paper-with-me

Papers

Knowledge Prompts: Injecting World Knowledge into Language Models through Soft Prompts

2022-10-10 · Cicero Nogueira dos santos, Zhe Dong, Daniel Cer, John Nham, Siamak Shakeri, Jianmo Ni, Yun-Hsuan Sung

Soft prompts have been recently proposed as a tool for adapting large frozen language models (LMs) to new tasks. In this work, we repurpose soft prompts to the task of injecting world knowledge into LMs. We introduce a method to train soft prompts via self-supervised learning on data from knowledge bases. The resulting soft knowledge prompts (KPs) are task independent and work as an external memory of the LMs. We perform qualitative and quantitative experiments and demonstrate that: (1) KPs can effectively model the structure of the training data; (2) KPs can be used to improve the performance of LMs in different knowledge intensive tasks.

📄 PDF Abstract BibTeX arXiv:2210.04726

Code (0)

등록된 구현이 없습니다.

Tasks

Self-Supervised LearningWorld Knowledge

Similar Papers 제목 키워드 기반

Learning and Unlearning of Fabricated Knowledge in Language Models

2024-10-29 · Chen Sun, Nolan Andrew Miller, Andrey Zhmoginov, Max Vladymyrov 외

What happens when a new piece of knowledge is introduced into the training data and how long does it last while a large language model (LM) continues to train? We investigate this question by injecting facts into LMs fro…

Data PoisoningLanguage ModelingLanguage ModellingLarge Language Model+1

Search Beyond What Can Be Taught: Evolving the Knowledge Boundary in Agentic Visual Generation

2026-07-06 · Haozhe Wang, Weijia Feng, Jinpeng Yu, Che Liu 외 arxiv

Visual generators excel at rendering, but they confidently fabricate what they do not know. User requests are unbounded, evolving, and deeply long-tailed: new characters, trending entities, post-cutoff events, and more. …

Enhancing Question Answering by Injecting Ontological Knowledge through Regularization

2020-11-01 · EMNLP (DeeLIO) 2020 11 · Travis Goodwin, Dina Demner-Fushman

Deep neural networks have demonstrated high performance on many natural language processing (NLP) tasks that can be answered directly from text, and have struggled to solve NLP tasks requiring external (e.g., world) know…

Question AnsweringSemantic CompositionWorld Knowledge

Research on World Models Is Not Merely Injecting World Knowledge into Specific Tasks

2026-02-02 · Bohan Zeng, Kaixin Zhu, Daili Hua, Bozhou Li 외 arxiv

World models have emerged as a critical frontier in AI research, aiming to enhance large models by infusing them with physical dynamics and world knowledge. The core objective is to enable agents to understand, predict, …

KORE: Enhancing Knowledge Injection for Large Multimodal Models via Knowledge-Oriented Controls

2025-10-22 · Kailin Jiang, Hongbo Jiang, Ning Jiang, Zhi Gao 외 arxiv

Large Multimodal Models encode extensive factual knowledge in their pre-trained weights. However, its knowledge remains static and limited, unable to keep pace with real-world developments, which hinders continuous knowl…

Data Augmentation