Summary of Influence Of Backdoor Paths on Causal Link Prediction, by Utkarshani Jaimini et al.
Influence of Backdoor Paths on Causal Link Prediction
by Utkarshani Jaimini, Cory Henson, Amit Sheth
First submitted to arxiv on: 12 Sep 2024
Categories
- Main: Artificial Intelligence (cs.AI)
- Secondary: None
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 novel approach, CausalLPBack, to predict causal links in knowledge graphs. The current method uses weighted causal relations, but this can lead to spurious results due to confounders. To address this, the authors suggest blocking these confounders using backdoor path adjustment. This involves removing non-causal association flows that connect the cause-entity to the effect-entity through other variables. The approach extends a neuro-symbolic framework to enable traditional causal AI concepts and methods. It uses a unique dataset splitting method called the Markov-based split, which is relevant for causal link prediction. Evaluation shows at least 30% improvement in MRR and 16% in Hits@K for causal link prediction. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary Causal links are important in knowledge graphs because they help us understand how things relate to each other. Right now, we can predict some of these links, but it’s not very accurate because sometimes there’s a problem called confounding. Confounding is when something that shouldn’t be related to the cause and effect actually is. This makes our predictions wrong. To fix this, scientists are using a new way to remove these problems. They’re calling it CausalLPBack. It uses a special kind of math called backdoor path adjustment. This helps us get rid of the confounding and make more accurate predictions. |
