2294 papers
This paper introduces TableEG, a framework that leverages fine-tuned large language models to generate authentic, distribution-aligned synthetic errors in tabular data, thereby addressing the scarcity of real-world error datasets and establishing a robust benchmark for evaluating data cleaning techniques.
This paper identifies a pervasive "agreement bias" in Multimodal LLM verifiers that causes them to over-validate agent behavior, and proposes a lightweight Self-Grounded Verification (SGV) method that significantly improves failure detection and task completion across web navigation, computer use, and robotics by decoupling prior generation from trajectory evaluation.
This paper presents the first comprehensive survey of flow matching applications in biology and life sciences, systematically reviewing its theoretical foundations and categorizing its recent advancements in biological sequence modeling, molecule design, and protein generation.
This paper proposes a tree-based Weak-to-Strong Generalization framework that leverages Monte Carlo Tree Search to organize both successful and failure trajectories from weak models, thereby significantly enhancing the reasoning and decision-making capabilities of strong models in complex interactive environments.
This paper investigates how malicious auditees can construct fairness-compliant yet representative-looking samples from non-compliant distributions to deceive auditors, formalizes these manipulation strategies using optimal transport and entropic projections, and proposes statistical tests to detect such distributional manipulation attacks.
This paper introduces a Dynamic, Automatic, and Systematic (DAS) red-teaming framework that exposes a critical "Benchmarking Gap" in medical large language models, revealing that despite high static benchmark scores, most models exhibit profound brittleness, privacy leaks, bias, and hallucinations when subjected to continuous, adversarial stress-testing.
The paper introduces CauKer, a novel algorithm that combines Gaussian Process kernel composition with Structural Causal Models to generate diverse, causally coherent synthetic time series, enabling sample-efficient pre-training of classification foundation models that exhibit clear scaling laws across varying dataset sizes and model capacities.
GraphProp is a two-phase framework for training graph foundation models that first learns structural generalization by predicting graph invariants and then leverages these representations as positional encodings to enhance cross-domain performance in graph-level tasks, particularly outperforming existing methods in scenarios with limited data or missing node attributes.
This paper demonstrates that gating mechanisms in recurrent neural networks act as data-driven preconditioners that couple state-space time-scales with parameter-space dynamics, inducing lag-dependent and anisotropic effective learning rates that complement optimizer-driven adaptivity to enhance trainability.
The paper introduces ECHO, a novel foundation model that leverages band-split architecture and frequency positional embeddings to achieve state-of-the-art performance in anomaly detection and fault classification across variable-length, arbitrary sampling rate machine signals without requiring padding or cropping.
This paper presents an inverse dynamic game algorithm that uses mixed-integer linear programs to learn parametric constraints from multi-agent interaction demonstrations by encoding Karush-Kuhn-Tucker conditions, thereby providing theoretical guarantees for recovering inner approximations of safe and unsafe sets to enable robust motion planning.
This paper proposes CbLDM, a Condition-based Latent Diffusion Model that utilizes conditional priors and Laplacian matrices to effectively and stably recover the nanostructures of monometallic nanoparticles from their atomic pair distribution functions, addressing the highly ill-posed nature of the inverse problem.
This paper proposes a unified training framework that combines entropy-driven curriculum learning, which sequences training from simple to complex trajectories based on Lempel-Ziv compression, with multi-task learning to simultaneously optimize location, distance, and direction predictions, thereby achieving state-of-the-art performance and significantly faster convergence in human mobility prediction.
This paper benchmarks Multivariate Imputation by Chained Equations (MICE) against deep generative models like Variational Autoencoders and Conditional Tabular GANs for synthetic ratemaking data, finding that MICE offers a simpler yet high-fidelity alternative that effectively preserves statistical distributions and supports robust Generalized Linear Model training.
This paper proposes the F²SA- method, which utilizes -th order finite differences to achieve a nearly optimal complexity for finding -stationary points in stochastic bilevel optimization with highly smooth objectives, thereby improving upon previous first-order bounds and matching the fundamental lower limit.
Through large-scale experiments with over 1.5 million LLM-generated behavioral sequences, this paper reveals a fundamental asymmetry in behavioral inference where agent motivations are nearly perfectly recoverable while belief systems remain largely opaque due to inherent information-theoretic limits and architectural constraints, particularly within a "neutral zone" of behavioral ambiguity.
This paper introduces a modular, multimodal framework that leverages generative AI to synthesize realistic residential building data from public images and information, thereby overcoming data accessibility and privacy barriers to advance energy modeling and machine learning research.
The paper introduces PANO, a physics-aware neural operator that performs direct, single-pass inversion of raw sensor data into high-quality 3D photoacoustic images, outperforming traditional algorithms and enabling real-time reconstruction across diverse sparse acquisition settings to facilitate the clinical translation of 3D PACT.
This paper presents a fast, scalable method that combines deep learning with Dynamic Input Conductances (DICs) to reconstruct diverse, degenerate populations of conductance-based neuron models directly from spike times, effectively bridging experimental recordings and mechanistic biophysical parameters.
This paper introduces MICA, a privacy-preserving, speech-interactive multi-agent system that leverages Adaptive Step Fusion and a safety-audited coordination topology to deliver robust, real-time industrial assistance for assembly and maintenance tasks on resource-constrained hardware.