Open Software Development Research Articles

Towards Optimal NLP Solutions: Analyzing GPT and LLaMA-2 Models Across Model Scale, Dataset Size, and Task Diversity

This study carries out a comprehensive comparison of fine-tuned GPT models (GPT-2, GPT-3, GPT-3.5) and LLaMA-2 models (LLaMA-2 7B, LLaMA-2 13B, LLaMA-2 70B) in text classification, addressing dataset sizes, model scales, and task diversity. Since its inception in 2018, the GPT series has been pivotal in advancing NLP, with each iteration introducing substantial enhancements. Despite its progress, detailed analyses, especially against competitive open-source models like the LLaMA-2 series in text classification, remain scarce. The current study fills this gap by fine-tuning these models across varied datasets, focusing on enhancing task-specific performance in hate speech and offensive language detection, fake news classification, and sentiment analysis. The learning efficacy and efficiency of the GPT and LLaMA-2 models were evaluated, providing a nuanced guide to choosing optimal models for NLP tasks based on architectural benefits and adaptation efficiency with limited data and resources. In particular, even with datasets as small as 1,000 rows per class, the F1 scores for the GPT-3.5 and LLaMA-2 models exceeded 0.9, reaching 0.99 with complete datasets. Additionally, the LLaMA-2 13B and 70B models outperformed GPT-3, demonstrating their superior efficiency and effectiveness in text classification. Both the GPT and LLaMA-2 series showed commendable performance on all three tasks, underscoring their ability to handle a diversity of tasks. Based on the size, performance, and resources required for fine-tuning the model, this study identifies LLaMA-2 13B as the most optimal model for NLP tasks.

Research Paper on Social Media Application Using MERN Stack

Abstract: With rapid increase in the technology it is very essential to understand about emerging computing technology in the world. This research paper will highlight the popular MERN stack which includes four open source technologies i.e MERN ( MongoDB, Express, ReactJs, NodeJs). The demand for skilled MERN stack developers is skyrocketing. Companies across industries are recognizing the power and flexibility of this technology stack and are constantly seeking talented individuals who can bring their ideas to life using MERN. We will be discussing MERN Stack Application Social Media App

Strong Fractionation of Deuterium and Helium in Sub-Neptune Atmospheres along the Radius Valley

We simulate atmospheric fractionation in escaping planetary atmospheres using IsoFATE, a new open-source numerical model. We expand the parameter space studied previously to planets with tenuous atmospheres that exhibit the greatest helium and deuterium enhancement. We simulate the effects of extreme-ultraviolet-driven photoevaporation and core-powered mass loss on deuterium–hydrogen and helium–hydrogen fractionation of sub-Neptune atmospheres around G, K, and M stars. Our simulations predict prominent populations of deuterium- and helium-enhanced planets along the upper edge of the radius valley with mean equilibrium temperatures of ≈370 K and as low as 150 K across stellar types. We find that fractionation is mechanism dependent, so constraining He/H and D/H abundances in sub-Neptune atmospheres offers a unique strategy to investigate the origin of the radius valley around low-mass stars. Fractionation is also strongly dependent on retained atmospheric mass, offering a proxy for planetary surface pressure as well as a way to distinguish between desiccated enveloped terrestrials and water worlds. Deuterium-enhanced planets tend to be helium dominated and CH4 depleted, providing a promising strategy to observe HDO in the 3.7 μm window. We present a list of promising targets for observational follow-up.

