Pith. sign in

REVIEW 3 major objections 4 minor 66 references

KERL: Knowledge-Enhanced Personalized Recipe Recommendation using Large Language Models

T0 review · 3 major / 4 minor · reviewed 2026-08-07 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗

Pith's one-line read A compact LLM grounded in a food knowledge graph beats much larger models on recipe recommendation, generation, and nutrition.

desk verdict Solid recipe and nutrition modules; the recommendation headline is confounded by an unfair context comparison and needs a matched baseline before the strong claims hold. read the letter →

arxiv 2505.14629 v1 pith:MLM7BOY4 submitted 2025-05-20 cs.LG cs.AIcs.CLcs.CV

Fnu Mohbat , Mohammed J Zaki This is my paper
classification cs.LGcs.AIcs.CLcs.CV
keywords foodrecommendationknowledgegraphlargelanguagemodelsLoRAretrieval-augmentedgenerationrecipenutritionestimationquestionansweringovergraphs
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

The paper tries to establish that a single compact LLM—Phi-3-mini—combined with the FoodKG knowledge graph and three task-specific LoRA adapters can handle the full job of personalized meal assistance: recommending recipes under ingredient and nutrition constraints, writing cooking instructions, and reporting micronutrient values. On the authors' newly built 77,900-question benchmark, the recommendation adapter reaches F1 0.973, a 56-point gain over the 128K-context Phi-3-mini base and a 26-point gain over the much larger Llama-2-7B. The recipe and nutrition adapters also beat their baselines, including the multimodal LLaVA-Chef. Because the authors release their code and benchmark datasets, the claim is checkable rather than merely asserted.

What carries the argument

The load-bearing mechanism is retrieval-then-rank: a template-driven entity parser and SPARQL generator first restrict the universe to recipes carrying one of 15 health tags and matching ingredient filters, and the LLM's job is to select, from those retrieved subgraphs, the recipes that satisfy all constraints. Each of the three tasks—recommendation, recipe generation, and nutrition generation—is a separate LoRA (low-rank adaptation) adapter on the same Phi-3-mini-128K backbone, so one deployed base model can switch adapters per task and run efficiently on a single GPU. The knowledge graph supplies both the candidate set and the ground-truth answers, which is what lets a small fine-tuned model outrank much larger pretrained baselines.

What would settle it

Construct a test set of paraphrased and free-form versions of the same constraints—using synonyms for ingredients, alternative tag names, and misspellings—and measure how often the template SPARQL retrieval returns a correct candidate set before the LLM is asked to rank. If a large fraction of such inputs yield empty or wrong retrieval, the reported F1 on template-generated questions would not reflect real-world use, and the central claim that the system is a complete personalized recommender would be undermined.

Watch

Extended reading notes

Core claim

The central claim is that retrieving recipe subgraphs from FoodKG and feeding them to an LLM as context—rather than asking the model to answer from memory—is what makes constrained food question answering work. KERL-Recom parses a user query into a health tag, liked ingredients, disliked ingredients, and numeric nutrient limits, converts these into a template SPARQL query, retrieves the matching recipe subgraphs, and is trained with a LoRA adapter to output only recipes that satisfy every condition. The paper reports F1 0.973 on its KGQA test set versus 0.41 for Phi-3-mini-128K and 0.713 for Llama-2-7B, and on the older pFoodReq benchmark it reports F1 0.854 versus 0.637 for pFoodReq itself. For generation, a second adapter produces cooking steps with higher BLEU, ROUGE, METEOR, and CIDEr scores than LLaVA-Chef while training far fewer parameters, and a third adapter lowers mean average error across ten micronutrients, especially when given the title, ingredients, and instructions together.

Load-bearing premise

The system assumes every user question can be expressed through fixed templates using one of 15 health tags and the exact ingredient names stored in FoodKG; since the benchmark questions are generated by filling those same templates, the test never exposes the system to free-form wording, synonyms, misspellings, or ingredient substitutions, and a real question that deviates can produce an empty or wrong candidate set that the model cannot recover from.

