QUBO formulation for simultaneous fragment docking with an added inter-fragment distance term, approximately doubles top-1 recovery of reconstruction-feasible pose pairs and places at least one feasible pair in the top-5 for over 90% of benchmark cases without losing pose accuracy.
-
Resource-efficient quantum approximate optimization algorithm via Bayesian optimization and maximum-probability evaluation
QAOA with max-probability bitstring cut value objective, Bayesian optimization, and dual-criteria adaptive shots matches conventional MaxCut quality while using fewer total measurements.
-
Evidence of an Emergent "Self" in Continual Robot Learning
Continual learning robots form a significantly more stable invariant subnetwork than constant-task controls, and preserving it improves adaptation while damaging it hurts performance.
-
EarthSight: A Distributed Framework for Low-Latency Satellite Intelligence
EarthSight reduces average compute time per image by 1.9x and 90th-percentile end-to-end latency from 51 to 21 minutes by distributing inference decisions between orbit and ground with shared backbones and early rejection filters.
-
Generalizing Reduced Rank Extrapolation to Low-Rank Matrix Sequences
Two generalizations of reduced rank extrapolation are derived for low-rank matrix sequences and iteration-dependent mapping functions, with numerical tests on Lyapunov and Riccati equations.
-
Toward a Standard Interface for User-Defined Scheduling in OpenMP
The paper identifies core components for user-defined scheduling in OpenMP and proposes two competing interfaces as candidates for standardization across C, C++, and Fortran.
-
On Surrogate Modeling of Static Response of AM Short-Fiber Thermoplastics Using Graph Neural Networks
A GNN-LSTM surrogate trained on Voronoi-cell homogenized nonlinear FE data predicts unseen SFT microstructure responses with R²≈0.98 and >100x speedup over direct FE.