CodeS: Towards Building Open-source Language Models for Text-to-SQL

Language models have shown promising performance on the task of translating natural language questions into SQL queries (Text-to-SQL). However, most of the state-of-the-art (SOTA) approaches rely on powerful yet closed-source large language models (LLMs), such as ChatGPT and GPT-4, which may have the limitations of unclear model architectures, data privacy risks, and expensive inference overheads. To address the limitations, we introduce CodeS, a series of pre-trained language models with parameters ranging from 1B to 15B, specifically designed for the text-to-SQL task. CodeS is a fully open-source language model, which achieves superior accuracy with much smaller parameter sizes. This paper studies the research challenges in building CodeS. To enhance the SQL generation abilities of CodeS, we adopt an incremental pre-training approach using a specifically curated SQL-centric corpus. Based on this, we address the challenges of schema linking and rapid domain adaptation through strategic prompt construction and a bi-directional data augmentation technique. We conduct comprehensive evaluations on multiple datasets, including the widely used Spider benchmark, the newly released BIRD benchmark, robustness-diagnostic benchmarks such as Spider-DK, Spider-Syn, Spider-Realistic, and Dr.Spider, as well as two real-world datasets created for financial and academic applications. The experimental results show that our CodeS achieves new SOTA accuracy and robustness on nearly all challenging text-to-SQL benchmarks.

An Internet of Things-Based Cluster System for Monitoring Lactating Sows’ Feed and Water Intake

Acquiring real-time feeding information for monitoring lactating sows and their feeding requirements is a challenging task. Real-time data represent an important input for numerous tasks, such as disease monitoring, nutritional regulation, and feeding modeling. However, concurrently monitoring large numbers of sows and processing the real-time information for modeling is challenging using existing platforms. In this paper, we describe the design and development of a system that monitors and processes sows’ feed and water consumption in real time. The system was custom-developed using open-source networking technologies. The system consists of three components: an electronic sow feeder connected to a central controller via a CAN network, an MQTT service cluster, and a data processing program. The MQTT service cluster uses Netty to develop a single service node, and it uses Zookeeper and Redis to complete node registration, discovery, and scheduling. The data processing program is based on Spark and Flink. We conducted comparative testing of three common codecs (Java Serializer, Marshalling, and Protostuff) to further speed up data transmission. The results of the experiment show that, with three service nodes, the system can concurrently monitor up to 20,000 sows. Moreover, the system achieves optimal performance when monitoring 10,000 sows at the same time, with a TPS of 6399 pcs/s and an RT of 643 ms.

Optimal timing of a colonoscopy screening schedule depends on adenoma detection, adenoma risk, adherence to screening and the screening objective: A microsimulation study.

Colonoscopy-based screening provides protection against colorectal cancer (CRC), but the optimal starting age and time intervals of screening colonoscopies are unknown. We aimed to determine an optimal screening schedule for the US population and its dependencies on the objective of screening (life years gained or incidence, mortality, or cost reduction) and the setting in which screening is performed. We used our established open-source microsimulation model CMOST to calculate optimized colonoscopy schedules with one, two, three or four screening colonoscopies between 20 and 90 years of age. A single screening colonoscopy was most effective in reducing life years lost from CRC when performed at 55 years of age. Two, three and four screening colonoscopy schedules saved a maximum number of life years when performed between 49-64 years; 44-69 years; and 40-72 years; respectively. However, for maximum incidence and mortality reduction, screening colonoscopies needed to be scheduled 4-8 years later in life. The optimum was also influenced by adenoma detection efficiency with lower values for these parameters favoring a later starting age of screening. Low adherence to screening consistently favored a later start and an earlier end of screening. In a personalized approach, optimal screening would start earlier for high-risk patients and later for low-risk individuals. In conclusion, our microsimulation-based approach supports colonoscopy screening schedule between 45 and 75 years of age but the precise timing depends on the objective of screening, as well as assumptions regarding individual CRC risk, efficiency of adenoma detection during colonoscopy and adherence to screening.

