Out-of-Distribution Generalization via Composition: A Lens Through Induction Heads in Transformers

Information Science, Computer Science, Artificial Intelligence,
out-of-distribution generalization   composition   large language models   emergent ability   mechanistic interpretability  

Study on Out-of-Distribution Generalization and Composition Mechanisms in Large Language Models Paper Background In recent years, large language models (LLMs) such as GPT-4 have demonstrated remarkable creativity in handling novel tasks, often solving problems with just a few examples. These tasks require models to generalize on distributions diffe...

Discounted Stable Adaptive Critic Design for Zero-Sum Games with Application Verifications

Information Science, Computer Science, Automation, Artificial Intelligence,
Adaptive Critic Design   Discounted Value Iteration   Zero-Sum Games   System Stability   Optimal Control  

Discounted Adaptive Critic Design with Application Verification in Zero-Sum Games Research Background In the field of control, optimal control is a core research direction aimed at designing and analyzing control systems to optimize system performance. As system complexity increases, traditional optimal control methods based on the Hamilton-Jacobi-...

Resource-Efficient Decentralized Sequential Planner for Spatiotemporal Wildfire Mitigation

Engineering, Aerospace Engineering, Information Science, Computer Science, Artificial Intelligence,
UAV   wildfire mitigation   decentralized planning   spatiotemporal tasks   resource optimization  

Efficient Decentralized Sequential Planner for Spatiotemporal Wildfire Mitigation Using Multiple UAVs Academic Background Wildfires pose a significant threat to global biodiversity and resource sustainability, especially in their early stages. If not controlled in time, wildfires can rapidly expand, leading to severe ecological damage. In recent ye...

A Projective Weighted DTW Based Monitoring Approach for Multi-Stage Processes with Unequal Durations

Information Science, Computer Science, Automation, Artificial Intelligence,
Dynamic Time Warping   Multi-Stage Processes   Process Monitoring   Unequal Durations   Projective Weighted  

Projective Weighted Dynamic Time Warping-Based Monitoring Method for Multi-Stage Processes with Unequal Durations Academic Background In modern manufacturing industries, online monitoring of multi-stage processes (such as batch and transition processes) is crucial for improving product quality and reducing failure risks. However, due to varying ope...

Effective Probabilistic Neural Networks Model for Model-Based Reinforcement Learning in USV

Engineering, Information Science, Computer Science, Automation,
Probabilistic Neural Networks   Model Predictive Control   Unmanned Surface Vehicles   Reinforcement Learning   Uncertainty Propagation  

A New Approach to Model Predictive Control for Unmanned Surface Vehicles (USV): A Probabilistic Neural Network-Based MBRL Framework Academic Background Unmanned Surface Vehicles (USVs) have seen rapid development in recent years within the field of marine science, finding extensive applications in scenarios such as marine transportation, environmen...

Oral-Anatomical Knowledge-Informed Semi-Supervised Learning for 3D Dental CBCT Segmentation and Lesion Detection

Information Science, Computer Science, Automation, Medicine, Oral Sciences, Medical Imaging,
dental CBCT   semi-supervised learning   lesion detection   oral-anatomical knowledge   3D image segmentation  

Academic Background and Research Motivation In the field of dental healthcare, Cone Beam Computed Tomography (CBCT) is a widely used three-dimensional imaging technique. CBCT provides three-dimensional images of the oral cavity and is particularly effective in diagnosing odontogenic lesions. However, the segmentation of CBCT images—labeling each vo...