Developing novel approaches to semantic understanding and information processing, creating AI systems that truly understand meaning and context.
Our Knowledge Representation research focuses on developing AI systems that can understand, reason about, and manipulate knowledge in ways that are meaningful to humans. We're working on the fundamental challenge of how to represent and process semantic information in computational systems.
By combining insights from linguistics, cognitive science, and computer science, we're building the foundation for the next generation of AI systems that can truly understand context, meaning, and the relationships between concepts.
Developing advanced embedding techniques that capture not just word similarities, but deep semantic relationships and contextual understanding.
Building dynamic knowledge graphs that can evolve and learn from new information while maintaining consistency and coherence.
Creating geometric representations of concepts that preserve semantic relationships and enable intuitive reasoning.
Developing models that can understand and adapt to different contexts, including cultural, temporal, and domain-specific variations.
Integrating knowledge from text, images, audio, and other modalities to create richer, more comprehensive representations.
Building systems that can perform logical inference and reasoning over knowledge representations to answer complex questions.
Our knowledge representation techniques are advancing the state-of-the-art in natural language understanding, enabling more accurate translation, summarization, and question-answering systems.
We're developing search systems that understand the intent and context behind queries, rather than just matching keywords, leading to more relevant and useful results.
Our research is enabling the development of AI assistants that can understand complex requests, maintain context across conversations, and provide more helpful responses.
By representing scientific knowledge in structured, computable forms, we're helping researchers discover new connections and insights across different fields.
Coming September 2025 — arXiv preprint
A novel framework for representing semantic knowledge using physics-inspired action principles and energy-based models.
Interested in advancing the state of knowledge representation? We're looking for researchers, linguists, and AI practitioners who want to help build AI systems that truly understand meaning.
Contact: [email protected]