11583 papers
This paper analyzes the asymptotic performance of best-of- LLM ensembling via majority voting as , proposing an adaptive generation scheme to efficiently allocate inference budgets and an optimal weighted ensemble method formulated as a mixed-integer linear program to outperform individual models.
The paper proposes CAD-Tokenizer, a framework that employs modality-specific tokenization via a sequence-based VQ-VAE to overcome the limitations of standard LLM tokenizers, thereby significantly enhancing the quality and instruction-following capabilities of unified text-guided CAD prototyping.
This paper proposes a lightweight, interpretable approach where reasoning-capable LLMs act as agents to induce decision trees for small tabular datasets, achieving competitive performance with state-of-the-art black-box models while offering human-readable reasoning traces and the ability to incorporate fairness and monotonicity constraints.
This paper proposes a scalable second-order cubic-regularized Riemannian Newton algorithm for -means clustering that reformulates the problem as a smooth unconstrained optimization on a product manifold, enabling linear-time subproblem solutions and achieving faster convergence with optimal statistical accuracy compared to state-of-the-art first-order methods.
This paper introduces Ssiuu, a novel unlearning method that employs attribution-guided regularization to eliminate spurious neurons and ensure the faithful, robust removal of sensitive knowledge from large language models, thereby preventing its resurfacing during subsequent retraining.
This paper argues that mainstream Class Incremental Learning evaluation protocols are biased due to insufficient sequence sampling, and proposes EDGE, a new protocol that leverages inter-task similarity to identify extreme sequences for accurately characterizing the full performance distribution.
Inspired by biological neural mechanisms, the paper proposes Uni-NTFM, a unified foundation model that integrates heterogeneous feature projection, topological embeddings, and a Mixture-of-Experts Transformer to achieve superior generalization across diverse EEG tasks through alignment with the brain's sparse coding and geometric topology.
This paper benchmarks eight ECG foundation models across 26 clinical tasks, revealing that architectural choices like ECG-CPC's compact state-space design often outweigh massive scale in performance and label efficiency, while highlighting significant remaining gaps in cardiac structure and outcome prediction.
This paper introduces the Online Learning in the Replay Setting to model systems training on self-annotated data, establishing the Extended Threshold dimension as the exact measure of learnability and proving that while proper learners may fail catastrophically, specific improper algorithms can achieve optimal mistake bounds against replay adversaries.
This paper presents improved algorithms for various kernel matrix linear algebra tasks, such as matrix-vector products and spectral norm computation, by leveraging kernel density estimation to achieve faster runtimes with reduced dependencies on the number of data points and error tolerance compared to existing methods.
FLOWR.root is an SE(3)-equivariant flow-matching foundation model that unifies structure-aware 3D ligand generation with multi-purpose affinity prediction and confidence estimation, achieving state-of-the-art performance through mixed-fidelity training and parameter-efficient finetuning for efficient, high-quality drug design.
The paper proposes Cell-Mechanistic Neural Networks (Cell-MNN), an end-to-end encoder-decoder architecture that utilizes locally linearized ODEs to efficiently model single-cell differentiation dynamics and explicitly learn interpretable, biologically consistent gene interactions, outperforming current state-of-the-art methods in scalability and interpretability.
The paper proposes ELMUR, a transformer architecture equipped with structured external memory that enables robotic agents to effectively handle long-horizon, partially observable tasks by significantly extending effective horizons and outperforming existing baselines on synthetic and real-world manipulation benchmarks.
This paper introduces Value Flows, a distributional reinforcement learning method that employs flexible flow-based models and a novel flow-matching objective to estimate full future return distributions and identify high-uncertainty states, achieving superior performance across diverse benchmarks compared to prior approaches.
This paper proposes a novel geometric framework that models LLM reasoning as smooth flows in representation space, demonstrating through empirical experiments that next-token prediction enables models to internalize logical invariants as higher-order geometry, thereby challenging the "stochastic parrot" hypothesis and suggesting a universal representational law underlying machine understanding.
This paper introduces ToMCLIP, a topology-aware framework that enhances multilingual vision-language alignment by applying persistent homology to preserve the global geometric structure of shared embedding spaces, thereby improving zero-shot accuracy and retrieval performance compared to existing instance-level methods.
This paper introduces Gym-TORAX, an open-source Python package that bridges Reinforcement Learning algorithms with the TORAX plasma control simulator by automatically generating Gymnasium-compatible environments for optimizing tokamak performance and stability, currently featuring an ITER ramp-up scenario.
This paper introduces WeightLens and CircuitLens, two complementary methods that advance mechanistic interpretability by analyzing feature weights and component interactions directly, thereby overcoming the limitations of activation-based approaches in scalability, robustness, and the ability to capture circuit-level dynamics without relying on external explainer models or datasets.
This paper introduces COGS, a data-efficient framework that synthesizes large-scale reasoning datasets by decomposing seed questions into primitive factors and recomposing them with new images, thereby significantly enhancing the visual reasoning capabilities of multi-modal large language models in annotation-scarce domains like charts and webpages.
This paper reveals that the reliability of Mahalanobis-based out-of-distribution detection is highly dependent on the geometric properties of the feature space, specifically within-class spectral structure and local intrinsic dimensionality, and proposes a radially scaled normalization method that dynamically adjusts feature radii to optimize detection performance based on these geometric signals.