LLMs Accelerate Annotation for Medical Information Extraction
Authors:
Akshay Goel,
Almog Gueta,
Omry Gilon,
Chang Liu,
Sofia Erell,
Lan Huong Nguyen,
Xiaohong Hao,
Bolous Jaber,
Shashir Reddy,
Rupesh Kartha,
Jean Steiner,
Itay Laish,
Amir Feder
Abstract:
The unstructured nature of clinical notes within electronic health records often conceals vital patient-related information, making it challenging to access or interpret. To uncover this hidden information, specialized Natural Language Processing (NLP) models are required. However, training these models necessitates large amounts of labeled data, a process that is both time-consuming and costly wh…
▽ More
The unstructured nature of clinical notes within electronic health records often conceals vital patient-related information, making it challenging to access or interpret. To uncover this hidden information, specialized Natural Language Processing (NLP) models are required. However, training these models necessitates large amounts of labeled data, a process that is both time-consuming and costly when relying solely on human experts for annotation. In this paper, we propose an approach that combines Large Language Models (LLMs) with human expertise to create an efficient method for generating ground truth labels for medical text annotation. By utilizing LLMs in conjunction with human annotators, we significantly reduce the human annotation burden, enabling the rapid creation of labeled datasets. We rigorously evaluate our method on a medical information extraction task, demonstrating that our approach not only substantially cuts down on human intervention but also maintains high accuracy. The results highlight the potential of using LLMs to improve the utilization of unstructured clinical data, allowing for the swift deployment of tailored NLP solutions in healthcare.
△ Less
Submitted 4 December, 2023;
originally announced December 2023.
Study of complex properties of binary system of ethanol-methanol at extreme concentrations
Authors:
K Nilavarasi,
Thejus R Kartha,
V Madhurima
Abstract:
At low concentrations of methanol in ethanol-methanol binary system, the molecular interactions are seen to be uniquely complex. It is observed that the ethanol aggregates are not strictly hydrogen-bonded complexes; dispersion forces also play a dominant role in the self- association of ethanol molecules. On the addition of small amount of methanol to ethanol, the dipolar association of ethanol is…
▽ More
At low concentrations of methanol in ethanol-methanol binary system, the molecular interactions are seen to be uniquely complex. It is observed that the ethanol aggregates are not strictly hydrogen-bonded complexes; dispersion forces also play a dominant role in the self- association of ethanol molecules. On the addition of small amount of methanol to ethanol, the dipolar association of ethanol is destroyed. The repulsive forces between the two moieties dominate the behavior of the binary system at lower concentration of methanol. At higher concentration of methanol (> 30%), the strength and extent (number) of formation of hydrogen bonds between ethanol and methanol increases. The geometry of molecular structure at high concentration favors the fitting of component molecules with each other. Intermolecular interactions in the ethanol-methanol binary system over the entire concentration range were investigated in detail using broadband dielectric spectroscopy, FTIR, surface tension and refractive index studies. Molecular Dynamics simulations show that the hydrogen bond density is a direct function of the number of methanol molecules present, as the ethanol aggregates are not strictly hydrogen-bond constructed which is in agreement with the experimental results.
△ Less
Submitted 18 November, 2016;
originally announced November 2016.
