Investigating Faithfulness in Large Audio Language Models
Framework evaluating faithfulness of chain-of-thought reasoning in large audio language models for multimodal tasks.
Framework evaluating faithfulness of chain-of-thought reasoning in large audio language models for multimodal tasks.
Flow-matching models for 3D point cloud generation using optimal transport and meanflow for single-step inference acceleration.
KAN-based feature selection framework for tabular data via spline-based importance scoring. Specialized ML technique.
Sub-quadratic attention algorithm removing bounded-entry restrictions for LLM inference speedup. Foundational LLM efficiency research.
Quantization technique for vision encoders using prefix registers to handle outliers. Optimization research for multimodal models.
Theoretical reinforcement learning on decision-estimation coefficients for adversarial MDPs. Pure RL theory.
Conditional flow matching for precipitation forecasting. Weather prediction ML, not core AI interests.
Diffusion-Transformer model converting images directly to G-code for 3D printing. Applied ML, domain-specific.
Genetic algorithm for sample reweighting to mitigate ML bias. Fairness-focused, not primary tech interests.
Continual learning research on replay buffer size impact on feature retention vs. classifier forgetting. Specialized ML theory.
Neural ODEs for quantum many-body dynamics simulation. Physics-focused ML, not core AI interests.
Algorithm extraction from Discrete Transformers via symbolic program synthesis. Addresses representation entanglement in interpretability.
EEG foundation model for brain-computer interfaces with biophysical grounding. Neuroscience domain, not AI/tech stack focused.
Research analyzing mechanistic changes when post-training autoregressive models into masked diffusion models. Studies model internals via circuit analysis.
Machine learning for materials science: multimodal models predict dielectric elastomer properties under limited data. Domain-specific ML, not AI-focused.
Unified theoretical framework for model merging explaining effectiveness across heterogeneous fine-tuning hyperparameters with scaling laws.
Mixed-precision training and compilation techniques for RRAM-based computing-in-memory ML accelerators with low bit-width constraints.
Application of sheaf neural networks to biomedical problems comparing performance against GCNs, GATs, and GraphSage.
Continuous-time Koopman autoencoder for surrogate modeling of time-dependent PDEs in fluid dynamics.
Krause Attention: principled attention mechanism addressing representation collapse and attention sink issues in transformers.
Multi-scale retrieval benchmark for time series language models addressing long-context temporal localization under computational constraints.
Position paper on causal inference requirements for valid and generalizable interpretability claims in LLM research.
Deep reinforcement learning stability improvement through isotropic Gaussian embeddings under non-stationary training dynamics.
Benchmark comparing state space models, transformers, and RNNs for US power grid electricity demand forecasting.
Graph neural network approach for spatial allocation in energy system coupling with mismatched resolutions.
CeRA: improved parameter-efficient fine-tuning method that surpasses LoRA's linear constraints via manifold expansion with gating and dropout.
Analysis of transformer training trajectories under AdamW showing low-dimensional drift directions and batch-gradient alignment patterns.
Explainable AI method for highlighting token attributions in text classification using transformers.
Framework for autonomous neural architecture and hyperparameter search using self-evaluating RL agents without human supervision.
Research on robust policy training in partially observable reinforcement learning under adversarial latent state distribution shifts.
Theoretical analysis connecting drifting models and score-based generative models through kernel-based transport discrepancy.
Systematic study of jailbreak attack scaling laws across LLM methods and model families using compute-bounded optimization framework.
Research on parameter-efficient fine-tuning for continual learning using representation-level optimization instead of weight-level black-box methods.
Zero-shot surgical duration prediction combining retrieval-augmented LLMs with Bayesian averaging for resource management.
Physics-informed autoencoder with frozen PDE solver for tracking continuum mechanics dynamics in video.
Survey of privacy-preserving machine learning mechanisms for IoT devices covering federated learning and edge computing approaches.
Analysis of transformer training dynamics via Spectral Edge Dynamics, identifying coherent optimization directions vs stochastic noise.
Virtual cell perturbation prediction model using optimal transport for in silico experimentation on genetic/chemical perturbations.
Diffusion-based reinforcement learning policy using flow matching with direct entropy regularization and efficient gradient computation.
Comprehensive review of AI methods in fashion including aesthetics, personalization, virtual try-on, and forecasting.
Novel framework for estimating heterogeneous causal contrasts combining T-learning and DR-learning approaches.
Deep learning framework to compensate for perception latency in vision-based autonomous vehicle lane-keeping.
Detects hallucinations in virtually-stained histology using latent space analysis and neural precursor method.
Method for evaluating synthetic chest X-ray quality using embedded characteristic scores.
Universal sparse autoencoders for discovering and aligning interpretable concepts across multiple neural networks.
Multifidelity simulation-based inference framework for parameter estimation with expensive simulators.
Algorithm for finding game equilibria under differential privacy constraints in polymatrix games.
Survey of AI-based methods for detecting and mitigating distributed denial-of-service attacks.
Ensemble of language models for automated tumor classification in cancer registry pathology reports.
Physics-informed neural networks for learning transferable friction models in robotics simulation.