# Semantics of the Black-Box: Can knowledge graphs help make deep learning systems more interpretable and explainable?

arXiv: https://arxiv.org/abs/2010.08660 · HTML: https://mlsanity.com/paper/2010.08660/

## Key facts

- Cited by 31 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 16 Oct 2020.

## Built on by (most cited first, 30 of 31)

- [Explainable AI (XAI): A Systematic Meta-Survey of Current Challenges and Future Opportunities](https://mlsanity.com/paper/2111.06420/) (2021)
- [Incorporating Domain Knowledge into Deep Neural Networks](https://mlsanity.com/paper/2103.00180/) (2021)
- [A Review on Explainability in Multimodal Deep Neural Nets](https://mlsanity.com/paper/2105.07878/) (2021)
- [A Decade of Knowledge Graphs in Natural Language Processing: A Survey](https://mlsanity.com/paper/2210.00105/) (2022)
- [Knowledge-intensive Language Understanding for Explainable AI](https://mlsanity.com/paper/2108.01174/) (2021)
- [Characterization of Time-variant and Time-invariant Assessment of Suicidality on Reddit using C-SSRS](https://mlsanity.com/paper/2104.04140/) (2021)
- [A Review on Knowledge Graphs for Healthcare: Resources, Applications, and Promises](https://mlsanity.com/paper/2306.04802/) (2023)
- [Towards Tractable Mathematical Reasoning: Challenges, Strategies, and Opportunities for Solving Math Word Problems](https://mlsanity.com/paper/2111.05364/) (2021)
- [KI-BERT: Infusing Knowledge Context for Better Language and Domain Understanding](https://mlsanity.com/paper/2104.08145/) (2021)
- [ProKnow: Process Knowledge for Safety Constrained and Explainable Question Generation for Mental Health Diagnostic Assistance](https://mlsanity.com/paper/2305.08010/) (2023)
- [Explainable, Domain-Adaptive, and Federated Artificial Intelligence in Medicine](https://mlsanity.com/paper/2211.09317/) (2022)
- [A Cross Attention Approach to Diagnostic Explainability using Clinical Practice Guidelines for Depression](https://mlsanity.com/paper/2311.13852/) (2023)
- [Comprehensible Artificial Intelligence on Knowledge Graphs: A survey](https://mlsanity.com/paper/2404.03499/) (2024)
- ["Is depression related to cannabis?": A knowledge-infused model for Entity and Relation Extraction with Limited Supervision](https://mlsanity.com/paper/2102.01222/) (2021)
- [KSAT: Knowledge-infused Self Attention Transformer -- Integrating Multiple Domain-Specific Contexts](https://mlsanity.com/paper/2210.04307/) (2022)
- [The Future of Cognitive Strategy-enhanced Persuasive Dialogue Agents: New Perspectives and Trends](https://mlsanity.com/paper/2402.04631/) (2024)
- [Causal Intersectionality and Dual Form of Gradient Descent for Multimodal Analysis: a Case Study on Hateful Memes](https://mlsanity.com/paper/2308.11585/) (2023)
- [A Survey on Interpretable Cross-modal Reasoning](https://mlsanity.com/paper/2309.01955/) (2023)
- [Semantic XAI for contextualized demand forecasting explanations](https://mlsanity.com/paper/2104.00452/) (2021)
- [Combining Transformers with Natural Language Explanations](https://mlsanity.com/paper/2110.00125/) (2021)
- [Requirements for Explainability and Acceptance of Artificial Intelligence in Collaborative Work](https://mlsanity.com/paper/2306.15394/) (2023)
- [Navigating Healthcare Insights: A Birds Eye View of Explainability with Knowledge Graphs](https://mlsanity.com/paper/2309.16593/) (2023)
- [Empowering machine learning models with contextual knowledge for enhancing the detection of eating disorders in social media posts](https://mlsanity.com/paper/2402.05536/) (2024)
- [Targeted aspect-based emotion analysis to detect opportunities and precaution in financial Twitter messages](https://mlsanity.com/paper/2404.08665/) (2024)
- [Intelligent interactive technologies for mental health and well-being](https://mlsanity.com/paper/2105.05306/) (2021)
- [A Review of Some Techniques for Inclusion of Domain-Knowledge into Deep Neural Networks](https://mlsanity.com/paper/2107.10295/) (2021)
- [M-ar-K-Fast Independent Component Analysis](https://mlsanity.com/paper/2108.07908/) (2021)
- [MMTM: Multi-Tasking Multi-Decoder Transformer for Math Word Problems](https://mlsanity.com/paper/2206.01268/) (2022)
- [Neuro-Symbolic Approaches for Context-Aware Human Activity Recognition](https://mlsanity.com/paper/2306.05058/) (2023)
- [IoT-Based Preventive Mental Health Using Knowledge Graphs and Standards for Better Well-Being](https://mlsanity.com/paper/2406.13791/) (2024)

## Abstract (arXiv)

The recent series of innovations in deep learning (DL) have shown enormous potential to impact individuals and society, both positively and negatively. The DL models utilizing massive computing power and enormous datasets have significantly outperformed prior historical benchmarks on increasingly difficult, well-defined research tasks across technology domains such as computer vision, natural language processing, signal processing, and human-computer interactions. However, the Black-Box nature of DL models and their over-reliance on massive amounts of data condensed into labels and dense representations poses challenges for interpretability and explainability of the system. Furthermore, DLs have not yet been proven in their ability to effectively utilize relevant domain knowledge and experience critical to human understanding. This aspect is missing in early data-focused approaches and necessitated knowledge-infused learning and other strategies to incorporate computational knowledge. This article demonstrates how knowledge, provided as a knowledge graph, is incorporated into DL methods using knowledge-infused learning, which is one of the strategies. We then discuss how this makes a fundamental difference in the interpretability and explainability of current approaches, and illustrate it with examples from natural language processing for healthcare and education applications.