Editorial extensions

If this is right

  • A 3.8B-parameter model with three task-specific adapters can outperform much larger pretrained LLMs on constrained recipe recommendation, indicating that knowledge-graph grounding plus task-specific tuning can matter more than parameter count for this task.
  • The same base model with switchable adapters can serve recommendation, cooking-step writing, and nutrition analysis in one deployment, which is more practical than running separate specialized systems.
  • Retrieving subgraphs from FoodKG first means any recommended recipe already satisfies the stated ingredient and nutrient constraints whenever such a recipe exists in the graph, giving hard constraint satisfaction rather than free-form guesswork.
  • The new 77,900-question KGQA benchmark is over an order of magnitude larger than the previous pFoodReq benchmark, giving the food-KG community a substantially bigger testbed for constrained question answering.
  • On the independent pFoodReq benchmark, KERL-Recom improves F1 by 21.7 points over the pFoodReq method itself, suggesting the approach transfers beyond the authors' own template-generated test set.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • Editorial inference: Because the benchmark questions are generated by filling the same templates the system parses, the reported F1 may overestimate performance on free-form user language; a human-written or paraphrased test set would show whether the KG grounding helps when entity parsing is harder.
  • Editorial inference: The multi-LoRA-plus-KG recipe could transfer to other constrained recommendation domains—such as travel, shopping, or medical meal planning—where a knowledge graph supplies candidates and each task gets its own adapter.
  • Editorial inference: The system's coverage is bounded by FoodKG's tag and ingredient vocabulary; enriching the graph or adding an ingredient-substitution mapping would extend the system to queries the current templates cannot parse.
  • Editorial inference: Because the recommendation module can only select within retrieved subgraphs, any SPARQL retrieval failure is unrecoverable; training a learned retriever on free-form questions could make the pipeline more robust than fixed templates.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

3 major / 4 minor

Summary. The paper presents KERL, a unified food system built on Phi-3-mini with separate LoRA adapters for recipe recommendation (KERL-Recom), recipe generation (KERL-Recipe), and nutrition generation (KERL-Nutri). Recommendation is framed as constrained question answering over FoodKG: template-based entity and constraint parsing generates SPARQL queries, retrieved recipe subgraphs are serialized into the LLM context, and the LoRA adapter selects recipe names satisfying the constraints. The authors also construct a 77,900-question KGQA benchmark from FoodKG tags, ingredients, and nutrient thresholds, and evaluate recipe-generation and nutrition benchmarks derived from Recipe1M. The headline result is that KERL-Recom reaches F1 0.973 on the KGQA test set, outperforming pretrained LLMs by 26-56 F1 points, and that the recipe and nutrition adapters improve over Phi-3 baselines.

Significance. If the recommendation comparison were clean, KERL would be a practical unified system and the KGQA benchmark would be a useful resource. Strengths include public code and benchmark datasets, a multi-LoRA design that shares one base model, and standard, clearly described recipe-generation and nutrition experiments. However, the central evidence for KG-augmented recommendation is currently confounded: the recommendation model is tested with retrieved KG context while the baselines are not, and no deterministic KG-query control is reported. The core scientific claim about the value of KG augmentation therefore requires additional experiments; with those controls the paper could make a solid contribution.

major comments (3)
  1. [§5.2, Table 5] The comparison between KERL-Recom and pretrained LLMs is not apples-to-apples. Section 3.1 states that KERL-Recom receives, for each inference call, the user query plus a serialized subgraph containing candidate recipe names, ingredient lists, and nutrient values retrieved by SPARQL from FoodKG. The baselines in Table 5 are described only as pretrained LLMs, and the paper does not state that they are given the same subgraph context; the natural reading is that they receive the question alone and must recall recipe titles from parametric memory. Because the KGQA benchmark is generated by the same template/constraint pipeline (Sections 4.1-4.2), every ground-truth recipe name and all values needed for constraint filtering are present in KERL-Recom's context, so a zero-shot model with that context could be near-perfect without LoRA. The reported 56-point and 26-point F1 gains therefore do not yet establish that the KG-augmented LLM, rather than the retrieval step, drives the improvement. Please add (i) a context-matched zero-shot baseline: the same serialized subgraphs and query given to Phi-3-mini, and ideally to the other LLMs, without LoRA; and (ii) an ablation with the same context but no adapter.
  2. [§5.2, Table 5; §4.1-4.2] No deterministic KG-query baseline is reported. Since R+(tj) is defined by the very ingredient-inclusion/exclusion and nutrient-threshold filters encoded in the SPARQL templates, a SPARQL query plus post-filter over FoodKG would be near-perfect by construction on this test set. Reporting this baseline is necessary to calibrate how much value the LLM selection adds over straightforward retrieval. As the paper stands, the conclusion that 'our proposed KG-augmented LLM significantly outperforms existing approaches' (Abstract) is not supported without this control.
  3. [§4.1-4.2 and §7] The benchmark's template-generated queries do not support the unqualified claim of a complete and coherent solution for personalized food recommendation. Every test question is generated by filling placeholders with exact FoodKG ingredient names and one of 15 closed-set tags, and the system parses entities from these same templates. Section 7 itself concedes that the system fails when context is empty or wrong. The claims about personalization would be strengthened by at least a small evaluation on paraphrased or free-form queries, synonym or misspelling variants, or tags outside the fixed list; as written, the system's behavior on realistic open-ended user input is untested.
minor comments (4)
  1. [Appendix D.1, Table 13] The per-tag results for KERL-Recom are labeled 'KERL-Nutri' in the model column (e.g., the lactose row with F1=0.935), which is inconsistent with the surrounding text and should be corrected.
  2. [Table 7] The row label 'KERN-Recom' is a typo for KERL-Recom.
  3. [Table 8] The Phi-3 Xt+Xing row is malformed: the numeric columns are merged ('0.0380.2160.042') and the final perplexity entry is incomplete ('0. 11.99'), making that row unreadable.
  4. [Introduction and Section 3.1] There are copyediting errors, including 'taks' for 'task' in the Introduction and 'The KERL-Recipe modules generates' in Section 3.1; these should be fixed.

