LM4OPT: Unveiling the potential of Large Language Models in formulating mathematical optimization problems
Why this work is in the frame
A frame that forgets how it found something cannot be audited. These are the routes that admitted this work.
Bibliographic record
Abstract
In the fast-paced domain of natural language processing, converting linguistic descriptions into mathematical optimization problems is a complex task, requiring profound comprehension and processing skills from Large Language Models (LLMs). In this study, various LLMs were evaluated, including GPT-3.5, GPT-4, and smaller variants with seven billion parameters: Llama-2, Falcon, Mistral, and Zephyr. This research investigated their performance in both zero-shot and one-shot settings for this task, revealing that GPT-4 outperformed others, particularly in the one-shot scenario. A core contribution of this study is the development of LM4OPT, a progressive fine-tuning framework specifically designed for smaller LLMs. This framework leverages noisy embeddings and specialized datasets to enhance the performance of the models. Regardless of the inherent limitations of smaller models in processing complex and lengthy input contexts, our experimental results indicate a significant reduction in the performance disparity between smaller and larger models when the former are fine-tuned using LM4OPT. Our empirical study, utilizing the NL4Opt dataset, unveils that GPT-4 surpasses the baseline performance established by previous research, achieving an accuracy of 63.30%, solely based on the problem description in natural language, and without relying on any additional named entity information. GPT-3.5 follows closely, both outperforming the progressively fine-tuned smaller models.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
Full frame distilled prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. Learned from the 10,348 direct Codex labels and 10,348 direct Gemma labels. Candidate is the union of thresholded teacher heads; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels or direct frontier model labels.
Codex and Gemma teacher scores by category
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.003 | 0.000 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.000 | 0.000 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.001 | 0.004 |
| Open science | 0.000 | 0.000 |
| Research integrity | 0.000 | 0.000 |
| Insufficient payload (model declined to judge) | 0.000 | 0.000 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from it