VISIR-2: ship weather routing in Python

Abstract. Ship weather routing, which involves suggesting low-emission routes, holds potential for contributing to the decarbonisation of maritime transport. However, including because of a lack of readily deployable open-source and open-language computational models, its quantitative impact has been explored only to a limited extent. As a response, the graph-search VISIR (discoVerIng Safe and effIcient Routes) model has been refactored in Python, incorporating novel features. For motor vessels, the angle of attack of waves has been considered, while for sailboats the combined effects of wind and sea currents are now accounted for. The velocity composition with currents has been refined, now encompassing leeway as well. Provided that the performance curve is available, no restrictions are imposed on the vessel type. A cartographic projection has been introduced. The graph edges are quickly screened for coast intersection via a K-dimensional tree. A least-CO2 algorithm in the presence of dynamic graph edge weights has been implemented and validated, proving a quasi-linear computational performance. The software suite's modularity has been significantly improved, alongside a thorough validation against various benchmarks. For the visualisation of the dynamic environmental fields along the route, isochrone-bounded sectors have been introduced. The resulting VISIR-2 model has been employed in numerical experiments within the Mediterranean Sea for the entirety of 2022, utilising meteo-oceanographic analysis fields. For a 125 m long ferry, the percentage saving of overall CO2 expenditure follows a bi-exponential distribution. Routes with a carbon dioxide saving of at least 2 % with respect to the least-distance route were found for prevailing beam or head seas. Two-digit savings, up to 49 %, were possible for about 10 d in a year. In the case of an 11 m sailboat, time savings increased with the extent of path elongation, particularly during upwind sailing. The sailboat's routes were made approximately 2.4 % faster due to optimisation, with the potential for an additional 0.8 % in savings by factoring in currents. VISIR-2 serves as an integrative model, uniting expertise from meteorology, oceanography, ocean engineering, and computer science, to evaluate the influence of ship routing on decarbonisation efforts within the shipping industry.

Building an open-source collaborative platform for migration research: A metadata modeling approach using XML

Public access to well-organized migration data repositories is rare. While open migration data are available, most are dispersed, challenging their organization. This work uses a metadata modeling approach to organize migration-related data and research components hosted in an open-source, collaborative research environment. The proposed platform integrates data, computational models, application libraries, and research projects in a unified environment for research. An overview of the metadata modeling process and the XML technology used to integrate the platform components was provided, along with descriptions of the platform’s essential features, such as standardized exchange of information for data organization and sharing and collaborative authoring. The design of a prototype collaborative, open-source migration data and modeling research platform using sample data and software components was built and hosted on GitHub for testing and refinement purposes. It allows the authors to collaborate, improve, and assess the platform’s functionalities and capabilities. Our platform, COSMOS (Collaborative Open-Source Modeling System), adopted the creation and use of open data formats for alignment with FAIR (Findable, Accessible, Interoperable, and Reusable) data principles to ensure that data at the prototype level is consistent with and can be used by the international migration research community in the future. By providing a standardized framework for data organization, data sharing, and collaboration, the platform can help accelerate scientific discovery and advance our understanding of complex social and cultural phenomena such as immigration.

DEVELOPMENT OF AN OPEN EDUCATION RESOURCES (OER) SYSTEM: A COMPARATIVE ANALYSIS AND IMPLEMENTATION APPROACH

Several institutions are collaborating on the development of a new web-based Open Education Resources (OER) system designed exclusively for non-commercial educational purposes. This initiative is underpinned by meticulous research aimed at constructing an OER system that optimizes user experiences across diverse user profiles. A significant emphasis is placed on utilizing open-source tools, frameworks, and technologies. The project includes a comparative analysis of the top five open-source Learning Management Systems (LMS), providing critical insights to inform the development process. The primary objective is to create a web-based system that facilitates the sharing of educational resources for non-commercial users, leveraging information and communication technologies. The project is structured around two key teams: a research team and a development team. This comprehensive approach is intended to establish a robust, user-centric OER system, informed by insights from existing platforms and the latest advancements in open education resource development. KEYWORDS: Open Education Resources (OER), Web-based System (WBS), Learning Management Systems (LMS)

