Papers
Topics
Authors
Recent
Gemini 2.5 Flash
Gemini 2.5 Flash
129 tokens/sec
GPT-4o
28 tokens/sec
Gemini 2.5 Pro Pro
42 tokens/sec
o3 Pro
4 tokens/sec
GPT-4.1 Pro
38 tokens/sec
DeepSeek R1 via Azure Pro
28 tokens/sec
2000 character limit reached

Cleared for Takeoff? Compositional & Conditional Reasoning may be the Achilles Heel to (Flight-Booking) Language Agents (2404.04237v1)

Published 5 Apr 2024 in cs.CL

Abstract: The rapid progress of LLMs has seen them excel and frequently surpass human performance on standard benchmarks. This has enabled many downstream applications, such as LLM agents, to rely on their sophisticated reasoning to navigate complex task requirements. However, LLMs are known to unexpectedly falter in simple tasks and under seemingly straightforward circumstances - underscoring the need for better and more diverse evaluation setups to measure their true capabilities. To this end, we choose to study compositional and conditional reasoning, two cornerstones of human cognition, and introduce GroundCocoa - a lexically diverse benchmark connecting these reasoning skills to the real-world problem of flight booking. Our task involves aligning detailed user preferences with available flight options presented in a multiple-choice format. Results indicate a significant disparity in performance among current state-of-the-art LLMs with even the best performing model, GPT-4 Turbo, not exceeding 67% accuracy despite advanced prompting techniques.