Circularity Check

0 steps flagged · score 0.0 of 10

No circular derivation found; the main concern is an evaluation asymmetry, not a circularity.

full rationale

KERL is an empirical systems paper rather than a formal derivation, and I could not exhibit any equation or fitted parameter that reduces to its own inputs. The closest candidate is KERL-Recom's evaluation setup: the benchmark answers R+(tj) are defined as the recipes satisfying the template-generated constraints (Section 4.1, 'Recipes that satisfy all constraints are considered ground truth answers'), and the inference context is retrieved from FoodKG using the same constraint-based SPARQL templates (Section 3.1, 'These entities are then used to generate SPARQL queries based on predefined templates'). This creates a legitimate evaluation-confounding concern: a deterministic SPARQL/post-filter baseline over FoodKG would likely achieve near-perfect F1, and the pretrained LLM baselines do not receive the retrieved subgraph, so Table 5's 56-point and 26-point F1 gains may reflect an asymmetric comparison rather than a learned KG-augmentation advantage. However, that is a missing-control / benchmark-validity issue, not a circular derivation in the taxonomy used here. No fitted parameter is renamed as a prediction, no self-cited uniqueness theorem is invoked to force a choice, and no ansatz is smuggled in via citation. The self-citations that exist (FoodKG, LLaVA-Chef, pFoodReq) are used as external knowledge sources, baseline systems, and prior benchmarks, not as unverified premises that the present argument depends on; the recipe and nutrition evaluations also rely on externally defined Recipe1M splits. The paper's own Limitations (Section 7) acknowledge failures when the KG context is empty or incorrect, which further confirms the system's behavior is driven by the retrieved context rather than by a circularly defined target. I therefore find no significant circularity and assign a score of 0.

Assumptions & free parameters 3 free parameters · 4 assumptions · 0 invented entities

No new physical or model entities are introduced. The central claim rests on two external datasets, a template-generated benchmark distribution, and human-chosen training hyperparameters. The FoodKG and pFoodReq resources were created in prior work involving one of the authors, which increases the importance of independent baselines.

free parameters (3)
  • K (context subset size) = at most K, with K/2 positives and K/2 negatives; exact value not stated
    Training and inference pass a subset of R(tj) to the LLM; K is chosen so the model fits in GPU memory and directly determines how many candidate recipes the model sees (Section 3.1).
  • LoRA configuration = r=64, alpha=16, dropout=0.5
    The same hyperparameters are used for all three adapters with no ablation, so their effect on the reported gains is not isolated (Section 5.1).
  • Nutritional constraint thresholds = sampled within mu +/- 2 sigma or ranges such as (0, x)
    Used to synthesize benchmark questions; different sampling would change question difficulty and answer sets (Section 4.1).
assumptions (4)
  • domain assumption FoodKG is a correct and sufficiently complete source of recipes, ingredients, tags, and nutrition values.
    All recommendation results are read from FoodKG subgraphs; errors or missing recipes are inherited directly (Sections 3.1 and 7).
  • domain assumption Recipe1M instruction and nutrient annotations are adequate ground truth for generation and nutrition tasks.
    Recipe and nutrition adapters are trained and evaluated against Recipe1M labels; the authors note extreme nutrition cases may be inaccurate (Sections 4.3, 4.4, and 7).
  • ad hoc to paper Template-generated constrained questions are a valid proxy for real personalized queries.
    The KGQA benchmark fills templates with random ingredient sets and nutrient limits, and ground truth is defined by the same constraints; this is the main test distribution (Sections 4.1 and 4.2).
  • domain assumption Entity parsing maps user text to one of 15 health tags and exact FoodKG ingredient names.
    SPARQL generation assumes exact entity names; parser accuracy is not separately evaluated (Section 3.1).

