arXiv:2609.09691v1 Announce Type: cross Abstract: When training data are limited, increasing parameter count is not the only way to improve language-model…
Tag: cs.AI updates on arXiv.org
Introducing Consort: A Spec-First Agent Framework for Enforced, Test-Driven Development on Live Database Branches
arXiv:2609.09671v1 Announce Type: cross Abstract: When an agent writes code, the development framework becomes the control system for a non-deterministic…
Cascading Gradient Inversion via LT-Code Inspired Peeling in Federated Learning
arXiv:2609.09659v1 Announce Type: cross Abstract: Federated learning shares model updates rather than raw data, yet these updates can be inverted to…
Which Medical Questions Deserve Rationales? Perturbation-Sensitive Selection for Robust QA
arXiv:2609.09684v1 Announce Type: cross Abstract: Medical question-answering datasets often contain answer labels, whereas high-quality rationales remain…
Hyperbolic Geometry for Open-World Object Detection in Remote Sensing Imagery
arXiv:2609.09626v1 Announce Type: cross Abstract: Open-world object detection (OWOD) extends closed-set detection by requiring models to identify unknown…
RouteBridge: Reliability-Routed Bidirectional Distillation Between Neural Radiance Fields and 3D Gaussian Splatting
arXiv:2609.09606v1 Announce Type: cross Abstract: Neural radiance fields (NeRFs) and 3D Gaussian Splatting (3DGS) encode a scene with complementary…
Watermarks Without Verification: AI Text Watermarking After the EU AI Act
arXiv:2609.09604v1 Announce Type: cross Abstract: On August 2, 2026, the obligations of Article 50 of the EU AI Act took effect, requiring generative AI…
Compact Visuotactile World Models for Lifting: Prediction, Reward Alignment, and Force Constraints
arXiv:2609.09597v2 Announce Type: cross Abstract: Accurate tactile forecasts need not improve force-constrained control. We study a 652,157-parameter…
Teacher Geometry Shapes Learnability in Teacher-Student Networks
arXiv:2609.09595v1 Announce Type: cross Abstract: Teacher-student systems, in which a teacher neural network generates training labels so that a student…
Learning with Synthetic Data via SGD in High-Dimensional Linear Regression
arXiv:2609.09572v1 Announce Type: cross Abstract: Synthetic data has become a promising way to scale model training beyond limited human-generated data…