Definition Search Book Streamline Icon: https://streamlinehq.com
References (59)
  1. MathQA: Towards interpretable math word problem solving with operation-based formalisms. In Jill Burstein, Christy Doran, and Thamar Solorio (eds.), Proceedings of the 2019 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, Volume 1 (Long and Short Papers), pp.  2357–2367, Minneapolis, Minnesota, June 2019. Association for Computational Linguistics. doi: 10.18653/v1/N19-1245. URL https://aclanthology.org/N19-1245.
  2. Abductive commonsense reasoning. In International Conference on Learning Representations, 2020. URL https://openreview.net/forum?id=Byg1v1HKDB.
  3. Piqa: Reasoning about physical commonsense in natural language. In Proceedings of the AAAI conference on artificial intelligence, volume 34, pp.  7432–7439, 2020.
  4. Think you have solved question answering? try arc, the ai2 reasoning challenge. arXiv preprint arXiv:1803.05457, 2018.
  5. Transformers as soft reasoners over language. In Proceedings of the Twenty-Ninth International Conference on International Joint Conferences on Artificial Intelligence, pp.  3882–3890, 2021.
  6. Training verifiers to solve math word problems. arXiv preprint arXiv:2110.14168, 2021.
  7. Mind2web: Towards a generalist agent for the web. Advances in Neural Information Processing Systems, 36, 2024.
  8. Shortcut learning of large language models in natural language understanding. Commun. ACM, 67(1):110–120, dec 2023. ISSN 0001-0782. doi: 10.1145/3596490. URL https://doi.org/10.1145/3596490.
  9. Faith and fate: Limits of transformers on compositionality. In Thirty-seventh Conference on Neural Information Processing Systems, 2023. URL https://openreview.net/forum?id=Fkckkr3ya8.
  10. Multimodal web navigation with instruction-finetuned foundation models. In The Twelfth International Conference on Learning Representations, 2023.
  11. Evaluating models’ local decision boundaries via contrast sets. In Findings of the Association for Computational Linguistics: EMNLP 2020, pp.  1307–1323, 2020.
  12. Shortcut learning in deep neural networks. Nature Machine Intelligence, 2(11):665–673, 2020.
  13. Robustness gym: Unifying the nlp evaluation landscape. In Proceedings of the 2021 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies: Demonstrations, pp.  42–55, 2021.
  14. Folio: Natural language reasoning with first-order logic. arXiv preprint arXiv:2209.00840, 2022. URL https://arxiv.org/abs/2209.00840.
  15. Measuring massive multitask language understanding. In International Conference on Learning Representations, 2021a. URL https://openreview.net/forum?id=d7KBjmI3GmQ.
  16. Measuring mathematical problem solving with the math dataset. NeurIPS, 2021b.
  17. Cogagent: A visual language model for gui agents, 2023.
  18. On the compositional generalization gap of in-context learning. In Jasmijn Bastings, Yonatan Belinkov, Yanai Elazar, Dieuwke Hupkes, Naomi Saphra, and Sarah Wiegreffe (eds.), Proceedings of the Fifth BlackboxNLP Workshop on Analyzing and Interpreting Neural Networks for NLP, pp.  272–280, Abu Dhabi, United Arab Emirates (Hybrid), December 2022. Association for Computational Linguistics. doi: 10.18653/v1/2022.blackboxnlp-1.22. URL https://aclanthology.org/2022.blackboxnlp-1.22.
  19. Cosmos QA: Machine reading comprehension with contextual commonsense reasoning. In Kentaro Inui, Jing Jiang, Vincent Ng, and Xiaojun Wan (eds.), Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing and the 9th International Joint Conference on Natural Language Processing (EMNLP-IJCNLP), pp.  2391–2401, Hong Kong, China, November 2019. Association for Computational Linguistics. doi: 10.18653/v1/D19-1243. URL https://aclanthology.org/D19-1243.
  20. Mistral 7b. arXiv preprint arXiv:2310.06825, 2023.
  21. Mixtral of experts. arXiv preprint arXiv:2401.04088, 2024.
  22. Wilds: A benchmark of in-the-wild distribution shifts. In International Conference on Machine Learning, pp.  5637–5664. PMLR, 2021.
  23. Sgd-x: A benchmark for robust generalization in schema-guided dialogue systems. In Proceedings of the AAAI Conference on Artificial Intelligence, volume 36, pp.  10938–10946, 2022.
  24. Transformers learn shortcuts to automata. In The Eleventh International Conference on Learning Representations, 2022.
  25. Text and patterns: For effective chain of thought, it takes two to tango. arXiv preprint arXiv:2209.07686, 2022.
  26. Sympy: symbolic computing in python. PeerJ Computer Science, 3:e103, January 2017. ISSN 2376-5992. doi: 10.7717/peerj-cs.103. URL https://doi.org/10.7717/peerj-cs.103.
  27. A diverse corpus for evaluating and developing English math word problem solvers. In Dan Jurafsky, Joyce Chai, Natalie Schluter, and Joel Tetreault (eds.), Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics, pp.  975–984, Online, July 2020. Association for Computational Linguistics. doi: 10.18653/v1/2020.acl-main.92. URL https://aclanthology.org/2020.acl-main.92.
  28. Investigating the limitations of transformers with simple arithmetic tasks. arXiv preprint arXiv:2102.13019, 2021.
  29. Cognition and conditionals: Probability and logic in human thought. 2010. URL https://api.semanticscholar.org/CorpusID:124383943.
  30. OpenAI. Gpt-4 technical report. ArXiv, abs/2303.08774, 2023. URL https://arxiv.org/abs/2303.08774.
  31. Anton Osika. gpt-engineer, April 2023. URL https://github.com/gpt-engineer-org/gpt-engineer.
  32. Limitations of language models in arithmetic and symbolic induction. pp.  9285–9298, 01 2023. doi: 10.18653/v1/2023.acl-long.516.
  33. Towards scalable multi-domain conversational agents: The schema-guided dialogue dataset. In Proceedings of the AAAI Conference on Artificial Intelligence, volume 34, pp.  8689–8696, 2020.
