Hello 👋 ! I am Kejing Xia (夏可敬), a first-year Ph.D. student in Computer Science at Georgia Tech, where I work in the GTISC Lab with Prof. Wenke Lee.

Prior to my Ph.D., I earned my B.Eng. in Communication Engineering (Artificial Intelligence Track) from Wuhan University. Before that, I studied at the No.1 Middle School Affiliated to Central China Normal University (华中师大一附中).

You can view my CV here. (Updated in Feb. 2026)

Research Interests

• Diffusion LLM : Decoding Acceleration, Hybrid-continuous dLLM.
• LLM Reasoning : Latent reasoning, LLM alignment.
• Efficient LLMs : Long-context modeling, Training-free optimization.
• 3D Vision : 3D Gaussian Splatting, Event-based vision, Neural Radiance Fields.

📢 News

• 2025.09 Our paper D²GS is accepted by NeurIPS 2025.
• 2025.05 Our paper LAMB is accepted by ACL 2025.
• 2025.02 Our paper LongMamba is accepted by ICLR 2025.
• 2024.06 Our paper A Spatio-temporal Event Transformer on Versatile Tasks for Human Behavior Analysis is accepted by IJCAI-W MiGA 2024.
• 2024.05 I won National Second Prize and 5th Place (top 9%) in RoboCup China Open as team leader.
• 2023.11 I won Province First Prize in China Undergraduate Mathematical Contest in Modeling.

📝 Publications

Image Placeholder

MetaState: Persistent Working Memory for Discrete Diffusion Language Models

Under Review

Paper Code Project

Image Placeholder

D²GS: Dense Depth Regularization for LiDAR-free Urban Scene Reconstruction

NeurIPS 2025

Paper

Image Placeholder

LAMB: A Training-Free Method to Enhance the Long-Context Understanding of SSMs via Attention-Guided Token Filtering

ACL 2025

Paper Code

Image Placeholder

LongMamba: Enhancing Mamba’s Long-Context Capabilities via Training-Free Receptive Field Enlargement

ICLR 2025

Paper Code Project

🔬 Research Experience

• GTISC Lab, Georgia Tech (Jan. 2026 – Present) - Working on MetaState: Persistent Working Memory for Discrete Diffusion Language Models.
• EIC Lab, Georgia Tech (May 2024 – Feb. 2026) - Led & deeply involved in projects on R²-dLLM, Soft Token Interpolation, LongMamba, and LAMB, proposing training-free approaches and novel decoding methods for efficient language models.
• Signal Processing Lab, Wuhan University (Oct. 2023 – Apr. 2024) - Co-led BRS-E²NeRF, developing event-enhanced NeRFs to reconstruct sharp images from rolling shutter blur.
• Center for Machine Vision and Signal Analysis (CMVS), University of Oulu (2023-2024) - Led research on event-based human behavior analysis with spatio-temporal transformers, presented at IJCAI 2024 (MiGA Workshop).

💼 Professional Experience

• Bosch Research & Innovation Software Development Co., Ltd (Jul. 2024 – May 2025) - Lead Research Intern for D²GS, combining diffusion depth enhancement with 3D Gaussian models for LiDAR-free scene reconstruction.

🏅 Competition

• RoboCup China Open (2024) - Captain, National First Prize.
• Student Innovation & Entrepreneurship Project (2023-2024) - National Prize for All-in-Focus Seeing Through Occlusions with Event and Frame.
• China Undergraduate Mathematical Contest in Modeling (2023) - Hubei Province First Prize.
• Huazhong Cup Mathematical Modeling Challenge (2023) - National First Prize.

🏆 Awards

• Wuhan University First-Class Academic Scholarship (2023-2024).
• “Luojia” Distinguished Exchange Student Scholarship (2023-2024).
• Wuhan University Merit Students (2021-2022, 2023-2024).
• Wuhan University Outstanding Student (2022-2023).
• Wuhan University Merit and Outstanding Academic Scholarships (2021-2023).