Yan-Shuo Liang 「梁宴硕」

Currently, I am a PhD candidate in Department of Computer Science and Technology at Nanjing University, where I am fortunate to be advised by Prof. Wu-Jun Li.

DBLP  /  Github

2024.03-Present: PhD candidate, Department of Computer Science and Technology at Nanjing University,

2020.09-2024.03: PhD student, Department of Computer Science and Technology at Nanjing University,

2016.09-2020.06: B.S., Department of Mathematics at Nanjing University,

My current research focuses on continual / lifelong / incremental learning. Continual learning requires the model to learn from a non-stationary environments. When learning on a new environment (task), the model should overcomes forgetting of the old tasks. Meanwhile, I am broadly interested in computer vision and machine learning topics.

we propose a new PEFT method, called InfLoRA, for continual learning. InfLoRA injects a small number of parameters to reparameterize the pre-trained weights and shows that fine-tuning these injected parameters is equivalent to fine-tuning the pre-trained weights within a subspace. Furthermore, InfLoRA designs this subspace to eliminate the interference of the new task on the old tasks, making a good trade-off between stability and plasticity.

We propose a simple yet effective method, which separates the two objectives for the new task by decoupling the loss of the new task, providing a way to obtain a better trade-off between stability and plasticity than those methods with coupled loss.

We propose a new method that evaluates the model's plasticity and then adaptively improve the model's plasticity for learning a new task adaptively.

Presidential Special Scholarship, Nanjing University, 2020

Huawei Scholarship, 2023

Data Structure. (For undergraduate students, Autumn, 2020)

Data Structure. (For undergraduate students, Spring, 2021)

Email: liangys [at] smail [dot] nju [dot] edu [dot] cn

Nanjing University Xianlin Campus, 163 Xianlin Avenue, Qixia District, Nanjing, Jiangsu 210023, China

Template