P-Guide: Efficient Single-Pass CFG Inference for AI Generation
Discover P-Guide, a novel framework that cuts inference latency by 50% in Classifier-Free Guidance while maintaining high-quality AI generation.
Heuristic Design with LLMs: Bridging Code and Knowledge
Explore how large language models transform heuristic design by prioritizing knowledge over code for better optimization and algorithm development.
Policy-Guided Model Routing for Efficient AI Reasoning
Discover a novel policy-guided stepwise model routing method that enhances AI reasoning efficiency and reduces inference costs in large language models.
CrossCult-KIBench: Benchmark for Cross-Cultural MLLM Knowledge
Discover CrossCult-KIBench, a benchmark for evaluating cross-cultural knowledge insertion in MLLMs across English, Chinese, and Arabic contexts.
Constraint-Driven Resource Allocation for Agentic AI Workflows
Discover how constraint-driven resource allocation improves agentic AI workflows to meet deadlines and budgets efficiently with Monte Carlo Portfolio Plann...