Containerized Microservices for Mobile Applications Deployed on Cloud Systems

This study explores the transformative role of containerized microservices in the sphere of mobile application development, especially within public cloud ecosystems. It focuses on how technologies such as Docker and Kubernetes contribute to improving deployment, scalability, and overall management of mobile applications, with an emphasis on containerizing backend services. We analyze their efficiency in streamlining deployment processes, focusing on how they improve the application’s performance and reliability. Additionally, we examine various alternative deployment strategies, such as blue-green, rolling, and canary releases, to emphasize their effectiveness in minimizing risks and facilitating smooth transitions in dynamic cloud environments. The study takes a comprehensive approach to achieve this goal, which includes a systematic review of existing literature, a thorough examination of relevant use cases, and an assessment of open-source technologies. Our findings reveal not only the practical benefits of these strategies but also their strategic application, offering important insights for software engineers and decision-makers. This study emphasizes the significance of integrating and optimizing containerized microservices in mobile app development to achieve more efficient, scalable, and manageable application lifecycles on cloud-based platforms.

Ecology and conservation researchers should adopt open source technologies

Ecology and conservation researchers should adopt open source technologies

Automated Data Quality Control System in Health and Demographic Surveillance System

The automated data quality management system serves as a comprehensive solution developed to enhance the precision, dependability, and uniformity of data within data-driven organizations. Such systems play an important role in eliminating the shortcomings associated with manual data quality management, which is prevalent in health and demographic surveillance systems (HDSS). The ongoing difficulty of ensuring data quality through manual processing hinders the HDSS's capacity to optimize data quality effectively. To address this challenge, our study adopted design science methodologies to provide guidelines for the design and implementation of the automated data quality control system. The open source technologies (Pentaho data integration, R Studio, SQL, Windows task scheduler) were used to facilitate the automation and validation of the incoming and database resident data. The quality of data has vastly improved since the implementation of the proposed system. The findings suggest that the automated data quality control system exhibited superior performance compared to the manual methods, thereby minimizing errors and time-wasting efforts.

TinyML4D: Scaling Embedded Machine Learning Education in the Developing World

Embedded machine learning (ML) on low-power devices, also known as "TinyML," enables intelligent applications on accessible hardware and fosters collaboration across disciplines to solve real-world problems. Its interdisciplinary and practical nature makes embedded ML education appealing, but barriers remain that limit its accessibility, especially in developing countries. Challenges include limited open-source software, courseware, models, and datasets that can be used with globally accessible heterogeneous hardware. Our vision is that with concerted effort and partnerships between industry and academia, we can overcome such challenges and enable embedded ML education to empower developers and researchers worldwide to build locally relevant AI solutions on low-cost hardware, increasing diversity and sustainability in the field. Towards this aim, we document efforts made by the TinyML4D community to scale embedded ML education globally through open-source curricula and introductory workshops co-created by international educators. We conclude with calls to action to further develop modular and inclusive resources and transform embedded ML into a truly global gateway to embedded AI skills development.

Using image segmentation models to analyse high-resolution earth observation data: new tools to monitor disease risks in changing environments.

