Publications
* denotes equal contribution
2026
RL Post-Training Diffusion Language Models Along Their Own Generation Paths
arXiv preprint arXiv:2610.17266
(2026)
Fixed State, Long Reach: What a Constant-Size Cache Buys Block Diffusion at Scale
arXiv preprint arXiv:2609.11998
(2026)
DiffuMamba: High-Throughput Diffusion LMs with Mamba Backbone
International Conference on Machine Learning (ICML)
(2026)
Dual-Phase Continual Learning: Supervised Adaptation Meets Unsupervised Retention
Transactions on Machine Learning Research (TMLR)
(2026)
2025
Beyond Cosine Decay: On the Effectiveness of Infinite Learning Rate Schedule for Continual Pre-training
Conference on Lifelong Learning Agents (CoLLAs)
(2025)
Oral presentation
2024
2022
On Spectral and Temporal Feature Encoding Behaviour in Stacked Architectures
NeurIPS 2022 ENLSP Workshop
(2022)