TDWI | Training & Research | Business Intelligence, Analytics, Big Data, Data Warehousing

Q&A: Can Neuro-Symbolic AI Solve AI’s Weaknesses?

Neuro-symbolic AI integrates several technologies to let enterprises efficiently solve complex problems and queries demanding reasoning skills despite having limited data. Dr. Jans Aasman, CEO of Franz, Inc., explains the benefits, downsides, and use cases of neuro-symbolic AI as well as how to know it’s time to consider the technology for your enterprise.

• By Upside Staff
• April 8, 2024

Upside: What is neuro-symbolic AI?

Dr. Jans Aasman: Neuro-symbolic AI is an advanced approach that integrates machine learning, neural network decision-making, symbolic logic and reasoning, and the capabilities of large language models. This combination enables the ability to efficiently solve complex problems and queries demanding reasoning skills while facilitating effective learning with limited data. 

By combining these approaches, the AI facilitates secondary reasoning, allowing for more nuanced inferences. This secondary reasoning not only leads to superior decision-making but also generates decisions that are understandable and explainable to humans, marking a substantial advancement in the field of artificial intelligence.

In what ways does neuro-symbolic AI extend or build upon existing AI methodologies?

Neuro-symbolic AI is designed to capitalize on the strengths of each approach to overcome their respective weaknesses, leading to AI systems that can both reason with human-like logic and adapt to new situations through learning. The tangible objective is to enhance trust in AI systems by improving reasoning, classification, prediction, and contextual understanding.

What benefits does neuro-symbolic AI offer?

The features of neuro-symbolic AI address the limitations of purely neural network-based or purely symbolic AI systems. Here are the main ones that come to mind:

• Enhanced reasoning and generalization. Neuro-symbolic AI can reason over knowledge represented symbolically, allowing it to generalize from fewer examples than neural networks, which often require large data sets to learn.


• Improved interpretability. Symbolic components allow the AI to explain its decisions and reasoning processes in a human-understandable way, addressing the "black box" issue commonly associated with deep learning models.


• Flexibility in data requirements. This approach can work with both big and small data. Symbolic AI can handle tasks with limited data through predefined rules and logic; neural networks can process and learn from large data sets.


• Error handling. Symbolic systems can be more resistant to errors because they follow explicit rules. Combining this with neural networks can create systems that are both adaptable and precise.


• Learning efficiency. Neuro-symbolic systems can potentially learn more efficiently, combining the rapid pattern recognition of neural networks with the structured knowledge representation of symbolic AI.


• Domain knowledge integration. The ability to integrate expert domain knowledge in a structured form, such as ontologies and knowledge graphs, allows these systems to make informed decisions in specific areas.


• Combination of learning and reasoning. These systems can both learn from data (just like neural networks) and reason about data (like symbolic systems do), making them versatile for a wide range of applications. 

What are the downsides of neuro-symbolic AI?

Although one of the goals of neuro-symbolic AI is to reduce the data requirements for training models, it still needs a significant amount of high-quality, structured data for the symbolic part, which is not always available.

Another consideration is available computational resources. The integration of symbolic reasoning into neural networks often involves additional computational overhead. This can make neuro-symbolic models more resource-intensive compared to pure neural networks or purely symbolic systems.

It’s also early days. Most neuro-symbolic AI tools are only now emerging from academic and industry research labs. The commercial and venture investment for AI startups is mostly being directed towards deep learning approaches and neuro-symbolic AI development will be slower until funding arrives.

How does neuro-symbolic AI work in conjunction with graph databases?

Graph databases play a crucial role in this by organizing data in a way that emphasizes the relationships and connections between different data points.

In the context of neuro-symbolic AI, graph databases can provide the structured symbolic representations required for symbolic reasoning. They store entities and their interrelations in a network structure that resembles human knowledge representation. This allows AI systems to understand and navigate complex relationships within the data.

Exactly how does neuro-symbolic AI address hallucinations?

Neuro-symbolic AI can address hallucinations -- instances where AI systems generate plausible yet incorrect or nonsensical outputs -- when employing knowledge graphs by leveraging their explicit, structured nature to corroborate and validate the outputs generated by neural networks.

The integration of knowledge graphs in neuro-symbolic AI systems can provide a more reliable foundation for AI reasoning, ensuring that the generated content is not only contextually relevant but also factually correct. The structured data within knowledge graphs acts as a reference point that AI systems can use to cross-verify the plausibility and accuracy of their outputs, thus addressing the issue of hallucinations more effectively.

I’ll give you some examples of how hallucinations are minimized using neuro-symbolic AI in the context of a knowledge graph:

• Structured knowledge validation. Knowledge graphs provide a structured form of data that symbolic AI can use to verify the outputs of neural networks. If a neural network generates an assertion, the symbolic AI can check this against the knowledge graph to confirm whether it aligns with known facts and relationships.


