Haoran Zhang

The paper introduces CityTrajBench, a unified benchmark framework that standardizes the evaluation of city-scale vehicle trajectory generation, demonstrating that no single generation model dominates all performance metrics.

This paper proposes a preconditioning layer for stable weight conditioning in LLM training.

This paper investigates the relationship between language model perplexity and word error rate in modern automatic speech recognition systems, studying the impact of external language models, encoder context length, and large language models.