← Search

Lu Pan

6 accepted papers

2026

Azimuth-LIO: Robust LiDAR-Inertial Odometry via Azimuth-Aware Voxelization and Probabilistic Fusion

RA-L 2026

Voxel-based LiDAR–inertial odometry (LIO) is accurate and efficient but can suffer from geometric inconsistencies when single-Gaussian voxel models indiscriminately merge observations from conflicting viewpoints. To address this limitation, we propose Azimuth-LIO, a robust voxel-based LIO framework

Cited by 0SourceScholar
2026

Navigating the Pareto Frontier of Alignment:Spectrum-Adaptive Fine-Tuning for LLMs

ICML 2026poster

Supervised Fine-Tuning (SFT) with Negative Log-Likelihood (NLL) remains the standard post-training paradigm for Large Language Models, yet it imposes an excessive penalty on low-probability target tokens. This focus forces the model to prioritize minimizing the loss of difficult samples over optimiz…

Cited by 0SourceScholar
2022

When Transfer Learning Meets Cross-City Urban Flow Prediction: Spatio-Temporal Adaptation Matters

IJCAI 2022poster

Urban flow prediction is a fundamental task to build smart cities, where neural networks have become the most popular method. However, the deep learning methods typically rely on massive training data that are probably inaccessible in real world. In light of this, the community calls for knowledge t…

Cited by 22SourcePDFScholar
2021

Simple or Complex? Complexity-controllable Question Generation with Soft Templates and Deep Mixture of Experts Model

EMNLP 2021finding

The ability to generate natural-language questions with controlled complexity levels is highly desirable as it further expands the applicability of question generation. In this paper, we propose an end-to-end neural complexity-controllable question generation model, which incorporates a mixture of e…

Cited by 17SourcePDFScholar