All papers

Reconstructs divergence-free 3D fluid velocity fields from sparse, irregular observations in a single forward pass using a set-to-field transformer.

Tests a hosted decision model to show that adding irrelevant options systematically shifts output probability distributions between top candidates.

Detects adversarial nodes on evolving graphs without fraud labels by measuring the consistency between node features and aggregated neighborhood contexts.

The paper proposes evaluating partial generation outputs to stop harmful model interactions early while avoiding unnecessary prompt rejections.

The paper presents TabWM, a tabular world model that infers system dynamics from past executions to predict database edits from actions.

The paper introduces InfoVis, a 9B multimodal model along with a benchmark and training dataset for slide image generation and editing.

The paper proposes an unsupervised drift monitor using internal embedding distances to flag unfamiliar inputs in deployed deepfake speech detectors.

The paper presents an empirical study evaluating syntactic language models at scales up to 1B parameters across linearizations and masks.

The paper presents multi-objective bandit algorithms to identify a small set of arms that minimizes degradation across multiple objectives.

The paper introduces a 30B tool-use model and an execution-grounded dataset to scale agentic tool supervision with verifiable environments.

The paper proposes Neural Bayesian Networks, replacing classical probability tables with differentiable tensor operations for scalable probabilistic reasoning on GPUs.

The paper analyzes legal eDiscovery review data to compare generative AI tools against continuous active machine learning and manual human review.

The paper analyzes how Cover Tree nearest-neighbor implementation details and query modes materially change runtime performance across datasets.

The paper investigates how fine-tuning language models on descriptions of user motivations influences their downstream trust and cooperative behavior.

The paper evaluates the impact of sarcasm on Turkish hate speech detection models using a targeted external test dataset.

The paper presents a structured step-by-step workflow designed to help non-programmer scientists implement and validate published machine learning models.

The paper proposes an offline reinforcement learning method that optimizes generative decision chunks in latent space using conditional velocity fields.

The paper proposes a manifold-aware graph diffusion network to generate valid functional brain connectomes while capturing fine-grained geometric structure.

The paper introduces ProbabilityBench to evaluate model uncertainty calibration against analytically computed, ground-truth instance-level probabilities across three text environments.

Introduces a framework that discovers and separates physical neural operators from unlabeled trajectory data containing multiple mechanisms.

Routes gradients from imperfect monitor flags into sacrificial parameters to mitigate reward hacking during reinforcement learning.

Shows that language agents often ignore identified cross-domain dependencies, causing unintended disruptive actions in network management.

Converts half-duplex speech models into streaming full-duplex conversational models capable of simultaneous listening and speaking.

Demonstrates that the scoring format chosen for an evaluation model strongly restricts its expressive capacity and ranking reliability.

Uses language-model perceptual descriptions of complex chemical mixtures to predict perceived odor similarity from molecular composition.

Discovers robot capabilities using a generative flow network that creates symbolic task skeletons grounded in physical robot affordances.

The paper describes a modular software backend architecture for extracting energy domain information and answering grounded questions from web sources.

The paper proposes information-theoretic tests to measure learned equivariance in intermediate neural network representations without knowing explicit group transformations.

The paper proposes clustering observations by feature importance profiles to incorporate latent factor representations into deep learning models.

The paper develops an approach to map AI semantic blind spots and evaluates the quality of distinct human ideas.

The paper presents a statistical mechanics framework using modern Hopfield models to analyze inference reliability and phase transitions in Transformer attention.

The paper presents a benchmark to evaluate whether formal specifications generated by language models accurately reflect natural-language requirements.

The paper proposes a hierarchical and attribute-aware supervision method to improve training efficiency and long-tail performance in generative recommendations.

The authors propose an evolutionary framework that optimizes instance coverage and updates language model weights to discover complementary optimization heuristics.

The paper introduces a test-time wrapper that distorts exposed output probabilities to defend black-box machine learning models from query-based adversarial attacks.

The authors introduce a topology-aware transformer meta-trained across multiple electricity markets to forecast day-ahead electricity prices for unseen grids.

The paper develops a sequential betting approach to reliably detect data distribution shifts in unstructured text and image streams over time.

The authors introduce a time-frequency diffusion framework that improves the generation of rare extreme events in time-series data.

The authors present a flight navigation policy for quadrotors that penalizes movement outside the camera view to avoid collisions in cluttered environments.

The authors formulate colon cancer treatment planning as an interactive reinforcement learning task in settings where actions are only partially verifiable.

The authors examine why language model agents fail by analyzing discrepancies between internal representations of state values and actual action choices.

The paper studies how much models contribute to multi-model teams, showing individual test scores do not correlate well with collaborative performance.

The paper evaluates seven machine learning and graph models on detecting misinformation within social media noise from conflict-related tweets.

The paper presents an in-context foundation model architecture for next-event prediction across diverse operational event logs.

The paper analyzes the failure modes of agent memory reuse and evaluates certified memory update strategies when facts change.

The paper investigates how indexed instruction wording affects the performance of lexical rankers for robot task retrieval in manipulation benchmarks.

The paper introduces an evaluation method that separates question answerability detection from answer span extraction in radiology report question answering.

The paper analyzes why dynamic graph neural networks fail on evolving nodes and introduces a change-aware classification framework.

The paper introduces a dataset and benchmark to evaluate vision-language models on recognizing and answering questions about Philippine cultural heritage sites.

The paper introduces a morphology-aware contrastive learning framework to predict five-year colorectal cancer survival from whole-slide pathology images.

The paper evaluates judge labels on agent steps by restarting failed trajectories from intermediate states to identify true failure points.

Presents SADDER, an inference-time text diffusion safety framework using directional embedding signals to steer unsafe tokens toward benign distributions.

Introduces DeltaTTG, a method using predicted remaining token count as a progress signal to allocate advantages in multi-turn reinforcement learning.

Introduces LatentRiskBench to evaluate safety risks in enterprise agents where action safety depends on latent workflow context.

Proposes MedEvo, an iterative training framework that repairs multimodal medical reasoning trajectories using minimal hints and taxonomy-guided feedback.

Introduces the PocketWatch benchmark and zPocket, an approach training pocket-prediction heads on frozen representations from protein co-folding models.

Presents LeanPRM, a method using the Lean proof assistant to generate step-level correctness labels for training natural-language mathematical proof verifiers.

Studies the trade-off between spatial smoothness and gene-level fidelity when mapping single-cell RNA sequencing profiles to images for neural networks.

Develops SCRIM, a method quantifying state criticality in imitation learning by measuring downstream value degradation under bounded action perturbations.

Proposes a recursive self-improvement framework for electrocardiogram analysis models combining verified synthetic waveform edits with targeted intervention selection.