{"ID":2822722,"CreatedAt":"2026-06-01T04:54:23.091178241Z","UpdatedAt":"2026-06-01T04:54:23.091178241Z","DeletedAt":null,"paper_url":"https://arxiv.org/abs/2601.02531","arxiv_id":"2601.02531","title":"Losses that Cook: Topological Optimal Transport for Structured Recipe Generation","abstract":"Cooking recipes are complex procedures that require not only a fluent and factual text, but also accurate timing, temperature, and procedural coherence, as well as the correct composition of ingredients. Standard training procedures are primarily based on cross-entropy and focus solely on fluency. Building on RECIPE-NLG, we investigate the use of several composite objectives and present a new topological loss that represents ingredient lists as point clouds in embedding space, minimizing the divergence between predicted and gold ingredients. Using both standard NLG metrics and recipe-specific metrics, we find that our loss significantly improves ingredient- and action-level metrics. Meanwhile, the Dice loss excels in time/temperature precision, and the mixed loss yields competitive trade-offs with synergistic gains in quantity and time. A human preference analysis supports our finding, showing our model is preferred in 62% of the cases.","short_abstract":"Cooking recipes are complex procedures that require not only a fluent and factual text, but also accurate timing, temperature, and procedural coherence, as well as the correct composition of ingredients. Standard training procedures are primarily based on cross-entropy and focus solely on fluency. Building on RECIPE-NL...","url_abs":"https://arxiv.org/abs/2601.02531","url_pdf":"https://arxiv.org/pdf/2601.02531v2","authors":"[\"Mattia Ottoborgo\",\"Daniele Rege Cambrin\",\"Paolo Garza\"]","published":"2026-01-05T20:09:41Z","proceeding":"cs.CL","tasks":"[\"cs.CL\",\"cs.AI\"]","methods":"[]","has_code":false}