  34. Winogrande: An adversarial winograd schema challenge at scale. Commun. ACM, 64(9):99–106, aug 2021. ISSN 0001-0782. doi: 10.1145/3474381. URL https://doi.org/10.1145/3474381.
  35. RobustLR: A diagnostic benchmark for evaluating logical robustness of deductive reasoners. In Yoav Goldberg, Zornitsa Kozareva, and Yue Zhang (eds.), Proceedings of the 2022 Conference on Empirical Methods in Natural Language Processing, pp.  9614–9631, Abu Dhabi, United Arab Emirates, December 2022. Association for Computational Linguistics. doi: 10.18653/v1/2022.emnlp-main.653. URL https://aclanthology.org/2022.emnlp-main.653.
  36. Social IQa: Commonsense reasoning about social interactions. In Kentaro Inui, Jing Jiang, Vincent Ng, and Xiaojun Wan (eds.), Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing and the 9th International Joint Conference on Natural Language Processing (EMNLP-IJCNLP), pp.  4463–4473, Hong Kong, China, November 2019. Association for Computational Linguistics. doi: 10.18653/v1/D19-1454. URL https://aclanthology.org/D19-1454.
  37. Abulhair Saparov and He He. Language models are greedy reasoners: A systematic formal analysis of chain-of-thought. In The Eleventh International Conference on Learning Representations, 2022.
  38. Abulhair Saparov and He He. Language models are greedy reasoners: A systematic formal analysis of chain-of-thought. In The Eleventh International Conference on Learning Representations, 2023. URL https://openreview.net/forum?id=qFVVBzXxR2V.
  39. Testing the general deductive reasoning capacity of large language models using OOD examples. CoRR, abs/2305.15269, 2023. doi: 10.48550/arXiv.2305.15269. URL https://doi.org/10.48550/arXiv.2305.15269.
  40. Toolformer: Language models can teach themselves to use tools. Advances in Neural Information Processing Systems, 36, 2024.
  41. Hugginggpt: Solving ai tasks with chatgpt and its friends in hugging face. Advances in Neural Information Processing Systems, 36, 2024.
  42. Significant Gravitas. AutoGPT. URL https://github.com/Significant-Gravitas/AutoGPT.
  43. Human problem solving: The state of the theory in 1970. American Psychologist, 26:145–159, 1971. URL https://api.semanticscholar.org/CorpusID:14405801.
  44. Why do universal adversarial attacks work on large language models?: Geometry might be the answer. In The Second Workshop on New Frontiers in Adversarial Machine Learning, 2023.
  45. Proofwriter: Generating implications, proofs, and abductive statements over natural language. In Findings of the Association for Computational Linguistics: ACL-IJCNLP 2021, pp.  3621–3634, 2021.
  46. CommonsenseQA: A question answering challenge targeting commonsense knowledge. In Jill Burstein, Christy Doran, and Thamar Solorio (eds.), Proceedings of the 2019 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, Volume 1 (Long and Short Papers), pp.  4149–4158, Minneapolis, Minnesota, June 2019. Association for Computational Linguistics. doi: 10.18653/v1/N19-1421. URL https://aclanthology.org/N19-1421.
  47. Towards benchmarking and improving the temporal reasoning capability of large language models. In Anna Rogers, Jordan Boyd-Graber, and Naoaki Okazaki (eds.), Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), pp.  14820–14835, Toronto, Canada, July 2023. Association for Computational Linguistics. doi: 10.18653/v1/2023.acl-long.828. URL https://aclanthology.org/2023.acl-long.828.
  48. Large language models can be lazy learners: Analyze shortcuts in in-context learning. pp.  4645–4657, 01 2023. doi: 10.18653/v1/2023.findings-acl.284.
  49. Gemini: a family of highly capable multimodal models. arXiv preprint arXiv:2312.11805, 2023.
  50. Diagnosing the first-order logical reasoning ability through LogicNLI. In Proceedings of the 2021 Conference on Empirical Methods in Natural Language Processing, pp.  3738–3747, Online and Punta Cana, Dominican Republic, November 2021. Association for Computational Linguistics. doi: 10.18653/v1/2021.emnlp-main.303. URL https://aclanthology.org/2021.emnlp-main.303.
  51. Llama 2: Open foundation and fine-tuned chat models. arXiv preprint arXiv:2307.09288, 2023.
  52. On the robustness of chatgpt: An adversarial and out-of-distribution perspective. In ICLR 2023 Workshop on Trustworthy and Reliable Large-Scale Machine Learning Models, 2023.
  53. Chain-of-thought prompting elicits reasoning in large language models. Advances in Neural Information Processing Systems, 35:24824–24837, 2022.
  54. Visual chatgpt: Talking, drawing and editing with visual foundation models. arXiv preprint arXiv:2303.04671, 2023.
  55. Can pretrained language models (yet) reason deductively? In Proceedings of the 17th Conference of the European Chapter of the Association for Computational Linguistics, pp.  1439–1454, 2023.
  56. Hellaswag: Can a machine really finish your sentence? In Proceedings of the 57th Annual Meeting of the Association for Computational Linguistics, pp.  4791–4800, 2019.
  57. On the paradox of learning to reason from data. In Proceedings of the 32nd International Joint Conference on Artificial Intelligence (IJCAI), aug 2023. URL http://starai.cs.ucla.edu/papers/ZhangIJCAI23.pdf.
  58. Gpt-4v (ision) is a generalist web agent, if grounded. arXiv preprint arXiv:2401.01614, 2024.
  59. “going on a vacation” takes longer than “going for a walk”: A study of temporal commonsense understanding. In Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing and the 9th International Joint Conference on Natural Language Processing (EMNLP-IJCNLP). Association for Computational Linguistics, 2019.

Summary

We haven't generated a summary for this paper yet.

Ai Sparkles Streamline Icon: https://streamlinehq.com Summarize Now