how reviews work

0 comments
Cite this review

Pith. "Pith review of KERL: Knowledge-Enhanced Personalized Recipe Recommendation using Large Language Models." pith.science (2026). https://pith.science/paper/MLM7BOY4

@misc{pith2026250514629,
  author       = {Pith},
  title        = {Pith review of: KERL: Knowledge-Enhanced Personalized Recipe Recommendation using Large Language Models},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/MLM7BOY4}},
  note         = {Machine review of arXiv:2505.14629}
}
read the original abstract

Recent advances in large language models (LLMs) and the abundance of food data have resulted in studies to improve food understanding using LLMs. Despite several recommendation systems utilizing LLMs and Knowledge Graphs (KGs), there has been limited research on integrating food related KGs with LLMs. We introduce KERL, a unified system that leverages food KGs and LLMs to provide personalized food recommendations and generates recipes with associated micro-nutritional information. Given a natural language question, KERL extracts entities, retrieves subgraphs from the KG, which are then fed into the LLM as context to select the recipes that satisfy the constraints. Next, our system generates the cooking steps and nutritional information for each recipe. To evaluate our approach, we also develop a benchmark dataset by curating recipe related questions, combined with constraints and personal preferences. Through extensive experiments, we show that our proposed KG-augmented LLM significantly outperforms existing approaches, offering a complete and coherent solution for food recommendation, recipe generation, and nutritional analysis. Our code and benchmark datasets are publicly available at https://github.com/mohbattharani/KERL.

Figures

Figures reproduced from arXiv: 2505.14629 by the authors.

Figure 1
Figure 1. KERL Overview: Given a natural language question (with constraints), the system parses entities and generates a SPARQL query to retrieve a subgraph from the KG. The question and this subgraph as context, are given as input to the recommendation model (KERL-Recom), which generates a list of recipe names that satisfy the constraints. The KERL-Recipe and KERL-Nutri models then generate cooking steps and micro-nutrients… view at source ↗
Figure 2
Figure 2. KERL Multi-LoRA Setup: With the same base model, a separate LoRA adapter is trained for each task. Dur￾ing inference, the desired adapter is activated while base model remains the same. 3 KERL: Food Recommendation System We propose KERL, a personalize food recommen￾dation system that unifies recommendation with cooking steps and nutrition details generation by leveraging multi-LoRA approach as illustrated in [PITH_… view at source ↗
Figure 3
Figure 3. F1 scores of different models across various recipe types. Our model, KERL-Recom, consistently [PITH_FULL_IMAGE:figures/full_fig_p008_3.png] view at source ↗
Figures from the paper (1 more)
Figure 4
Figure 4. Figure 4: FoodKG Recipe sample: left panel shows a 2-hop KG subgraph for the recipe node shown on the right. P = |T P| |T P| + |F P| R = |T P| |T P| + |F N| F1 = 2P R P + R We compute precision at rank r for all relevant recipes M and average them to get average preci￾sion (AP).…

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

66 extracted references · 48 canonical work pages

  1. [1]

    Marah Abdin, Sam Ade Jacobs, Ammar Ahmad Awan, Jyoti Aneja, Ahmed Awadallah, Hany Awadalla, Nguyen Bach, Amit Bahree, Arash Bakhtiari, Harkirat Behl, et al. 2024. Phi-3 technical report: A highly capable language model locally on your phone. arXiv preprint arXiv:2404.14219

    arXiv 2024

  2. [2]

    Josh Achiam, Steven Adler, Sandhini Agarwal, Lama Ahmad, Ilge Akkaya, Florencia Leoni Aleman, Diogo Almeida, Janko Altenschmidt, Sam Altman, Shyamal Anadkat, et al. 2023. Gpt-4 technical report

    work page 2023

  3. [3]

    Caio Viktor S Avila, V \^a nia MP Vidal, Wellington Franco, and Marco A Casanova. 2024. Experiments with text-to-sparql based on chatgpt. In The 18th International Conference on Semantic Computing, pages 277--284. IEEE

    work page 2024

  4. [4]

    Debayan Banerjee, Sushil Awale, Ricardo Usbeck, and Chris Biemann. 2023. Dblp-quad: A question answering dataset over the dblp scholarly knowledge graph. 13th International Workshop on Bibliometric-enhanced Information Retrieval

    work page 2023

  5. [5]

    Zheng Cai, Maosong Cao, Haojiong Chen, Kai Chen, Keyu Chen, Xin Chen, Xun Chen, Zehui Chen, Zhi Chen, Pei Chu, et al. 2024. Internlm2 technical report. arXiv preprint arXiv:2403.17297

    arXiv 2024

  6. [6]

