Tag
AptMQL-Bench proposes a coding-agent-driven, human-in-the-loop pipeline to convert the BIRD text-to-SQL benchmark into a MongoDB-native text-to-MQL benchmark via access-pattern-based schema design and lossless data migration, showing that existing mechanical conversions fail or degrade performance, and that even Claude OpOpus 4.5 only reaches 57.38-70.34% accuracy on the new task.
该论文提出PlanPool方法，通过将澄清计划外部化为可变问题池，解决agentic Text-to-SQL系统中过早终止澄清的问题，从而减少静默失败并提升歧义覆盖能力。
DeSQ is a decomposition-based framework for generating SPARQL queries from natural language questions. It breaks complex questions into atomic constraints, maps them to SPARQL fragments, and assembles them into complete queries, outperforming state-of-the-art on four out of five benchmarks.
Introduces PQR, a framework that automatically generates diverse and realistic user queries to uncover failures in LLM-based QA agents, achieving 23-78% more unhelpful responses compared to prior methods.