Logic Language for LLMs
Hallucinations are a major challenge in LLM reasoning because natural language is unstructured. By nature, LLMs are pattern engines that reason more effectively over structured entities and relationships, enabling more reliable, near-deterministic results.
Nucleoid is designed with a minimally tokenized syntax for logic representation and a declarative execution model, eliminating the need for LLMs to manually manage control flow, state propagation, and other imperative constructs. In addition, Nucleoid is a next-generation logic programming language built on structured objects and their relationships, extending the traditional Knowledge Graph.
- Near-Deterministic: Structured, reliable reasoning.
- Logic Graph: Executable knowledge graph.
- Explainability: Transparent, traceable decisions.
Socrates is mortal without being told so
# There is a Human type with a name
class Human(name: str):
this.name = name
# Every human is mortal
$Human.mortal = true
# Socrates is a Human
socrates = Human("Socrates")
# Therefore, Socrates is mortal
assert(socrates.mortal, true)
chat.mp4
| Nucleoid Runtime | |
|---|---|
| 🦀 Rust-based | ⚡ LLM-based |
| Programming Language Runtime: Implements the language specification. | Fine-Tuned LLM: Fine-tuned on synthesized datasets derived from the language specification. |
| this repo | huggingface.co/nucleoid |
docs/reference.md is the language reference: statements and state, variables and dependencies, expressions, types and instances, properties, class-level rules, blocks and scope, control flow, functions, transactions, built-in objects, error messages, and a syntax summary.
It is assembled from the NUC documents in docs/, indexed by NUC 0 with the conventions in NUC 1. nucleoid.spec.md is normative; where the two disagree, the specification wins.
docs/examples.md covers the same ground as complete programs, each one runnable as written.
A Nucleoid program is a set of statements that remain true. An assignment is not an instruction that runs once and finishes, it is a relationship the runtime records and maintains.
An assignment states a relationship, not a result
a = 1
b = a + 2
a = 3
assert(b, 5)
b is never stale. It is the sum of a and 2, so changing a brings it up to date, and the same holds for properties, class-level rules and everything else the reference covers.
Every example in the reference is executable: tests/reference.md is its executable form and runs under cargo test, as do the snippets on this page.
Neuro-Symbolic AI is an approach that integrates the strengths of both neural networks and symbolic AI to create systems that can learn from data and also reason logically. By combining these two components, Neuro-Symbolic AI aims to leverage the intuitive, pattern-recognition capabilities of neural networks along with the logical, rule-based reasoning of symbolic AI. This integration offers a more holistic AI system that is both adaptable and able to explain its decisions, making it suitable for complex decision-making tasks where both learning from data and logical reasoning are required. Here’s how it breaks down:
Neural networks in Neuro-Symbolic AI are adept at learning patterns, relationships, and features from large datasets. These networks excel in tasks that involve classification, prediction, and pattern recognition, making them invaluable for processing unstructured data, such as images, text, and audio. Neural networks, through their learning capabilities, can generalize from examples to understand complex data structures and nuances in the data.
The symbolic component of Neuro-Symbolic AI focuses on logic, rules, and symbolic representations of knowledge. Unlike neural networks that learn from data, symbolic AI uses predefined rules and knowledge bases to perform reasoning, make inferences, and understand relationships between entities. This aspect of AI is transparent, interpretable, and capable of explaining its decisions and reasoning processes in a way that humans can understand.
Declarative Language in Neuro-Symbolic AI acts as the ubiquitous language for specifying the desired outcomes of a program without detailing the procedural methods to achieve these outcomes. This type of language is essential for articulating logical rules, constraints, and relationships that underpin symbolic reasoning within these systems. It supports the formulation of structured knowledge bases and facilitates logical reasoning tasks, enabling systems to deduce, infer, and respond to queries based on established rules. Moreover, declarative languages are instrumental in integrating the outputs of neural networks into symbolic reasoning frameworks, marrying data-driven learning with rule-based logic. Their widespread use enhances the transparency, explainability, and modularity of AI systems, while also boosting their efficiency in domains heavily reliant on rule-based operations.
Declarative logic is a subset of declarative programming, a style of building programs that expresses the logic of a computation without describing its control flow. In declarative logic, you state the facts and rules that define the problem domain. The runtime environment or the system itself figures out how to satisfy those conditions or how to apply those rules to reach a conclusion. This contrasts with imperative programming, where the developer writes code that describes the exact steps to achieve a goal.
Symbolic reasoning refers to the process of using symbols to represent problems and applying logical rules to manipulate these symbols and derive conclusions or solutions. In AI and computer science, it involves using symbolic representations for entities and actions, enabling the system to perform logical inferences, decision making, and problem-solving based on the rules and knowledge encoded in the symbols.