    Jing-Jing Chen, Chong-Wah Ngo, Fu-Li Feng, and Tat-Seng Chua. 2018. Deep understanding of cooking procedure for cross-modal recipe retrieval. In Proceedings of the 26th ACM International Conference on Multimedia

    work page 2018

  7. [7]

    Yu Chen, Ananya Subburathinam, Ching-Hua Chen, and Mohammed J Zaki. 2021. Personalized food recommendation as constrained question answering over a large-scale food knowledge graph. In Proceedings of the 14th ACM International Conference on Web Search and Data Mining, pages 544--552

    work page 2021

  8. [8]

    Prateek Chhikara, Dhiraj Chaurasia, Yifan Jiang, Omkar Masur, and Filip Ilievski. 2024. Fire: Food image to recipe generation. In Proceedings of the Winter Conference on Applications of Computer Vision, pages 8184--8194

    work page 2024

Show all 66 references
  1. [9]

    Damion M Dooley, Emma J Griffiths, Gurinder S Gosal, Pier L Buttigieg, Robert Hoehndorf, Matthew C Lange, Lynn M Schriml, Fiona SL Brinkman, and William WL Hsiao. 2018. Foodon: a harmonized food ontology to increase global food traceability, quality control and data integratio...

    2018

  2. [10]

    Desmond Elliott and Frank Keller. 2013. Image description using visual dependency representations. In Proceedings of the Conference on Empirical Methods in Natural Language Processing

    2013

  3. [11]

    Swetha Eppalapally, Daksh Dangi, Chaithra Bhat, Ankita Gupta, Ruiyi Zhang, Shubham Agarwal, Karishma Bagga, Seunghyun Yoon, Nedim Lipka, Ryan A Rossi, et al. 2024. Kapqa: Knowledge-augmented product question-answering. Proceedings of the 3rd Workshop on Knowledge Augmented Met...

    2024

  4. [12]

    Dec 16, 2023

    Mehrdad Farahani, Kartik Godawat, Haswanth Aekula, Deepak Pandian, and Nicholas Broad. Dec 16, 2023. Chef transformer. https://huggingface.co/flax-community/t5-recipe-generation

    2023

  5. [13]

    Xiaoyan Gao, Fuli Feng, Heyan Huang, Xian-Ling Mao, Tian Lan, and Zewen Chi. 2022. Food recommendation with graph convolutional network. Information Sciences, 584:170--183

    2022

  6. [14]

    Shijie Geng, Shuchang Liu, Zuohui Fu, Yingqiang Ge, and Yongfeng Zhang. 2022. Recommendation as language processing (rlp): A unified pretrain, personalized prompt & predict paradigm (p5). In Proceedings of the 16th ACM Conference on Recommender Systems, pages 299--315

    2022

  7. [15]

    Aaron Grattafiori, Abhimanyu Dubey, Abhinav Jauhri, Abhinav Pandey, Abhishek Kadian, Ahmad Al-Dahle, Aiesha Letman, Akhil Mathur, Alan Schelten, Alex Vaughan, et al. 2024. The llama 3 herd of models. arXiv preprint arXiv:2407.21783

    2024 arXiv

  8. [16]

    Lee, Ke Shu, Palakorn Achananuparp, Philips Kokoh Prasetyo, Yue Liu, Ee-Peng Lim, and Lav R Varshney

    Helena H. Lee, Ke Shu, Palakorn Achananuparp, Philips Kokoh Prasetyo, Yue Liu, Ee-Peng Lim, and Lav R Varshney. 2020. Recipegpt: Generative pre-training based cooking recipe generation and evaluation system. In Companion Proceedings of the Web Conference

    2020

  9. [17]

    Steven Haussmann, Oshani Seneviratne, Yu Chen, Yarden Ne’eman, James Codella, Ching-Hua Chen, Deborah L McGuinness, and Mohammed J Zaki. 2019. Foodkg: a semantics-driven knowledge graph for food recommendation. In The Semantic Web--ISWC: 18th International Semantic Web Confere...

