9 papers
Outcome-Guided On-Policy Self-Distillation
ZheXu Wang, Mao-Lin Luo, Yankun Hong +6
On-policy self-distillation (OPSD) provides denser token-level supervision and better computational efficiency than Reinforcement Learning with Verifiable Rewards (RLVR). However,…
Agentic Kernel Optimization: Generating State-of-the-Art GPU Kernels Without Hand-Written CUDA
Mao Luo, Hongbin Li, Feng Lin +2
We study whether general-purpose code agents can produce state-of-the-art GPU kernels without any manually written CUDA code. We investigate this question using representative work…
RL Forgets! Towards Continual Policy Optimization
Mao-Lin Luo, Zhe-Xu Wang, Zi-Hao Zhou +4
Continual post-training is becoming a central paradigm for adapting vision-language models to evolving tasks. Recent studies report that reinforcement learning is less prone to for…
KeepLoRA++: Continual Learning with Layer-Scaled Residual Gradient Adaptation
Mao-Lin Luo, Yi-Lin Zhang, Zi-Hao Zhou +5
Continual learning for pre-trained vision-language models requires balancing three competing objectives: retaining pre-trained knowledge, preserving knowledge from a sequence of le…
Bridging Coarse and Fine Recognition: A Hybrid Approach for Open-Ended Multi-Granularity Object Recognition in Interactive Educational Games
Hanling Yi, Feng Lin, Mao Luo +3
Recent advances in Multimodal Large Language Models (MLLMs) have enabled open-ended object recognition, yet they struggle with fine-grained tasks. In contrast, CLIP-style models ex…
DC-Merge: Improving Model Merging with Directional Consistency
Han-Chen Zhang, Zi-Hao Zhou, Mao-Lin Luo +3
Model merging aims to integrate multiple task-adapted models into a unified model that preserves the knowledge of each task. In this paper, we identify that the key to this knowled…
Spectral Imbalance Causes Forgetting in Low-Rank Continual Adaptation
Hao Gu, Mao-Lin Luo, Zi-Hao Zhou +3
Parameter-efficient continual learning aims to adapt pre-trained models to sequential tasks without forgetting previously acquired knowledge. Most existing approaches treat continu…
KeepLoRA: Continual Learning with Residual Gradient Adaptation
Mao-Lin Luo, Zi-Hao Zhou, Yi-Lin Zhang +3
Continual learning for pre-trained vision-language models requires balancing three competing objectives: retaining pre-trained knowledge, preserving knowledge from a sequence of le…
LADA: Scalable Label-Specific CLIP Adapter for Continual Learning
Mao-Lin Luo, Zi-Hao Zhou, Tong Wei +1
Continual learning with vision-language models like CLIP offers a pathway toward scalable machine learning systems by leveraging its transferable representations. Existing CLIP-bas…