• Logic-based reasoning. Neuro-symbolic AI can incorporate logic-based reasoning with knowledge graphs. This means that any new information or output is not just based on pattern recognition but also on logical inferences, which can significantly reduce hallucinations that arise from pattern-based errors.


• Uncertainty modeling. With the symbolic component, neuro-symbolic AI can assign confidence levels to information by understanding the strengths of relationships in a knowledge graph. This helps in recognizing and highlighting outputs that are less certain, thus alerting users to potential hallucinations.


• Constraint satisfaction. Knowledge graphs can enforce constraints that outputs must satisfy, which helps in filtering out hallucinations. If a generated output violates these constraints, it can be discarded or flagged for review.


• Incompleteness tolerance. Knowledge graphs, when used with neuro-symbolic AI, can handle incomplete information by not making false assumptions. This open-world assumption means the system doesn’t deduce false information from a lack of data, thereby reducing the chance of creating hallucinations.

What types of organizations and/or applications would benefit the most from using neuro-symbolic AI?

Neuro-symbolic AI can significantly approve applications in many industries:

• Healthcare can benefit from its data-efficient, interpretable models for disease diagnosis and drug repurposing


• Finance and business management sectors can utilize it for risk assessment and internal control automation


• Multimedia gains enhanced content processing with explainable AI


• Improved robotics that support advanced perception and decision-making capabilities can be applied in many industries


• Cybersecurity, education systems, smart cities, and recommendation systems can benefit from the improved reasoning and learning capabilities

How does neuro-symbolic AI fit within the trajectory of AI?

Neuro-symbolic AI represents a leap forward in the trajectory of AI advancements. The approach integrates the statistical prowess of machine learning with the structured, rule-based reasoning of symbolic AI. This fusion is seen as a promising path toward achieving artificial general intelligence.

The ultimate goal is to create AI systems that can reason, learn, and cognitively model in a manner similar to humans, filling the gaps that current AI systems face, particularly in terms of reasoning, generalization, and explainability.

Neuro-symbolic AI solves symbolic problems with deep learning, uses symbolic knowledge bases to enhance deep learning systems, provides explainability through background knowledge, and solves complex problems by coupling deep learning and symbolic components. The integration of expressive metadata and knowledge bases is particularly crucial because it provides additional context that can significantly improve the performance of deep learning systems. This increased capability for reasoning and problem-solving is what sets neuro-symbolic AI apart and makes it a critical step in the progression of AI technology.

How does an enterprise know it’s time to explore neuro-symbolic AI?

Enterprises that recognize the limitations of purely statistical AI models in addressing their complex, data-driven challenges and those aiming for greater AI transparency, integration, and innovation are likely at a stage where exploring neuro-symbolic AI could be highly beneficial. If an enterprise is faced with any of the following challenges, neuro-symbolic AI would be advantageous.

• Complex decision-making. Enterprises facing challenges in making sense of vast, complex data sets or those needing to automate intricate decision-making processes could benefit significantly from neuro-symbolic AI. This AI paradigm excels at integrating deep learning's predictive capabilities with symbolic AI's logical reasoning to address complex problems.


• Requirement for explainable AI. Industries subject to strict regulatory requirements or those where trust and transparency are paramount may find neuro-symbolic AI particularly advantageous. This is because neuro-symbolic AI offers a more interpretable form of AI, providing insights into how decisions are made or conclusions are reached, which is essential for compliance and building trust with stakeholders.


• Data integration and reasoning. Organizations with diverse data ecosystems looking to integrate and reason across various data formats and sources could leverage neuro-symbolic AI. By combining knowledge graphs for structured data representation with neural networks for unstructured data analysis, neuro-symbolic AI facilitates a comprehensive understanding of data landscapes.


• Enhanced customer interaction and personalization. Enterprises aiming to elevate customer experience through personalized interactions, recommendations, or services can use neuro-symbolic AI. This technology's ability to understand and process natural language at a nuanced level, combined with logical reasoning, enables more sophisticated and contextually relevant customer engagements.


• Innovation and competitive advantage. For businesses seeking to innovate and maintain a competitive edge, adopting neuro-symbolic AI can unlock new capabilities and applications. This could include developing advanced recommendation systems, improving predictive maintenance, or creating more dynamic and intelligent virtual assistants.

[Editor’s note: Jans Aasman is a Ph.D. psychologist and expert in cognitive science as well as CEO of Franz, Inc., an early innovator in artificial intelligence and provider of knowledge graph solutions. As both a scientist and CEO, Dr. Aasman continues to break ground in the areas of neuro-symbolic artificial intelligence and knowledge graphs as he works hand in hand with numerous Fortune 500 organizations as well as government entities worldwide.  You can reach Dr. Aasman via LinkedIn.]