    2019

  10. [18]

    Xiaoxin He, Yijun Tian, Yifei Sun, Nitesh Chawla, Thomas Laurent, Yann LeCun, Xavier Bresson, and Bryan Hooi. 2024. G-retriever: Retrieval-augmented generation for textual graph understanding and question answering. Advances in Neural Information Processing Systems, 37:132876--132907

    2024

  11. [19]

    Yu Hou and Rui Zhang. 2024. Enhancing dietary supplement question answer via retrieval-augmented generation (rag) with llm. medRxiv, pages 2024--09

    2024

  12. [20]

    Edward J Hu, Yelong Shen, Phillip Wallis, Zeyuan Allen-Zhu, Yuanzhi Li, Shean Wang, Lu Wang, and Weizhu Chen. 2022. https://openreview.net/forum?id=nZeVKeeFYf9 Lo RA : Low-rank adaptation of large language models . In International Conference on Learning Representations

    2022

  13. [21]

    Mojan Javaheripi, S \'e bastien Bubeck, Marah Abdin, Jyoti Aneja, Sebastien Bubeck, Caio C \'e sar Teodoro Mendes, Weizhu Chen, Allie Del Giorno, Ronen Eldan, Sivakanth Gopi, et al. 2023. Phi-2: The surprising power of small language models. Microsoft Research Blog, 1:3

    2023

  14. [22]

    Albert Q Jiang, Alexandre Sablayrolles, Arthur Mensch, Chris Bamford, Devendra Singh Chaplot, Diego de las Casas, Florian Bressand, Gianna Lengyel, Guillaume Lample, Lucile Saulnier, et al. 2023. Mistral 7b. arXiv preprint arXiv:2310.06825

    2023 arXiv

  15. [23]

    Daniel Kirk, Elise van Eijnatten, and Guido Camps. 2023. Comparison of answers between chatgpt and human dieticians to common nutrition questions. Journal of Nutrition and Metabolism, 2023:5548684

    2023

  16. [24]

    Akio Kobayashi, Shotaro Mori, Akira Hashimoto, Tetsuo Katsuragi, and Takahiro Kawamura. 2024. Functional food knowledge graph-based recipe recommendation system focused on lifestyle-related diseases. In 18th International Conference on Semantic Computing, pages 261--268. IEEE

    2024

  17. [25]

    Xin Lai, Zhuotao Tian, Yukang Chen, Yanwei Li, Yuhui Yuan, Shu Liu, and Jiaya Jia. 2024. Lisa: Reasoning segmentation via large language model. In Proceedings of the Conference on Computer Vision and Pattern Recognition, pages 9579--9589

    2024

  18. [26]

    Diya Li and Mohammed J Zaki. 2022. Food knowledge representation learning with adversarial substitution. In Proceedings of the 2nd Conference of the Asia-Pacific Chapter of the Association for Computational Linguistics and the 12th International Joint Conference on Natural Lan...

    2022

  19. [27]

    Diya Li, Mohammed J Zaki, and Ching-hua Chen. 2023. Health-guided recipe recommendation over knowledge graphs. Journal of Web Semantics, 75:100743

    2023

  20. [28]

    Chin-Yew Lin. 2004. Rouge: A package for automatic evaluation of summaries. In Text summarization branches out

    2004

  21. [29]

    Chin-Yew Lin and Franz Josef Och. 2004. Automatic evaluation of machine translation quality using longest common subsequence and skip-bigram statistics. In Proceedings of the 42nd Annual Meeting of the Association for Computational Linguistics, pages 605--612

    2004

  22. [30]

    a rbel Kn \

    Yabo Ling, Jian-Yun Nie, Daiva Nielsen, B \"a rbel Kn \"a uper, Nathan Yang, and Laurette Dub \'e . 2022. Following good examples-health goal-oriented food recommendation based on behavior data. In Proceedings of the ACM Web Conference 2022, pages 3745--3754

    2022

  23. [31]

    Guoshan Liu, Hailong Yin, Bin Zhu, Jingjing Chen, Chong-Wah Ngo, and Yu-Gang Jiang. 2025. Retrieval augmented recipe generation. In 2025 Winter Conference on Applications of Computer Vision, pages 2453--2463. IEEE

    2025

  24. [32]

    Haotian Liu, Chunyuan Li, Qingyang Wu, and Yong Jae Lee. 2024. Visual instruction tuning. Advances in Neural Information Processing Systems, 36

    2024

  25. [33]

    Linhao Luo, Yuan-Fang Li, Gholamreza Haffari, and Shirui Pan. 2023. Reasoning on graphs: Faithful and interpretable large language model reasoning. arXiv preprint arXiv:2310.01061

    2023 arXiv

  26. [34]

    Peihua Ma, Shawn Tsai, Yiyang He, Xiaoxue Jia, Dongyang Zhen, Ning Yu, Qin Wang, Jaspreet KC Ahuja, and Cheng-I Wei. 2024. Large language models in food science: Innovations, applications, and future. Trends in Food Science & Technology, page 104488

    2024

  27. [35]

    Puneet Mathur, Zhe Liu, Ke Li, Yingyi Ma, Gil Karen, Zeeshan Ahmed, Dinesh Manocha, and Xuedong Zhang. 2024. Doc-rag: Asr language model personalization with domain-distributed co-occurrence retrieval augmentation. In Proceedings of the Joint International Conference on Comput...

