Papers
Topics
Authors
Recent
Search
2000 character limit reached

CookingSense: A Culinary Knowledgebase with Multidisciplinary Assertions

Published 1 May 2024 in cs.AI and cs.CL | (2405.00523v1)

Abstract: This paper introduces CookingSense, a descriptive collection of knowledge assertions in the culinary domain extracted from various sources, including web data, scientific papers, and recipes, from which knowledge covering a broad range of aspects is acquired. CookingSense is constructed through a series of dictionary-based filtering and LLM-based semantic filtering techniques, which results in a rich knowledgebase of multidisciplinary food-related assertions. Additionally, we present FoodBench, a novel benchmark to evaluate culinary decision support systems. From evaluations with FoodBench, we empirically prove that CookingSense improves the performance of retrieval augmented LLMs. We also validate the quality and variety of assertions in CookingSense through qualitative analysis.

Definition Search Book Streamline Icon: https://streamlinehq.com
References (65)
  1. Uncommonsense: Informative negative knowledge about everyday concepts. In Proceedings of the 31st ACM International Conference on Information & Knowledge Management, CIKM ’22, page 37–46, New York, NY, USA. Association for Computing Machinery.
  2. PIQA: Reasoning about physical commonsense in natural language. In Proceedings of the AAAI conference on artificial intelligence, volume 34, pages 7432–7439.
  3. COMET: Commonsense transformers for automatic knowledge graph construction. In Proceedings of the 57th Annual Meeting of the Association for Computational Linguistics, pages 4762–4779, Florence, Italy. Association for Computational Linguistics.
  4. TLDR: Extreme summarization of scientific documents. In Findings of the Association for Computational Linguistics: EMNLP 2020, pages 4766–4777, Online. Association for Computational Linguistics.
  5. Chatgpt to replace crowdsourcing of paraphrases for intent classification: Higher diversity and comparable model robustness. arXiv preprint arXiv:2305.12947.
  6. The receptors and cells for mammalian taste. Nature, 444(7117):288–294.
  7. KitchenScale: Learning to predict ingredient quantities from recipe contexts. Expert Systems with Applications, 224:120041.
  8. Scaling instruction-finetuned language models. Computing Research Repository, arXiv:2210.11416. Version 5.
  9. Ernest Davis and Gary Marcus. 2015. Commonsense reasoning and commonsense knowledge in artificial intelligence. Communications of the ACM, 58(9):92–103.
  10. BERT: Pre-training of deep bidirectional transformers for language understanding. In 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), pages 4171–4186, Minneapolis, Minnesota. Association for Computational Linguistics.
  11. Andrew Dornenburg and Karen Page. 2008. The Flavor Bible: The Essential Guide to Culinary Creativity, Based on the Wisdom of America’s Most Imaginative Chefs. Little, Brown.
  12. Sema Ekincek and Semra Günay. 2023. A recipe for culinary creativity: Defining characteristics of creative chefs and their process. International Journal of Gastronomy and Food Science, 31:100633.
  13. USDA’s FoodData Central: what is it and why is it needed today? The American journal of clinical nutrition, 115(3):619–624.
  14. FlavorDB: a database of flavor molecules. Nucleic Acids Research, 46(D1):D1210–D1216.
  15. RecipeMind: Guiding ingredient choices from food pairing to recipe completion using cascaded set transformer. In Proceedings of the 31st ACM International Conference on Information & Knowledge Management, CIKM ’22, page 3092–3102, New York, NY, USA. Association for Computing Machinery.
  16. RecipeBowl: A cooking recommender for ingredients and recipes using set transformer. IEEE Access, 9:143623–143633.
  17. Q-Chef: The impact of surprise-eliciting systems on food-related decision-making. In Proceedings of the 2022 CHI Conference on Human Factors in Computing Systems, CHI ’22, New York, NY, USA. Association for Computing Machinery.
  18. FoodKG: a semantics-driven knowledge graph for food recommendation. In The Semantic Web – ISWC 2019, pages 146–162, Cham. Springer International Publishing.
  19. How far can we extract diverse perspectives from large language models? criteria-based diversity prompting! arXiv preprint arXiv:2311.09799.
  20. spaCy: Industrial-strength natural language processing in Python.
  21. DEER: Descriptive knowledge graph for explaining entity relationships. In Proceedings of the 2022 Conference on Empirical Methods in Natural Language Processing, pages 6686–6698, Abu Dhabi, United Arab Emirates. Association for Computational Linguistics.
  22. Towards smart healthcare management based on knowledge graph technology. In Proceedings of the 2019 8th International Conference on Software and Computer Applications, ICSCA ’19, page 330–337, New York, NY, USA. Association for Computing Machinery.
  23. COMET-ATOMIC 2020: On symbolic and neural commonsense knowledge graphs. In Proceedings of the AAAI Conference on Artificial Intelligence, volume 35, pages 6384–6392.
  24. Mistral 7b. arXiv preprint arXiv:2310.06825.
  25. Justin M. Johnson and Taghi M. Khoshgoftaar. 2019. Survey on deep learning with class imbalance. Journal of Big Data, 6(27).
  26. FactKG: Fact verification via reasoning on knowledge graphs. In Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), pages 16190–16206, Toronto, Canada. Association for Computational Linguistics.
  27. Is the suggested food your desired?: Multi-modal recipe recommendation with demand-based knowledge graph. Expert Systems with Applications, 186:115708.
