StreamLLM: Enhancing Constraint Programming with LLM-Generated Streamliners | Florentina Voboril
💬 Discord: https://discord.gg/h8NVzwnysW 💻 GitHub: https://github.com/centaurinstitute 🤝 LinkedIn: https://www.linkedin.com/company/centaur-ai-institute/ 📢 New Initiative: Neuro-Symbolic Agentic Protocol ⭐ Star us on GitHub https://github.com/centaurinstitute/neuro-symbolic-agentic-protocol — Solving complex constraint programming problems is challenging, as the search space grows rapidly with problem size. Additional constraints, so-called streamliners, can narrow the search space and improve solving efficiency. Traditionally, streamliners are crafted manually or require expensive offline testing. This talk presents StreamLLM, a novel approach that uses Large Language Models (LLMs) to quickly propose effective streamliners. The generated constraints are empirically evaluated on small training instances, and their performance is fed back into the LLM. The results show substantial runtime improvements across several constraint satisfaction problems. This work demonstrates how combining symbolic reasoning with sub-symbolic AI can improve the efficiency of constraint solving. Links for one or two papers you plan to cover: https://www.computer.org/csdl/proceedings-article/nse/2025/146000a017/27FR6lMQSPe https://www.jair.org/index.php/jair/article/view/18965 [2025 IEEE/ACM 1st International Workshop on Neuro-Symbolic Software Engineering (NSE); Journal of Artificial Intelligence Research (JAIR)] About the speaker: Florentina Voboril is a PhD student in the Algorithms and Complexity Group at TU Wien, supervised by Prof. Stefan Szeider. In her research, she explores how LLMs can be used to improve the efficiency of solving constraint programming problems. She has presented her work at international venues in artificial intelligence and constraint programming, including the Constraint Programming conference (CP) and the Neuro-Symbolic Software Engineering workshop (NSE), and has published in the Journal of Artificial Intelligence Research (JAIR). Beyond her research, she is committed to inspiring children to engage with STEM through workshops, her book, and visits to schools as a role model. — 🗨️ Join us for an interactive session exploring Neuro-Symbolic AI, the emerging paradigm that blends the strengths of neural networks with symbolic reasoning. We will discuss how hybrid approaches can enhance generalization, interpretability, and reasoning, and how these methods are shaping the future of intelligent systems. Whether you’re a researcher, engineer, or simply curious about the cutting edge of AI, you’ll find an engaging space to learn, connect, and exchange ideas. Event Format: 🔍 Reading Group 🎙️ Panel Discussion 🧩 Tutorials/Workshop Series 💬 Research Roundtable 🛠️ Open-Source Project Review 📄 Paper Pitch 🤝 Collaboration Hour 🍵 Networking Mixer #AI #NeuroSymbolic #FutureOfAI