    2024

  28. [36]

    Weiqing Min, Chunlin Liu, Leyi Xu, and Shuqiang Jiang. 2022. Applications of knowledge graphs for food science and industry. Patterns, 3

    2022

  29. [37]

    Fnu Mohbat and Mohammed J Zaki. 2024. Llava-chef: A multi-modal generative model for food recipes. In ACM International Conference on Information and Knowledge Management

    2024

  30. [38]

    Michael Moor, Qian Huang, Shirley Wu, Michihiro Yasunaga, Yash Dalmia, Jure Leskovec, Cyril Zakka, Eduardo Pontes Reis, and Pranav Rajpurkar. 2023. Med-flamingo: a multimodal medical few-shot learner. In Machine Learning for Health (ML4H)

    2023

  31. [39]

    Kishore Papineni, Salim Roukos, Todd Ward, and Wei-Jing Zhu. 2002. Bleu: a method for automatic evaluation of machine translation. In Proceedings of the 40th annual meeting of the Association for Computational Linguistics

    2002

  32. [40]

    Matt Post. 2018. A call for clarity in reporting BLEU scores. In Proceedings of the Third Conference on Machine Translation: Research Papers, pages 186--191, Brussels, Belgium. Association for Computational Linguistics

    2018

  33. [41]

    Zhixiao Qi, Yijiong Yu, Meiqi Tu, Junyi Tan, and Yongfeng Huang. 2023. Foodgpt: A large language model in food testing domain with incremental pre-training and knowledge graph prompt. arXiv preprint arXiv:2308.10173

    2023 arXiv

  34. [42]

    Alec Radford, Jeffrey Wu, Rewon Child, David Luan, Dario Amodei, Ilya Sutskever, et al. 2019. Language models are unsupervised multitask learners. OpenAI blog

    2019

  35. [43]

    Julio C Rangel, Tarcisio Mendes de Farias, Ana Claudia Sima, and Norio Kobayashi. 2024. Sparql generation: an analysis on fine-tuning openllama for question answering over a life science knowledge graph. arXiv preprint arXiv:2402.04627

    2024 arXiv

  36. [44]

    Muhammad Saad Razzaq, Fahad Maqbool, Muhammad Ilyas, and Hajira Jabeen. 2023. Evorecipes: A generative approach for evolving context-aware recipes. IEEE Access

    2023

  37. [45]

    Anja Reusch, Alexander Weber, Maik Thiele, and Wolfgang Lehner. 2021. Recipegm: A hierarchical recipe generation model. In The 37th International Conference on Data Engineering Workshops, pages 24--29. IEEE

    2021

  38. [46]

    Ali Rostami, Ramesh Jain, and Amir M Rahmani. 2024. Food recommendation as language processing (f-rlp): A personalized and contextual paradigm. arXiv preprint arXiv:2402.07477

    2024 arXiv

  39. [47]

    Amaia Salvador, Nicholas Hynes, Yusuf Aytar, Javier Marin, Ferda Ofli, Ingmar Weber, and Antonio Torralba. 2017. Learning cross-modal embeddings for cooking recipes and food images. In Proceedings of the Conference on Computer Vision and Pattern Recognition. IEEE

    2017

  40. [48]

    Sola S Shirai, Oshani Seneviratne, Ching-Hua Chen, Daniel M Gruen, and Deborah L McGuinness. 2021. Healthy food recommendation and explanation generation using a semantically-enabled framework? In International Semantic Web Conference: Posters, Demos, and Industry Tracks. CEUR-WS

    2021

  41. [49]

    Jiashuo Sun, Chengjin Xu, Lumingyuan Tang, Saizhuo Wang, Chen Lin, Yeyun Gong, Heung-Yeung Shum, and Jian Guo. 2023. Think-on-graph: Deep and responsible reasoning of large language model with knowledge graph. arXiv preprint arXiv:2307.07697

    2023 arXiv

  42. [50]

    Tilahun Abedissa Taffa and Ricardo Usbeck. 2023. Leveraging llms in scholarly knowledge graph question answering. In Scholarly QALD Challenge at The 22nd International Semantic Web Conference

    2023

  43. [51]

    Hikaru Tanabe and Keiji Yanai. 2024. Caloriellava: Image-based calorie estimation with multimodal large language models. In Proceedings of the Proceedings of ICPR Workshop on Multimedia Assisted Dietary Management, Kolkata, India, volume 1

    2024

  44. [52]