In the near future, the incidence of mosquito-borne diseases may expand to new sites due to changes in temperature and rainfall patterns caused by climate change. Therefore, there is a need to use recent technological advances to improve vector surveillance methodologies. Unoccupied Aerial Vehicles (UAVs), often called drones, have been used to collect high-resolution imagery to map detailed information on mosquito habitats and direct control measures to specific areas. Supervised classification approaches have been largely used to automatically detect vector habitats. However, manual data labelling for model training limits their use for rapid responses. Open-source foundation models such as the Meta AI Segment Anything Model (SAM) can facilitate the manual digitalization of high-resolution images. This pre-trained model can assist in extracting features of interest in a diverse range of images. Here, we evaluated the performance of SAM through the Samgeo package, a Python-based wrapper for geospatial data, as it has not been applied to analyse remote sensing images for epidemiological studies. We tested the identification of two land cover classes of interest: water bodies and human settlements, using different UAV acquired imagery across five malaria-endemic areas in Africa, South America, and Southeast Asia. We employed manually placed point prompts and text prompts associated with specific classes of interest to guide the image segmentation and assessed the performance in the different geographic contexts. An average Dice coefficient value of 0.67 was obtained for buildings segmentation and 0.73 for water bodies using point prompts. Regarding the use of text prompts, the highest Dice coefficient value reached 0.72 for buildings and 0.70 for water bodies. Nevertheless, the performance was closely dependent on each object, landscape characteristics and selected words, resulting in varying performance. Recent models such as SAM can potentially assist manual digitalization of imagery by vector control programs, quickly identifying key features when surveying an area of interest. However, accurate segmentation still requires user-provided manual prompts and corrections to obtain precise segmentation. Further evaluations are necessary, especially for applications in rural areas.

A novel full-resolution convolutional neural network for urban-fringe-rural identification: A case study of urban agglomeration region

Monitoring urbanization processes is important because they are often accompanied by intensive landscape pattern transitions and pluralistic socioeconomic changes. To effectively monitor urban expansion and support regional planning, it is essential to develop a fast, accurate and universal urban–rural classification model, especially identifying the dynamic spatial patterns of urban, urban–rural fringe and rural areas. Although deep learning can effectively detect land cover changes, its applications in urban–rural identification have received little attention due to a lack of high-quality training datasets. In this study, we develop a novel transferable full-resolution convolutional neural network (FR-Net) to identify urban-fringe-rural areas. A large-scale training dataset was constructed using field surveys and aerial photography, and a data cube was stacked by multiple typical socio-natural indicators. We took the Beijing-Tianjin-Hebei (BTH) urban agglomeration region in China as a case study and identified spatiotemporal evolutions of urban-fringe-rural areas from 2000 to 2020. The results indicated that over the past two decades, the urban–rural fringe expanded outward with urban areas, and both areas gradually increased, with an inverted U-shaped growth rate. Accurate identification of these fringes can benefit regional urban–rural planning and social governance. Based on the identification results, complex socio-ecological impacts of urbanization could be further explored. Testing demonstrated that the developed FR-Net model has high accuracy and robustness. Our developed open-source FR-Net model exhibits transferability and can be applied to multi-scale urbanized areas.

Accelerated MRI reconstructions via variational network and feature domain learning

We introduce three architecture modifications to enhance the performance of the end-to-end (E2E) variational network (VarNet) for undersampled MRI reconstructions. We first implemented the Feature VarNet, which propagates information throughout the cascades of the network in an N-channel feature-space instead of a 2-channel feature-space. Then, we add an attention layer that utilizes the spatial locations of Cartesian undersampling artifacts to further improve performance. Lastly, we combined the Feature and E2E VarNets into the Feature-Image (FI) VarNet, to facilitate cross-domain learning and boost accuracy. Reconstructions were evaluated on the fastMRI dataset using standard metrics and clinical scoring by three neuroradiologists. Feature and FI VarNets outperformed the E2E VarNet for 4×\\documentclass[12pt]{minimal} \\usepackage{amsmath} \\usepackage{wasysym} \\usepackage{amsfonts} \\usepackage{amssymb} \\usepackage{amsbsy} \\usepackage{mathrsfs} \\usepackage{upgreek} \\setlength{\\oddsidemargin}{-69pt} \\begin{document}$$\ imes$$\\end{document}, 5×\\documentclass[12pt]{minimal} \\usepackage{amsmath} \\usepackage{wasysym} \\usepackage{amsfonts} \\usepackage{amssymb} \\usepackage{amsbsy} \\usepackage{mathrsfs} \\usepackage{upgreek} \\setlength{\\oddsidemargin}{-69pt} \\begin{document}$$\ imes$$\\end{document} and 8×\\documentclass[12pt]{minimal} \\usepackage{amsmath} \\usepackage{wasysym} \\usepackage{amsfonts} \\usepackage{amssymb} \\usepackage{amsbsy} \\usepackage{mathrsfs} \\usepackage{upgreek} \\setlength{\\oddsidemargin}{-69pt} \\begin{document}$$\ imes$$\\end{document} Cartesian undersampling in all studied metrics. FI VarNet secured second place in the public fastMRI leaderboard for 4×\\documentclass[12pt]{minimal} \\usepackage{amsmath} \\usepackage{wasysym} \\usepackage{amsfonts} \\usepackage{amssymb} \\usepackage{amsbsy} \\usepackage{mathrsfs} \\usepackage{upgreek} \\setlength{\\oddsidemargin}{-69pt} \\begin{document}$$\ imes$$\\end{document} Cartesian undersampling, outperforming all open-source models in the leaderboard. Radiologists rated FI VarNet brain reconstructions with higher quality and sharpness than the E2E VarNet reconstructions. FI VarNet excelled in preserving anatomical details, including blood vessels, whereas E2E VarNet discarded or blurred them in some cases. The proposed FI VarNet enhances the reconstruction quality of undersampled MRI and could enable clinically acceptable reconstructions at higher acceleration factors than currently possible.

