Gradient Concentration, Not Weight Saliency, Explains Representation-Level Class Unlearni…
Explorar
Noticias de IA
22116 elementos — filtrados, clasificados y sin duplicados
How Many Bits Can an Adapter Write? Measuring the Capacity and Memorization of Parameter-…
Quality-Aware Multimodal Fusion Reveals Implicit Identity in Valence-Arousal Features
M$^3$-Gen: Interpretable Multimodal Generation of Gene Expression Profiles Using Clinical…
Phonetic forced alignment for low-resource language varieties: Model training and evaluat…
SlerpFlow: Spherical Trajectory Correction for Rectified Flow Inversion
GRADRAG: Cross-Component Prompt Adaptation for Coordinated Multi-Agent RAG
DARPA, U.S. Air Force fly AI-controlled F-16
PC-Edit: Prompt-Contrastive Region Discovery and Region-Guided Editing
Filter Learning for Subgraphs: Algebras and Performance Risk Bounds
How AI helps scientists design the next generation of medicines
How Rules Represent Causal Knowledge: Causal Modeling with Probabilistic Logic Programming
A New Well-Supported Semantics for Description Logic Programs
Towards a Certifying Grounder
Automated Synthesis and Adversarial Validation of Executable Causal Research Pipelines
CRAG-MM-Diagnostics: Enabling Stage-Wise Analysis of Knowledge-Intensive VQA
V-DEAL: Diagnosing Video Safety De-Calibration as an Understanding-Refusal Coupling Failu…
DTIF: Robust Loop Closure Detection via Delaunay Triangle Topology in Complex Forests
Do Pathology Vision-Language Models Truly See Pathology?
QuantiBias: Benchmarking Quantization-Induced Bias in LLMs
MVEI & EmObserver: Empowering MLLM-Oriented Visual Emotional Intelligence via Emotion Sta…
Achieving Text-based Person Retrieval with Any Granularity
GroupVideo: Multi-Identity Customized Text-to-Video Generation
WAT3R: Feedforward Underwater 3D Reconstruction
ProCap: Prominence-guided Object Rectification for Faithful and Comprehensive Video Capti…
Naju: A Native Discrete State-Space Model with Independent Retention and Writing for Long…
Latent Variable-Mediated Cross-Learning for Few-Shot Acoustic Impedance Imaging
Distribution-Alignment Bridge for Uncertainty-Aware Text-to-Video Retrieval
Beyond Independent Optimization: Compression, MoE Routing, and Quantization Interactions …
China Cuts AI Gap as Moonshot Shows Zhipu Isn’t One-Off, BI Says