  28. BART: Denoising sequence-to-sequence pre-training for natural language generation, translation, and comprehension. In Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics, pages 7871–7880, Online. Association for Computational Linguistics.
  29. Retrieval-augmented generation for knowledge-intensive NLP tasks. In Advances in Neural Information Processing Systems, volume 33, pages 9459–9474. Curran Associates, Inc.
  30. S2ORC: The semantic scholar open research corpus. In Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics, pages 4969–4983, Online. Association for Computational Linguistics.
  31. J. Kenji López-Alt. 2015. The food lab: better home cooking through science. WW Norton & Company.
  32. Jacqueline B. Marcus. 2013. Culinary nutrition: the science and practice of healthy cooking. Academic Press.
  33. Recipe1M+: A dataset for learning cross-modal embeddings for cooking recipes and food images. IEEE Transactions on Pattern Analysis and Machine Intelligence, 43(1):187–203.
  34. Applications of knowledge graphs for food science and industry. Patterns, 3(5):100484.
  35. Advanced semantics for commonsense knowledge extraction. In Proceedings of the Web Conference 2021, WWW ’21, page 2636–2647, New York, NY, USA. Association for Computing Machinery.
  36. Knowledge graph-based neurodegenerative diseases and diet relationship discovery. Computing Research Repository, arXiv:2109.06123. Version 2.
  37. Can generalist foundation models outcompete special-purpose tuning? case study in medicine. arXiv preprint arXiv:2311.16452.
  38. Training language models to follow instructions with human feedback. In Advances in Neural Information Processing Systems, volume 35, pages 27730–27744. Curran Associates, Inc.
  39. FlavorGraph: a large-scale food-chemical graph for generating food representations and recommending food pairings. Scientific Reports, 11(931).
  40. KitcheNette: Predicting and ranking food ingredient pairings using siamese neural network. In Proceedings of the Twenty-Eighth International Joint Conference on Artificial Intelligence, IJCAI-19, pages 5930–5936. International Joint Conferences on Artificial Intelligence Organization.
  41. Teresa Pizzuti and Giovanni Mirabelli. 2013. FTTO: an example of food ontology for traceability purpose. In 2013 IEEE 7th International Conference on Intelligent Data Acquisition and Advanced Computing Systems (IDAACS), volume 1, pages 281–286. IEEE.
  42. Exploring the limits of transfer learning with a unified text-to-text transformer. Journal of Machine Learning Research, 21(140):1–67.
  43. Diet-Right: A smart food recommendation system. KSII Transactions on Internet and Information Systems, 11(6):2910–2925.
  44. The probabilistic relevance framework: BM25 and beyond. Foundations and Trends® in Information Retrieval, 3(4):333–389.
  45. Assorted, archetypal and annotated two million (3A2M) cooking recipes dataset based on active learning. In Machine Intelligence and Emerging Technologies, pages 188–203, Cham. Springer Nature Switzerland.
  46. Multitask prompted training enables zero-shot task generalization. In International Conference on Learning Representations.
  47. Chakkrit Snae and Michael Bruckner. 2008. FOODS: a food-oriented ontology-driven system. In 2008 2nd ieee international conference on digital ecosystems and technologies, pages 168–176. IEEE.
  48. Food for thought: how nutrition impacts cognition and emotion. npj Science of Food, 1(1):7.
  49. X-scitldr: cross-lingual extreme summarization of scholarly documents. In Proceedings of the 22nd ACM/IEEE Joint Conference on Digital Libraries, pages 1–12.
  50. Gemma: Open models based on gemini research and technology. arXiv preprint arXiv:2403.08295.
  51. Llama 2: Open foundation and fine-tuned chat models. arXiv preprint arXiv:2307.09288.
  52. Improvements to bm25 and language models examined. In Proceedings of the 19th Australasian Document Computing Symposium, ADCS ’14, page 58–65, New York, NY, USA. Association for Computing Machinery.
  53. Sweet apple, company? or food? adjective-centric commonsense knowledge acquisition with taxonomy-guided induction. Knowledge-Based Systems, 280:110988.
  54. Finetuned language models are zero-shot learners. In International Conference on Learning Representations.
  55. Chain-of-thought prompting elicits reasoning in large language models. Advances in neural information processing systems, 35:24824–24837.
  56. A broad-coverage challenge corpus for sentence understanding through inference. In Proceedings of the 2018 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, Volume 1 (Long Papers), pages 1112–1122. Association for Computational Linguistics.
  57. A self-enhancement approach for domain-specific chatbot training via knowledge mining and digest. arXiv preprint arXiv:2311.10614.
  58. RecipeDB: A resource for exploring recipes. Oxford Academic. PID https://cosylab.iiitd.edu.in/recipedb/.
  59. GenericsKB: A knowledge base of generic statements. PID https://allenai.org/data/genericskb.
  60. FooDB. 2020. FooDB Version 1.0. PID https://foodb.ca/.
  61. Refined commonsense knowledge from large-scale web contents. PID https://ascentpp.mpi-inf.mpg.de/.
  62. Extracting Cultural Commonsense Knowledge at Scale. PID https://candle.mpi-inf.mpg.de/.
  63. FORK: A Bite-Sized Test Set for Probing Culinary Cultural Biases in Commonsense Reasoning Models. Association for Computational Linguistics. PID https://github.com/shramay-palta/FORK_ACL2023.
  64. Commonsense properties from query logs and question answering forums. PID https://quasimodo.mpi-inf.mpg.de/.
  65. Conceptnet 5.5: An open multilingual graph of general knowledge. PID https://conceptnet.io/.
Citations (1)
View on Semantic Scholar

