Lune

ACL Findings2026

Inventive Problem Solving with LLMs: A Benchmark for TRIZ Reasoning

Zhu Wang, Brian Uzzi

Abstract

Large language models are increasingly used in inventive problem-solving, but effective support requires more than open-ended idea generation. Inventive problem-solving requires improving one aspect of a technical system without unintentionally worsening another. TRIZ (Theory of Inventive Problem Solving) provides a unique and structured framework for this setting by representing engineering tradeoffs as contradictions and linking them to standardized inventive principles. However, prior TRIZ-LLM evaluations are typically smallscale, case studies in focused areas of technology, and rarely grounded in patent text, which makes it difficult to assess structured reasoning at scale. We introduce TRIZBENCH, a dataset and benchmark for TRIZ reasoning grounded in open technical sources and U.S. patents. TRIZBENCH evaluates the core TRIZ workflow through three tasks: contradiction prediction, inventive principle prediction, and grounded TRIZ reasoning. Experiments with multiple LLM baselines show that detecting contradictions is easier than recovering correct trade-off pairs, while principle prediction benefits from explicitly exploiting TRIZ structure. Our findings further underscore the importance of grounding. We show that semantic retrieval enables evidence-based justifications and helps explain why LLMs fail. Dataset and code are available at https://github.com/ ellenzhuwang/trizbench.

Ask Lune

Know this paper in minutes, not an afternoon.

Lune can read every page of it and the top-tier papers that cite it, and gives you its claims, its evidence and its limits, each one cited to the paper it comes from. No skimming, no AI slop.

Questions to start from

Lune can draw on

the full paper · related work

Get the answer

Free to start. No credit card required.

Lineage

Builds onCited byLeft to right by year; the more cited, the higher.

Builds on3

  1. MPNet: Masked and Permuted Pre-training for Language UnderstandingKaitao Song, Xu Tan, Tao Qin, Jianfeng Lu et al.NeurIPS 2020 · 1,957 citations
  2. ReasonEmbed: Enhanced Text Embeddings for Reasoning-Intensive Document RetrievalJianlyu Chen, Junwei Lan, Chaofan Li, Defu Lian et al.ACL 2026 · 13 citations
  3. A Survey on Patent Analysis: From NLP to Multimodal AIHomaira Huda Shomee, Zhu Wang, Sathya N. Ravi, Sourav MedyaACL 2025 · 12 citations

Related papers

  1. SciBench: Evaluating College-Level Scientific Problem-Solving Abilities of Large Language ModelsXiaoxuan Wang, Ziniu Hu, Pan Lu, Yanqiao Zhu et al.ICML 2024 · 220 citations
  2. SATBench: Benchmarking LLMs' Logical Reasoning via Automated Puzzle Generation from SAT FormulasAnjiang Wei, Yuheng Wu, Yingjia Wan, Tarun Suresh et al.EMNLP 2025 · 1 citation
  3. LRBench and Judge-R1: Principled Evaluation and Training of LLM-Based Judges for Long-Context ReasoningXinyi Zhao, Haoqi Hu, Ziyu Wang, Jinfeng XiaoACL Findings 2026
  4. EngiBench: A Benchmark for Evaluating Large Language Models on Engineering Problem SolvingXiyuan Zhou, Xinlei Wang, Yirui He, Ruixi Zou et al.ACL Findings 2026 · 2 citations
  5. ProJudge: A Multi-Modal Multi-Discipline Benchmark and Instruction-Tuning Dataset for Mllm-Based Process JudgesJiaxin Ai, Pengfei Zhou, Zhaopan Xu, Ming Li et al.ICCV 2025 · 9 citations
Dusk over the sea between two cliffs drawn in fine vertical lines