papers

Publications (18)

cs.CV2026

E3AD: An Emotion-Aware Vision-Language-Action Model for Human-Centric End-to-End Autonomous Driving

Yihong Tang, Haicheng Liao, Tong Nie +7

End-to-end autonomous driving (AD) systems increasingly adopt vision-language-action (VLA) models, yet they typically ignore the passenger's emotional state, which is central to co…

cs.LG2024

Activity-aware Human Mobility Prediction with Hierarchical Graph Attention Recurrent Network

Yihong Tang, Junlin He, Zhan Zhao

Human mobility prediction is a fundamental task essential for various applications in urban planning, location-based services and intelligent transportation systems. Existing metho…

cs.RO2026

World Models as Adversaries: Multi-Agent Self-Play Fine-Tuning for Robust Motion Planning

Tong Nie, Yuewen Mei, Junlin He +3

Robust motion planning in dense traffic requires autonomous vehicles to interact in rare and safety-critical scenarios that are underrepresented in naturalistic driving data. Altho…

cs.DB2022

HQANN: Efficient and Robust Similarity Search for Hybrid Queries with Structured and Unstructured Constraints

Wei Wu, Junlin He, Yu Qiao +3

The in-memory approximate nearest neighbor search (ANNS) algorithms have achieved great success for fast high-recall query processing, but are extremely inefficient when handling h…

cs.LG2025

Solving Oversmoothing in GNNs via Nonlocal Message Passing: Algebraic Smoothing and Depth Scalability

Weiqi Guan, Junlin He

The relationship between Layer Normalization (LN) placement and the oversmoothing phenomenon remains underexplored. We identify a critical dilemma: Pre-LN architectures avoid overs…

cs.LG2024

Preventing Dimensional Collapse in Self-Supervised Learning via Orthogonality Regularization

Junlin He, Jinxiao Du, Wei Ma

Self-supervised learning (SSL) has rapidly advanced in recent years, approaching the performance of its supervised counterparts through the extraction of representations from unlab…

cs.CL2026

Reasoning-preserved Efficient Distillation of Large Language Models via Activation-aware Initialization

Junlin He, Yihong Tang, Tong Nie +5

Efficient Distillation (EDistill) compresses large language models (LLMs) by structured pruning parameters and tuning lightweight modules with high training efficiency. Although th…

cs.LG2026

Bridge: Retrieval-Augmented Spatiotemporal Modeling for Urban Delivery Demand

Yihong Tang, Tong Nie, Junlin He +3

Forecasting urban delivery demand becomes substantially more challenging when newly added service regions lack historical records. Existing spatiotemporal forecasters effectively m…

stat.AP2025

Estimating Real Demand Using a Flipped Queueing Model: A Case of Shared Micro-Mobility Services

Binyu Yang, Jinxiao Du, Junlin He +2

The spatial-temporal imbalance between supply and demand in shared micro-mobility services often leads to observed demand being censored, resulting in incomplete records of the und…

cs.AI2024

Geolocation Representation from Large Language Models are Generic Enhancers for Spatio-Temporal Learning

Junlin He, Tong Nie, Wei Ma

In the geospatial domain, universal representation models are significantly less prevalent than their extensive use in natural language processing and computer vision. This discrep…

cs.AI2026

Steerable Adversarial Scenario Generation through Test-Time Preference Alignment

Tong Nie, Yuewen Mei, Yihong Tang +5

Adversarial scenario generation is a cost-effective approach for safety assessment of autonomous driving systems. However, existing methods are often constrained to a single, fixed…

cs.LG2026

ADV-0: Closed-Loop Min-Max Adversarial Training for Long-Tail Robustness in Autonomous Driving

Tong Nie, Yihong Tang, Junlin He +5

Deploying autonomous driving systems requires robustness against long-tail scenarios that are rare but safety-critical. While adversarial training offers a promising solution, exis…

cs.CV2026

PLAF: Pixel-wise Language-Aligned Feature Extraction for Efficient 3D Scene Understanding

Junjie Wen, Junlin He, Fei Ma +1

Accurate open-vocabulary 3D scene understanding requires semantic representations that are both language-aligned and spatially precise at the pixel level, while remaining scalable…

cs.LG2025

Joint Estimation and Prediction of City-wide Delivery Demand: A Large Language Model Empowered Graph-based Learning Approach

Tong Nie, Junlin He, Yuewen Mei +4

The proliferation of e-commerce and urbanization has significantly intensified delivery operations in urban areas, boosting the volume and complexity of delivery demand. Data-drive…

cs.AI2026

EvoDrive: Pareto Evolution for Safety-Critical Autonomous Driving via Self-Improving LLM Agents

Tong Nie, Yuewen Mei, Yihong Tang +4

Generating safety-critical scenarios is essential for validating and improving autonomous driving systems, yet it inherently requires maximizing adversariality to expose failures w…

cs.AI2026

MobEvolve: An Agentic Self-Evolving Heuristic System for Interpretable Human Mobility Generation

Junlin He, Yihong Tang, Tong Nie +6

Human mobility generation aims to synthesize realistic trip chains for target populations based on individual features. Existing paradigms, including deep generative models, LLM-ba…

cs.LG2026

LLMSynthor: Macro-Aligned Micro-Records Synthesis with Large Language Models

Yihong Tang, Menglin Kong, Junlin He +3

Macro-aligned micro-records are crucial for credible simulations in social science and urban studies. For example, epidemic models are only reliable when individual-level mobility…

cs.LG2024

Preventing Model Collapse in Deep Canonical Correlation Analysis by Noise Regularization

Junlin He, Jinxiao Du, Susu Xu +1

Multi-View Representation Learning (MVRL) aims to learn a unified representation of an object from multi-view data. Deep Canonical Correlation Analysis (DCCA) and its variants shar…