About ICLARS

The International Conference on Learning and Reasoning Systems (ICLARS) is the premier venue dedicated to advancing the scientific foundations and engineering practice of tightly integrated learning–reasoning systems. ICLARS brings together researchers and professionals committed to building safe, interpretable, and domain-grounded AI systems that move beyond today’s brittle pattern-matching paradigms.

Traditional neurosymbolic approaches often treat neural and symbolic components as separate, loosely connected modules. In contrast, ICLARS promotes a unified vision where neural and symbolic representations co-evolve: neural models support the evolution of symbolic structures and abstractions, while symbolic knowledge constrains and guides neural training, enabling sample-efficient learning and explicit domain knowledge acquisition. This deeper integration offers principled safety guarantees that cannot be achieved through data alignment alone.

ICLARS serves as the global forum for publishing cutting-edge theory, methods, and systems that demonstrate real-world impact—particularly in autonomous systems, cyber-physical systems, robotics, and any domain requiring reliable reasoning under uncertainty.

Topics of Interest

ICLARS invites contributions across all aspects of integrated learning and reasoning, including but not limited to:

Foundations & Theory

• Unified paradigms for neuro-symbolic integration
• Executable semantics for learning–reasoning systems
• Concept formation, grounding, and knowledge induction
• Theoretical guarantees for safety, robustness, and correctness

Learning–Reasoning Integration

• Symbolically guided neural training and inference
• Neural models that induce, evolve, or refine symbolic rules
• Structured, low-data, and knowledge-driven learning
• Program synthesis, DSLs, and architecture-level unification

Trustworthy & Safe AI

• Formal verification for integrated AI systems
• Runtime monitoring and explainable reasoning pipelines
• Safety guardrails for ML-enabled autonomous systems
• Robustness, causality, and out-of-distribution generalization

Systems & Platforms

• Scalable architectures for neuro-symbolic pipelines
• Compilers, runtimes, and execution engines
• Benchmarks, datasets, and evaluation methodologies
• Real-world deployments in CPS, robotics, ADS, UAVs, and AI-enabled software

Applications

• Autonomous driving and aerial robotics
• Human–AI teaming and interactive agents
• Scientific discovery, digital twins, and simulation
• Safety-critical systems in healthcare, finance, and infrastructure