Related Concept Videos
Sample Handling
Samples should be transported carefully from collection points to the laboratory. They should be properly sealed and clearly labeled to prevent cross-contamination. To preserve the sample integrity, optimal temperature conditions during transport are essential. This could involve using...
Quantifying and Rejecting Outliers: The Grubbs Test
Positive, Negative, and Zero Work
Downsampling
The Fourier transform of the decimated sequence reveals a combination of scaled and shifted versions of the original spectrum. This...
Difference from Background: Limit of Detection
The LOD indicates the presence or absence...
Sampling Theorem
You might also read
Related Articles
Articles linked to this work by shared authors, journal, and citation graph.
Rational design of next-generation vaccine adjuvants: From molecular mechanisms to hybrid delivery platforms.
Broadening the horizon: The promise of PROTACs in non-malignant disorders.
MERBench: A Unified Evaluation Benchmark for Multimodal Emotion Recognition.
Related Experiment Video
Updated: Jun 25, 2025

A Knowledge Graph Approach to Elucidate the Role of Organellar Pathways in Disease via Biomedical Reports
Published on: October 13, 2023
M2ixKG: Mixing for harder negative samples in knowledge graph.
1Department of Automation, Tsinghua University, Beijing, China.
This study introduces M²ixKG, a novel mixing strategy for knowledge graph embeddings (KGE). M²ixKG generates higher-quality negative samples, significantly improving KGE model performance and robustness.
Area of Science:
- Artificial Intelligence
- Machine Learning
- Data Science
Background:
- Knowledge Graph Embedding (KGE) maps entities and relations to low-dimensional vectors.
- Effective KGE relies on distinguishing positive from negative triplets.
- Current KGE methods struggle to generate high-quality negative samples, hindering model performance.
Purpose of the Study:
- To address the limitations of existing negative sampling strategies in KGE.
- To introduce a novel mixing strategy, M²ixKG, for generating harder negative samples.
- To enhance the robustness and generalization capabilities of KGE models.
Main Methods:
- M²ixKG employs a mixing strategy for generating negative samples in knowledge graphs.
- It involves mixing heads and tails within triplets sharing the same relation.
- It also mixes high-scoring negative samples to create more challenging ones.
Main Results:
- Experiments on three datasets demonstrate M²ixKG's superior performance.
- M²ixKG significantly outperforms previous negative sampling algorithms.
- The strategy enhances the robustness and generalization of entity embeddings.
Conclusions:
- M²ixKG offers a significant advancement in negative sampling for KGE.
- The proposed mixing strategy effectively generates harder negatives, improving model performance.
- This approach provides a more robust and generalizable solution for knowledge graph embedding.