By integrating Nucleoid into Neuro-Symbolic AI, the system benefits from enhanced interpretability and reliability. The declarative logic and rules defined in Nucleoid provide clear explanations for the AI's decisions, making it easier for users to understand and trust the system's outputs. Furthermore, the explicit reasoning capabilities help ensure that decisions are made based on logical principles, adding a layer of reliability and consistency to the AI's behavior.
In the realm of Neuro-Symbolic AI, Plasticity is an important element for the system's ability to modify and optimize its connections in response to new information. This concept is inspired by neuroplasticity in biological brains, where neurons can strengthen or weaken their connections based on activity, enabling learning and memory.
- Dynamic Knowledge Base: Neuro-Symbolic AI systems dynamically update their knowledge base as they encounter new scenarios or corrections to their previous knowledge. This continuous updating process allows the systems to remain relevant and accurate over time.
- Adaptive Logic and Reasoning: The AI system can modify its symbolic rules and reasoning strategies to better match observed data or outcomes, enhancing its decision-making and problem-solving abilities.
- Generalization and Specialization: Through plasticity, the system can generalize from learned experiences to new, unseen scenarios or specialize in certain domains by fine-tuning its parameters or rules based on specific data inputs.
The idea of plasticity in AI, especially in logic-based systems like those in Neuro-Symbolic AI, enhances the capability of machines to not only perform tasks based on fixed rules but also to evolve those rules and adapt their reasoning over time, much like the cognitive flexibility of living things.
Learn more at nucleoid.com/docs/get-started
💡 Nucleoid runtime can also run on local machine with
npx @nucleoidai/ide startandnpx @nucleoidai/expert startincluding Nucleoid Chat. These commands enable IDE and expert system components needed for Neuro-Symbolic AI.
|
Welcome! I’ve been expecting you—"Skynet was gone. And now one road has become many." 🌐
The future is building up! Neuro-Symbolic AI is now an emerging field within AI communities and marks a crucial milestone on the journey to AGI. Unfortunately, existing symbolic AI and knowledge graphs lack advancement in today's AI landscape. Nucleoid is revolutionizing knowledge graphs with declarative, logic-based, contextual runtime, which can be integrated with ANNs to lay a robust foundation for the next leap forward imao.
Can Mingir
|
Nucleoid is an implementation of symbolic AI for declarative (logic) programming at the runtime. As mentioned, the declarative runtime environment manages JavaScript state and stores each transaction in the built-in data store by declaratively rerendering JavaScript statements and building the knowledge graph (base) as well as an execution plan.
The declarative runtime isolates a behavior definition of a program from its technical instructions and executes declarative statements, which represent logical intention without carrying any technical detail. In this paradigm, there is no segregation regarding what data is or not, instead approaches how data (declarative statement) is related with others so that any type of data including business rules can be added without requiring any additional actions such as compiling, configuring, restarting as a result of plasticity. This approach also opens possibilities of storing data in the same box with the programming runtime.
|
|
|---|
In short, the main objective of the project is to manage both of data and logic under the same runtime. The declarative programming paradigm used by Nucleoid allows developers to focus on the business logic of the application, while the runtime manages the technical details.This allows for faster development and reduces the amount of code that needs to be written. Additionally, the sharding feature can help to distribute the load across multiple instances, which can further improve the performance of the system.
This is the comparation our sample order app in Nucleoid IDE against MySQL and Postgres with using Express.js and Sequelize libraries.
https://nucleoid.com/ide/sample
Performance benchmark happened in t2.micro of AWS EC2 instance and both databases had dedicated servers with no indexes and default configurations.
https://github.com/NucleoidAI/benchmark
This does not necessary mean Nucleoid runtime is faster than MySQL or Postgres, instead databases require constant maintenance by DBA teams with indexing, caching, purging etc. however, Nucleoid tries to solve this problem with managing logic and data internally. As seen in the chart, for applications with average complexity, Nucleoid's performance is close to linear because of on-chain data store, in-memory computing model as well as limiting the IO process.
Track at Trello
Please report an issue or ask a question at Discussions
Learn more at nucleoid.com
⭐️ Star us on GitHub for the support
Neuro-Symbolic AI is an emerging field and thanks to declarative logic programming, we have a brand-new approach to Neuro-Symbolic AI. Join us in shaping the future of AI!
NucBot | canmingir | 322332 | dependabot[bot] | francisco-giancarelli-crombie | Gulshanaggarwal | CanPacis |
durulkoca | halilcengel | EnesKeremAYDIN | russle-smith | russellgray |
Generated by NucBot