Summary of Knowledge Graph Reasoning with Self-supervised Reinforcement Learning, by Ying Ma et al.
Knowledge Graph Reasoning with Self-supervised Reinforcement Learning
by Ying Ma, Owen Burns, Mingqiu Wang, Gang Li, Nan Du, Laurent El Shafey, Liqiang Wang, Izhak Shafran, Hagen Soltau
First submitted to arxiv on: 22 May 2024
Categories
- Main: Computation and Language (cs.CL)
- Secondary: Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
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 The paper proposes a self-supervised pre-training method for reinforcement learning (RL) to improve the performance of RL in incomplete knowledge graphs (KGs). The proposed method, called self-supervised RL (SSRL), combines the strengths of supervised learning (SL) and RL by using SL-generated labels to train the policy network. This approach increases the information density of the SL objective and prevents the agent from getting stuck with early-rewarded paths. Experimental results show that SSRL meets or exceeds state-of-the-art results on four large benchmark KG datasets, outperforming two baseline RL models, MINERVA and MultiHopKG. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary The paper creates a new way to train machines for incomplete knowledge graphs. It uses reinforcement learning to help the machine learn from what it knows, but also adds a special step before that makes it better at understanding the graph. This helps the machine find correct answers more often. The method is tested on four big datasets and does well compared to other methods. |
Keywords
» Artificial intelligence » Reinforcement learning » Self supervised » Supervised