Summary

  • The paper presents CookingSense, a large-scale culinary knowledgebase that integrates diverse data sources to enhance culinary decision-making using the FoodBench framework.
  • The evaluation demonstrates that CookingSense outperforms traditional knowledgebases in tasks such as flavor prediction and culinary question answering with modern language models.
  • The work offers practical benefits for personalized recipe recommendations, culinary education, and dietary management, paving the way for AI-driven cooking assistants.

Exploring CookingSense: A Versatile Culinary Knowledgebase

Introduction to CookingSense

The paper presents "CookingSense," a novel large-scale culinary knowledgebase (KB) that combines various sources to cover a broad spectrum of cooking-related information. This initiative addresses the existing challenge where most culinary KBs are focused on narrow aspects like recipes or nutrition information. CookingSense integrates data from web content, academic papers, and user-generated recipes, presenting an enriched, versatile platform for culinary information.

Key Contributions and Evaluations

Key Contributions:

  • Construction of Cooking Sense: A comprehensive dataset compiled from diverse sources, providing a rich mix of culinary facts and details beneficial for various culinary applications.
  • Development of FoodBench: A benchmarking framework created to evaluate the utility of CookingSense in supporting culinary decisions.
  • Effectiveness of CookingSense: Demonstrated through comparative analysis with other KBs, using the FoodBench framework underpinned by recent LLMs.

Evaluations:

  • The performance enhancement via CookingSense is evident when integrated with LLMs, showing improved decision-making capabilities in culinary contexts. This is quantitatively supported through various benchmark tasks like flavor prediction and culinary question answering, where CookingSense outperforms traditional and domain-specific KBs.

Practical Implications

The practical usability of CookingSense extends to several real-world applications:

  • Enhanced Recipe Recommendation Systems: By understanding diverse culinary techniques and ingredients, systems can offer more tailored recommendations.
  • Educational Tools for Culinary Training: CookingSense can serve as a comprehensive resource for students and professionals looking to deepen their culinary knowledge.
  • Support for Dietary Management: With broad coverage on nutrition and health, it can aid in designing better dietary plans tailored to individual health needs.

Theoretical Implications

From a theoretical perspective, CookingSense contributes to the body of knowledge in AI by demonstrating how multifaceted KBs can significantly enhance the performance of decision support systems in specialized domains such as cooking. It highlights the importance of integrating diverse data sources to avoid biases and ensure comprehensive coverage.

Future Directions

Looking forward, the enhancement of CookingSense could involve:

  • Integration of More Diverse Data Sources: Including more culturally and linguistically varied data to cover global culinary practices more comprehensively.
  • Application in AI-Driven Cooking Assistants: Utilizing this KB to support AI in performing more complex tasks like dynamic recipe adaptation or ingredient substitution based on available kitchen inventory or dietary restrictions.

Conclusion and Ethical Considerations

CookingSense sets a foundation for future research and development in culinary AI applications, proposing a robust framework that can be scaled and adapted for various needs. Ethically, the developers of CookingSense are urged to continuously monitor and filter the data to avoid potential biases that could arise from its diverse data sources, ensuring the KB remains a reliable and neutral resource. By addressing these aspects, CookingSense can become an indispensable tool for both AI researchers and culinary professionals, contributing positively to technological advancements in the culinary arts.

Paper to Video (Beta)

No one has generated a video about this paper yet.

Sign Up to Generate All Videos Subscribe on YouTube

Whiteboard

No one has generated a whiteboard explanation for this paper yet.

Sign Up to Generate

Open Problems

We haven't generated a list of open problems mentioned in this paper yet.

Generate Now

Collections

Sign up for free to add this paper to one or more collections.

Sign Up

Tweets

Sign up for free to view the 6 tweets with 11 likes about this paper.

Sign Up for Free