arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2410.22296cs.LGq-bio.QM
收藏

通才与专才:在高约束生物物理序列优化任务上评估大语言模型

Generalists vs. Specialists: Evaluating LLMs on Highly-Constrained Biophysical Sequence Optimization Tasks

Center for Data Science, New York University, New York City, U.S.A.(数据科学中心,纽约大学,纽约市,美国)

Angelica Chen, Samuel D. Stanton, Frances Ding, Robert G. Alberstein, Andrew M. Watkins, Richard Bonneau, Vladimir Gligorijević, Kyunghyun Cho, Nathan C. Frey

AI总结:

该研究针对大语言模型与专用求解器在生物物理序列优化上的对比难题,提出合成测试套件Ehrlich函数及双层优化方法LLOME,发现大模型偶有优势但专用求解器仍具竞争力且开销更低。

摘要:

尽管大语言模型(LLMs)在生物分子优化问题中展现出应用前景,但它们计算成本高昂,且难以满足精确约束。另一方面,LaMBO-2等专用求解器兼具效率与细粒度控制能力,但需要更多领域专业知识。由于实验室验证成本高昂且合成基准不足,对这两类方法进行比较颇具挑战。我们通过引入Ehrlich函数(一套捕捉生物物理序列优化问题几何结构的合成测试套件)解决了这一问题。仅使用提示工程时,现成LLMs难以优化Ehrlich函数。对此,我们提出LLOME(基于边际期望的语言模型优化)——一种用于在线黑箱优化的双层优化流程。结合一种新型偏好学习损失后,我们发现LLOME不仅能学会求解部分Ehrlich函数,在中等难度的Ehrlich变体上甚至能达到与LaMBO-2相当或更优的性能。不过,LLMs也存在一定的似然-奖励校准偏差,且在无明确奖励时表现不佳。我们的结果表明,LLMs偶尔能带来显著收益,但专用求解器仍具竞争力且开销更低。

英文摘要:

Although large language models (LLMs) have shown promise in biomolecule optimization problems, they incur heavy computational costs and struggle to satisfy precise constraints. On the other hand, specialized solvers like LaMBO-2 offer efficiency and fine-grained control but require more domain expertise. Comparing these approaches is challenging due to expensive laboratory validation and inadequate synthetic benchmarks. We address this by introducing Ehrlich functions, a synthetic test suite that captures the geometric structure of biophysical sequence optimization problems. With prompting alone, off-the-shelf LLMs struggle to optimize Ehrlich functions. In response, we propose LLOME (Language Model Optimization with Margin Expectation), a bilevel optimization routine for online black-box optimization. When combined with a novel preference learning loss, we find LLOME can not only learn to solve some Ehrlich functions, but can even perform as well as or better than LaMBO-2 on moderately difficult Ehrlich variants. However, LLMs also exhibit some likelihood-reward miscalibration and struggle without explicit rewards. Our results indicate LLMs can occasionally provide significant benefits, but specialized solvers are still competitive and incur less overhead.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