Ablate and Rescue: A Causal Analysis of Residual Stream Hyper-Connections
Causal analysis of residual stream hyper-connections in multi-stream transformer architectures exploring mechanistic interpretability.
Causal analysis of residual stream hyper-connections in multi-stream transformer architectures exploring mechanistic interpretability.
SafeDriver-IQ framework transforming binary crash prediction to continuous risk quantification for driver safety feedback.
Continual learning framework for toxicity detection adapting to evolving evasive perturbations in online content.
Multimodal weather foundation model combined with satellite data for fine-grained solar irradiance forecasting.
LLM-based decompilation tool translating pseudocode to compilable executable code with runtime correctness verification.
Architecture-agnostic defense mechanism against heterogeneous generative threats including diffusion models and GANs.
Vision-based dual-phase system for real-time traffic intersection control using vehicle detection.
Sample-efficient hypergradient estimation for decentralized bi-level reinforcement learning with leader-follower agent dynamics.
Generative data-augmentation framework for immunoglobulin-antigen binding prediction with synthetic training data.
GAN-based physics-driven approach for seismic full-waveform inversion under complex geological conditions.
LLM-based automated essay scoring using decision-level ordinal modeling for multimodal inputs with trait-specific visual relevance.
Signal Detection Theory analysis of LLM calibration metrics, decomposing sensitivity and bias components beyond ECE.
Post-hoc explanation method using informative perturbation selection for uncertainty-aware model interpretability.
Overview of spectrogram-based representations for deep learning audio and speech analysis systems.
Lightweight routing mechanism for transformer attention heads with mechanistic interpretability analysis of computational pathways.
ML fairness framework addressing gender bias in critical healthcare models with explainability integration.
Framework for question-aware keyframe selection in video question answering using synthetic supervision.
Multi-agent simulation framework generating synthetic corporate corpora with verifiable ground truth for RAG pipeline evaluation.
Theoretical analysis of log-barrier regularization for improving exploration in policy optimization algorithms.
Framework combining chemical structure analysis with biological insights for scalable drug discovery screening.
Flow matching models for generating pseudo-GPS trajectory data at nation-wide scale while preserving privacy.
Framework and architectural patterns for describing agentic AI systems with multi-agent collaboration and artifact exchange.
Replication study on AI-generated text detection using multilingual models and SHAP-based explainability analysis.
Vision-Language-Action model using state space models for efficient language-guided robotic manipulation tasks.
Latent-space reasoning approach for LLMs that reduces inference cost by computing reasoning steps implicitly rather than generating verbose traces.
Analysis of error sources in global feature effect estimation methods like PD and ALE plots for model interpretation.
Open-source biomedical knowledge graphs with federation and AI agent access for cross-referencing siloed databases.
Dataset and model for generating listener body motions responding to speaker utterances in multimodal interactions.
Framework for comparative legal studies connecting Japanese legal data to international standards via XML schema conversion.
Proposes Safe Flow Q-Learning for offline safe reinforcement learning with reachability-based safety constraints.
Research on improving LLMs' code generation using private libraries through better knowledge integration beyond API documentation retrieval.
HindSight evaluation framework measuring AI-generated research idea quality by matching against future publications and citation impact.
Graph learning framework with cross-attention fusion of multimodal brain imaging (fMRI and MRI) for autism spectrum disorder classification.
Hybrid approach combining Iterative Learning Control with deep reinforcement learning for safe and convergent batch process control.
Token Coherence framework applying MESI cache protocols to reduce synchronization overhead in multi-agent LLM orchestration systems.
CATFormer combines continual learning with spiking transformers using dynamic thresholds to mitigate catastrophic forgetting.
Study on architectural design choices for end-to-end autonomous driving planners using bird's eye view representations at scale.
ADV-0 closed-loop min-max adversarial training for autonomous driving robustness in long-tail safety-critical scenarios.
In-context symbolic regression for extracting interpretable analytical expressions from Kolmogorov-Arnold Networks in scientific ML.
RESTA defense extended to vision-language models for robustness against multi-modal jailbreaking attacks in trustworthy agentic AI.
Code-centric learning approach for LLM-based ICD medical coding improving generalization to unseen codes with better interpretability.
Scalable simulation-based model inference framework with test-time complexity control for selecting among large families of forward models.
CCTU benchmark evaluating LLM tool use under complex constraints, testing function calling, instruction following, and self-refinement.
TAGARELA Portuguese speech dataset with 8,972 hours of podcast audio for ASR and TTS model training.
NV-Bench benchmark for nonverbal vocalization synthesis in text-to-speech systems with 1,651 multilingual utterances and functional taxonomy.
Rectified flow-based generative model for fast seismic inversion in geophysical exploration, balancing sampling efficiency and accuracy.
FuXiWeather2 neural framework for weather forecasting combining data assimilation and prediction without reanalysis product emulation.
SKILLS benchmark framework evaluating LLM agents on 37 telecom operations workflows with real API interfaces, testing structured knowledge injection.
Hybrid XAI framework combining counterfactual explanations and feature attribution for neural network interpretability in healthcare/finance.
Analysis of beam search in LLMs showing wider beams can hurt output quality due to overestimation bias, grounded in Extreme Value Theory.