Summary of Preserving Knowledge in Large Language Model with Model-agnostic Self-decompression, by Zilun Zhang et al.
Preserving Knowledge in Large Language Model with Model-Agnostic Self-Decompression
by Zilun Zhang, Yutao Sun, Tiancheng Zhao, Leigang Sha, Ruochen Xu, Kyusong Lee, Jianwei Yin
First submitted to arxiv on: 17 Jun 2024
Categories
- Main: Computation and Language (cs.CL)
- Secondary: Artificial Intelligence (cs.AI); Computer Vision and Pattern Recognition (cs.CV)
GrooveSquid.com Paper Summaries
GrooveSquid.com’s goal is to make artificial intelligence research accessible by summarizing AI papers in simpler terms. Each summary below covers the same AI paper, written at different levels of difficulty. The medium difficulty and low difficulty versions are original summaries written by GrooveSquid.com, while the high difficulty version is the paper’s original abstract. Feel free to learn from the version that suits you best!
| Summary difficulty | Written by | Summary |
|---|---|---|
| High | Paper authors | High Difficulty Summary Read the original abstract here |
| Medium | GrooveSquid.com (original content) | Medium Difficulty Summary This paper proposes a novel self-decompression method called Tree Generation (TG) to address catastrophic forgetting in Large Language Models (LLMs). Specifically, it presents TG-SFT, a model-agnostic approach that generates synthetic training data for instruction tuning steps. The authors aim to reduce the forgetting problem by incorporating the dumped corpus during supervised fine-tuning (SFT) for Multimodal LLMs. By doing so, they achieve significant improvements in performance on language benchmarks compared to single-modality models. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary This paper helps us solve a big problem with computers that can learn from words and pictures. These “large language models” are very good at answering questions and completing tasks, but sometimes they forget what they already knew when learning new things. The authors of this paper came up with a clever way to keep the old knowledge while still learning new things. They created a special method called Tree Generation that helps computers remember what they learned before, so they don’t forget it as easily. This is important because we want our computers to be able to learn and adapt without losing their memory. |
Keywords
» Artificial intelligence » Fine tuning » Instruction tuning » Supervised