    Hikaru Tanabe and Keiji Yanai. 2025. Calorievol: Integrating volumetric context into multimodal large language models for image-based calorie estimation. In International Conference on Multimedia Modeling, pages 353--365. Springer

    2025

  45. [53]

    Karan Taneja, Richard Segal, and Richard Goodwin. 2024. Monte carlo tree search for recipe generation using gpt-2. The 14th International Conference on Computational Creativity

    2024

  46. [54]

    Quin Thames, Arjun Karpur, Wade Norris, Fangting Xia, Liviu Panait, Tobias Weyand, and Jack Sim. 2021. Nutrition5k: Towards automatic nutritional understanding of generic food. In Proceedings of the Conference on Computer Vision and Pattern Recognition, pages 8903--8911

    2021

  47. [55]

    Hugo Touvron, Louis Martin, Kevin Stone, Peter Albert, Amjad Almahairi, Yasmine Babaei, Nikolay Bashlykov, Soumya Batra, Prajjwal Bhargava, Shruti Bhosale, et al. 2023. Llama 2: Open foundation and fine-tuned chat models

    2023

  48. [56]

    Ramakrishna Vedantam, C Lawrence Zitnick, and Devi Parikh. 2015. Cider: Consensus-based image description evaluation. In Proceedings of the Conference on Computer Vision and Pattern Recognition, pages 4566--4575. IEEE

    2015

  49. [57]

    Muntasir Wahed, Xiaona Zhou, Tianjiao Yu, and Ismini Lourentzou. 2024. Fine-grained alignment for cross-modal recipe retrieval. In Proceedings of the Winter Conference on Applications of Computer Vision, pages 5584--5593. IEEE

    2024

  50. [58]

    Haiwen Wang, Le Zhou, Weinan Zhang, and Xinbing Wang. 2021. Literatureqa: A qestion answering corpus with graph knowledge on academic literature. In Proceedings of the 30th ACM International Conference on Information & Knowledge Management, pages 4623--4632

    2021

  51. [59]

    Shijie Wu, Ozan Irsoy, Steven Lu, Vadim Dabravolski, Mark Dredze, Sebastian Gehrmann, Prabhanjan Kambadur, David Rosenberg, and Gideon Mann. 2023. Bloomberggpt: A large language model for finance. arXiv preprint arXiv:2303.17564

    2023 arXiv

  52. [60]

    Zhentao Xu, Mark Jerome Cruz, Matthew Guevara, Tie Wang, Manasi Deshpande, Xiaofeng Wang, and Zheng Li. 2024 a . Retrieval-augmented generation with knowledge graphs for customer service question answering. In Proceedings of the 47th International ACM SIGIR Conference on Resea...

    2024

  53. [61]

    Ziwei Xu, Sanjay Jain, and Mohan Kankanhalli. 2024 b . Hallucination is inevitable: An innate limitation of large language models. arXiv preprint arXiv:2401.11817

    2024 arXiv

  54. [62]

    Dongyu Yao, Keling Yao, Junhong Zhou, and Yinghao Zhang. 2024. Caloraify: Calorie estimation with visual-text pairing and lora-driven visual language models. arXiv preprint arXiv:2412.09936

    2024 arXiv

  55. [63]

    Yuehao Yin, Huiyan Qi, Bin Zhu, Jingjing Chen, Yu-Gang Jiang, and Chong-Wah Ngo. 2023. Foodlmm: A versatile food assistant using large multi-modal model. arXiv preprint arXiv:2312.14991

    2023 arXiv

  56. [64]

    Zheyuan Zhang, Zehong Wang, Tianyi Ma, Varun Sameer Taneja, Sofia Nelson, Nhi Ha Lan Le, Keerthiram Murugesan, Mingxuan Ju, Nitesh V Chawla, Chuxu Zhang, et al. 2024. Mopi-hfrs: A multi-objective personalized health-aware food recommendation system with llm-enhanced interpreta...

    2024

  57. [65]

    online" 'onlinestring :=

    ENTRY address archivePrefix author booktitle chapter edition editor eid eprint eprinttype howpublished institution journal key month note number organization pages publisher school series title type volume year doi pubmed url lastchecked label extra.label sort.label short.list...

  58. [66]

    write newline

    " write newline "" before.all 'output.state := FUNCTION n.dashify 't := "" t empty not t #1 #1 substring "-" = t #1 #2 substring "--" = not "--" * t #2 global.max substring 't := t #1 #1 substring "-" = "-" * t #2 global.max substring 't := while if t #1 #1 substring * t #2 gl...

Pith tools

Reviewed August 7, 2026 · model on record in the stance chip above.