English
Related papers

Related papers: Unlocking [CLS] Features for Continual Post-Traini…

200 papers

Continual Instruction Tuning (CIT) is adopted to continually instruct Large Models to follow human intent data by data. It is observed that existing gradient update would heavily destroy the performance on previous datasets during CIT…

Machine Learning · Computer Science 2025-12-15 Jingyang Qiao , Zhizhong Zhang , Xin Tan , Yanyun Qu , Shouhong Ding , Yuan Xie

Safety alignment in large language models (LLMs) is commonly implemented as a single static policy embedded in model parameters. However, real-world deployments often require context-dependent safety rules that vary across users, regions,…

Artificial Intelligence · Computer Science 2026-03-18 Jingyu Peng , Hongyu Chen , Jiancheng Dong , Maolin Wang , Wenxi Li , Yuchen Li , Kai Zhang , Xiangyu Zhao

Transformers have become the backbone of modern AI, yet their high computational demands pose critical system challenges. While sparse training offers efficiency gains, existing methods fail to preserve critical structural relationships…

Machine Learning · Computer Science 2025-11-18 Jinqi Xiao , Cheng Luo , Lingyi Huang , Cheng Yang , Yang Sui , Huy Phan , Xiao Zang , Yibiao Ying , Zhexiang Tang , Anima Anandkumar , Bo Yuan

Pre-training Large Language Models (LLMs) on web-scale datasets becomes fundamental for advancing general-purpose AI. In contrast, enhancing their predictive performance on downstream tasks typically involves adapting their knowledge…

Inspired by the success of Self-supervised learning (SSL) in learning visual representations from unlabeled data, a few recent works have studied SSL in the context of continual learning (CL), where multiple tasks are learned sequentially,…

Computer Vision and Pattern Recognition · Computer Science 2023-03-15 Li Yang , Sen Lin , Fan Zhang , Junshan Zhang , Deliang Fan

Continual learning (CL) is crucial for evaluating adaptability in learning solutions to retain knowledge. Our research addresses the challenge of catastrophic forgetting, where models lose proficiency in previously learned tasks as they…

Using task-specific components within a neural network in continual learning (CL) is a compelling strategy to address the stability-plasticity dilemma in fixed-capacity models without access to past data. Current methods focus only on…

Machine Learning · Computer Science 2022-07-07 Ghada Sokar , Decebal Constantin Mocanu , Mykola Pechenizkiy

With the increased attention to model efficiency, post-training sparsity (PTS) has become more and more prevalent because of its effectiveness and efficiency. However, there remain questions on better practice of PTS algorithms and the…

Machine Learning · Computer Science 2024-12-11 Zining Wnag , Jinyang Guo , Ruihao Gong , Yang Yong , Aishan Liu , Yushi Huang , Jiaheng Liu , Xianglong Liu

Supervised fine-tuning is the most common method to adapt large language models (LLMs) to downstream tasks, but full fine-tuning LLMs requires massive computational resources. Recently, parameter-efficient fine-tuning (PEFT) methods have…

Computation and Language · Computer Science 2024-02-27 Xiangdi Meng , Damai Dai , Weiyao Luo , Zhe Yang , Shaoxiang Wu , Xiaochen Wang , Peiyi Wang , Qingxiu Dong , Liang Chen , Zhifang Sui

Learning classifier systems (LCSs) originated from cognitive-science research but migrated such that LCS became powerful classification techniques. Modern LCSs can be used to extract building blocks of knowledge to solve more difficult…

Neural and Evolutionary Computing · Computer Science 2020-06-03 Isidro M. Alvarez , Trung B. Nguyen , Will N. Browne , Mengjie Zhang

Online continual learning (CL) aims to learn new knowledge and consolidate previously learned knowledge from non-stationary data streams. Due to the time-varying training setting, the model learned from a changing distribution easily…

Machine Learning · Computer Science 2023-08-15 Quanziang Wang , Renzhen Wang , Yichen Wu , Xixi Jia , Deyu Meng

Large-scale multimodal foundation models, particularly Contrastive Captioners (CoCa), have achieved state-of-the-art results by unifying contrastive alignment with generative captioning. While zero-shot transfer capabilities are…

Computer Vision and Pattern Recognition · Computer Science 2025-12-16 N. K. B. M. P. K. B. Narasinghe , Uthayasanker Thayasivam

Sparse neural networks are becoming increasingly important as the field seeks to improve the performance of existing models by scaling them up, while simultaneously trying to reduce power consumption and computational footprint.…

Machine Learning · Computer Science 2021-06-08 Siddhant M. Jayakumar , Razvan Pascanu , Jack W. Rae , Simon Osindero , Erich Elsen

Continual learning refers to the problem where the training data is available in sequential chunks, termed "tasks". The majority of progress in continual learning has been stunted by the problem of catastrophic forgetting, which is caused…

Computer Vision and Pattern Recognition · Computer Science 2023-11-07 Rajas Chitale , Ankit Vaidya , Aditya Kane , Archana Ghotkar

In continual learning, where task data arrives in a sequence, fine-tuning on later tasks will often lead to performance degradation on earlier tasks. This is especially pronounced when these tasks come from diverse domains. In this setting,…

Machine Learning · Computer Science 2025-01-13 Anat Kleiman , Gintare Karolina Dziugaite , Jonathan Frankle , Sham Kakade , Mansheej Paul

To adapt to real-world data streams, continual learning (CL) systems must rapidly learn new concepts while preserving and utilizing prior knowledge. When it comes to adding new information to continually-trained deep neural networks (DNNs),…

Machine Learning · Computer Science 2025-07-02 Md Yousuf Harun , Christopher Kanan

Fixed representational capacity is a fundamental constraint in continual learning: practitioners must guess an appropriate model width before training, without knowing how many distinct concepts the data contains. We propose LACE…

Machine Learning · Computer Science 2026-03-31 Shivnath Tathe

Parallel Continual Learning (PCL) tasks investigate the training methods for continual learning with multi-source input, where data from different tasks are learned as they arrive. PCL offers high training efficiency and is well-suited for…

Machine Learning · Computer Science 2024-07-12 Li Yuepan , Fan Lyu , Yuyang Li , Wei Feng , Guangcan Liu , Fanhua Shang

Current visual text generation models struggle with the trade-off between text accuracy and overall image coherence. We find that achieving high text accuracy can reduce aesthetic quality and instruction-following capability. Although…

Computer Vision and Pattern Recognition · Computer Science 2026-04-28 Yaohou Fan , Qingzhong Wang , Yongsong Huang , Junyi Liu , Tomo Miyazaki , Shinichiro Omachi

Post-training with reinforcement learning (RL) typically optimizes a single scalar objective and ignores structure in how solutions are produced. We ask whether a scalar hint toward a canonical solver ordering, used only during RL…

Machine Learning · Computer Science 2026-05-06 Prakhar Gupta , Vaibhav Gupta