从具有有限注释的科学文献中提取物质性质测量.
Jessica Kong1, Gihan Panapitiya1, Emily Saldanha1
1Pacific Northwest National Laboratory, Richland, Washington 99354 , United States.
Journal of chemical information and modeling
|May 13, 2025
概括
提取物质属性数据对科学至关重要. 像GPT-4o这样的大型语言模型 (LLM) 在减少命名实体识别 (NER) 的数据标签和提高准确性方面显示出希望.
科学领域:
- 化学 化学 化学
- 材料科学 材料科学 材料科学
- 数据科学数据科学数据科学
背景情况:
- 从科学文献中提取物质属性数据对于数据驱动的研究至关重要.
- 使用监督命名实体识别 (NER) 的传统方法需要广泛的数据注释,这构成了重大障碍.
- 开发高效的专门数据提取方法是一个关键的挑战.
研究的目的:
- 为了比较传统的监督NER与少量学习和大语言模型 (LLM) 方法进行物质属性提取.
- 评估LLMs在减少对大型标记训练数据集的需求方面的有效性.
- 评估GPT-4o在直接财产提取和数据增强方面的性能.
主要方法:
- 对受监督的NER,少量学习架构和基于LLM的方法进行比较分析.
- 使用GPT-4o直接提取物质属性,仅有有限的例子.
- 采用LLM进行数据增强,以增强监督学习模型.
- 进行错误和数据质量评估.
主要成果:
- 性能最好的LLM (GPT-4o) 使用有限的示例有效提取材料属性.
- 通过数据增强,LLM,特别是GPT-4o,可以显著增强监督学习.
- 基于LLM的方法减轻了NER任务广泛数据标签的障碍.
- 错误和数据质量分析为提取性能因素提供了洞察力.
结论:
- 在物质财产提取方面,LLM为传统的监督NER提供了强大的替代方案.
- 在直接提取和增强监督模型方面,GPT-4o表现出卓越的性能.
- 这些发现为材料科学和化学中更有效和更可扩展的数据提取铺平了道路.
相关概念视频
03:02Measurement: Derived Units
The International System of Units or SI system, by international agreement, has fixed measurement units for seven fundamental properties: length, mass, time, temperature, electric current, amount of substance, and luminosity. These are called the SI base units.
02:46Uncertainty in Measurement: Reading Instruments
Counting is the type of measurement that is free from uncertainty, provided the number of objects being counted does not change during the process. Such measurements result in exact numbers. By counting the eggs in a carton, for instance, one can determine exactly how many eggs are there in the carton. Similarly, the numbers of defined quantities are also exact. For example, 1 foot is exactly 12 inches, 1 inch is exactly 2.54 centimeters, and 1 gram is exactly 0.001 kilograms. Quantities...
01:19Mass Spectrometry: Overview
Mass spectrometry is an analytical technique used to determine the molecular mass and molecular formula of a compound. The basic principle of mass spectrometry is to generate ions from the analyte molecule and measure these ion abundances against their molecular mass. One common type of ionization, known as electron ionization or EI, bombards the analyte molecules in the gas phase with high-energy electron beams. The electron beams displace an electron from the molecule and leave behind a...
01:09Empirical Method to Interpret Standard Deviation
The empirical rule, also known as the three-sigma rule, allows a statistician to interpret the standard deviation in a normally distributed dataset. The rule states that 68% of the data lies within one standard deviation from the mean, 95% lies within two standard deviations from the mean, and 99.7% lies within three standard deviations from the mean. Additionally, this rule is also called the 68-95-99.7 rule.
This rule is used widely in statistics to calculate the proportion of data values...
This rule is used widely in statistics to calculate the proportion of data values...
01:08Titrimetric Methods: Types and Commonly Used Strategies
In chemistry, titrimetric methods are broadly classified into three types: volumetric, gravimetric, and coulometric. Volumetric titrations involve measuring the volume of a titrant of known concentration that is required to react completely with an analyte. In gravimetric titrations, the standard solution reacts with the analyte to form an insoluble precipitate, which is filtered, dried, and weighed. In coulometric titrations, current is applied to an electrochemical reaction until the reaction...
01:05Estimation of the Physical Quantities
On many occasions, physicists, other scientists, and engineers need to make estimates of a particular quantity. These are sometimes referred to as guesstimates, order-of-magnitude approximations, back-of-the-envelope calculations, or Fermi calculations. The physicist Enrico Fermi was famous for his ability to estimate various kinds of data with surprising precision. Estimating does not mean guessing a number or a formula at random. Instead, estimation means using prior experience and sound...