LFPT5: A Unified Framework for Lifelong Few-shot Language Learning Based on Prompt Tuning of T5

By Chengwei Qin, Shafiq R. Joty. 2021

136 citations recorded by the source corpus.

Existing approaches to lifelong language learning rely on plenty of labeled data for learning a new task, which is hard to obtain in most real scenarios. Considering that humans can continually learn new tasks from a handful of examples, we expect the models also to be able to generalize well on new few-shot tasks without forgetting the previous ones. In this work, we define this more challenging yet practical problem as Lifelong Few-shot Language Learning (LFLL) and propose a unified framework for it based on prompt tuning of T5. Our framework called LFPT5 takes full advantage of PT's strong few-shot learning ability, and simultaneously trains the model as a task solver and a data generator. Before learning a new domain of the same task type, LFPT5 generates pseudo (labeled) samples of previously learned domains, and later gets trained on those samples to alleviate forgetting of previous knowledge as it learns the new domain. In addition, a KL divergence loss is minimized to achieve label consistency between the previous and the current model. While adapting to a new task type, LFPT5 includes and tunes additional prompt embeddings for the new task. With extensive experiments, we demonstrate that LFPT5 can be applied to various different types of tasks and significantly outperform previous methods in different LFLL settings.

Subfields: Artificial Intelligence, Computer Vision and Pattern Recognition, Prompting & Reasoning, Meta & Continual Learning

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

1. Han-Jia Ye
2. De-chuan Zhan
3. Da-Wei Zhou
4. Zixuan Ke
5. Bin Liu
6. Hang Su
7. Jun Zhu
8. Ziwei Liu

In Computer Vision and Pattern Recognition

1. Han-Jia Ye
2. De-chuan Zhan
3. Jun Zhu
4. Da-Wei Zhou
5. Hang Su
6. Ziwei Liu
7. Zixuan Ke
8. Bin Liu

In Prompting & Reasoning

1. Bin Liu
2. Zixuan Ke
3. Chengwei Qin
4. Liyuan Wang
5. Jun Zhu
6. De-chuan Zhan
7. Xuanjing Huang
8. Michihiro Yasunaga

In Meta & Continual Learning

1. Zixuan Ke
2. Bin Liu
3. Han-Jia Ye
4. De-chuan Zhan
5. Da-Wei Zhou
6. Liyuan Wang
7. Xingxing Zhang
8. Hang Su

Citations in the corpus

• A Comprehensive Survey of Continual Learning: Theory, Method and Application (2023, background): “The applications of continual learning in NLP are also highly diverse, such as dialogue system [130], [131], [260], [260], [282], [298], [437], text classification [87], [172], [325], [460], sentence generation [246], [298], [499], relation learning [343], [357], [455], [465], neural machine transla”
• Progressive Prompts: Continual Learning for Language Models (2023, methodology): “Hence, we did not include adapter baselines in this study, but instead focused on a more parameter-efficient and better-performing method – LFPT5.”
• Continual Learning of Natural Language Processing Tasks: A Survey (2022, methodology): “ELLE [Qin et al., 2022] expands layers whenever a new task comes.”
• Continual Training of Language Models for Few-Shot Learning (2022, background): “Several researchers have studied this problem recently (An-toniou et al., 2020; Qin and Joty, 2021; Jin et al., 2021; Xia et al., 2021; Yin et al., 2022).”
• Continual Few-shot Relation Learning via Embedding Space Regularization and Data Augmentation (2022, background): “searchers have considered lifelong language learning (Sun et al., 2020; Qin and Joty, 2022), where the model is expected to continually learn from different types of tasks.”
• Prompt Conditioned VAE: Enhancing Generative Replay for Lifelong Learning in Task-Oriented Dialogue (2022, background): “Some recent studies employ prompt tuning on continual learning for dialogue state tracking (Zhu et al., 2022) and few-shot learning (Qin and Joty, 2022).”
• Is GPT-3 a Good Data Annotator? (2022, unknown): “In practice, this paradigm has proven remarkably effective in low-data regimes (Scao and Rush, 2021; Gao et al., 2021; Qin and Joty, 2022b).”
• Orthogonal Subspace Learning for Language Model Continual Learning (2023, methodology): “To compare O-LoRA to the recent CL approaches (Wang et al., 2022c; Qin and Joty, 2021), we use the pre-trained T5-large model.”
• Is ChatGPT a General-Purpose Natural Language Processing Task Solver? (2023, background): “The strong performance of LLMs has led to a significant amount of work analyzing the abilities and behaviors of LLMs (Webson and Pavlick, 2022; Min et al., 2022; Liang et al., 2022).”
• A Survey on Few-Shot Class-Incremental Learning (2023, background): “Qin and Joty (2021) propose a unified framework for lifelong few-shot language learning, LFPT5, based on prompt tuning of T5.”
• LFS-GAN: Lifelong Few-Shot Image Generation (2023, background): “Previous studies on lifelong few-shot learning have primarily focused on discriminative tasks [1, 32, 21, 40, 38, 41].”
• Deep Class-Incremental Learning: A Survey (2023, unknown): “, semantic segmentation [164], [165], [166], [167], [174], object detection [175], [176], [177], [178], [179], [180], [181], video understanding [95], [133], medical surgery [182], [183], language model [184], [185], [186], [187], language-vision model [188], [189], etc.”
• Class-Incremental Learning: A Survey (2023, background): “…[137], [242], [243], [244], [245], [246], [247], video understanding [91], [161], medical surgery [248], [249], language model [250], [251], [252], [253], language-vision model [217], [254], etc. Specifically, object detection and semantic segmentation are two popular vision tasks that may…”