Development of an IoT-Based Alcohol Detector Using Blynk and NodeMCU

Although the consumption of alcoholic beverages has been a part of human society for a very long time, it is also a big contributor to accidents and health problems. Driving under the influence of alcohol is a leading contributor to death and injury on the roads across the globe. Individuals continue to operate motor vehicles while under the influence of alcohol, despite the existence of stringent regulations and awareness efforts, which poses a risk to both themselves and others who are on the road. With the intention of preventing drunk driving and promoting road safety, an Internet of Things (IoT)-based alcohol detector has been developed utilising Blynk and NodeMCU. This detector was developed to address this issue. The amount of alcohol that is present in a person's blood, urine, or breath can be determined with the help of a device known as an alcohol detector. Using a gas sensor, the device is able to determine the amount of alcohol that is present in the breath of the user. A connection has been made between the sensor and the NodeMCU, which is tied to the Blynk application. The NodeMCU receives the information that is gathered by the gas sensor that is incorporated within the device. This sensor monitors the amount of alcohol that is breathed by the user. Through the use of a smartphone application, Blynk is an Internet of Things platform that gives consumers the ability to control and monitor connected devices. This alcohol detector that is based on the Internet of Things is extremely portable, which is one of its most significant features. A cost-effective solution has been developed through the utilisation of open-source technologies such as Blynk and NodeMCU. This has enabled the device to maintain its accuracy and quality without compromising its functionality.

Artificial intelligence-based tools with automated segmentation and measurement on CT images to assist accurate and fast diagnosis in acute pancreatitis.

To develop an artificial intelligence (AI) tool with automated pancreas segmentation and measurement of pancreatic morphological information on CT images to assist improved and faster diagnosis in acute pancreatitis. This study retrospectively contained 1124 patients suspected for AP and received non-contrast and enhanced abdominal CT examination between September 2013 and September 2022. Patients were divided into training (N = 688), validation (N = 145), testing dataset [N = 291; N = 104 for normal pancreas, N = 98 for AP, N = 89 for AP complicated with PDAC (AP&PDAC)]. A model based on convolutional neural network (MSAnet) was developed. The pancreas segmentation and measurement were performed via eight open-source models and MSAnet based tools, and the efficacy was evaluated using dice similarity coefficient (DSC) and intersection over union (IoU). The DSC and IoU for patients with different ages were also compared. The outline of tumour and oedema in the AP and were segmented by clustering. The diagnostic efficacy for radiologists with or without the assistance of MSAnet tool in AP and AP&PDAC was evaluated using receiver operation curve and confusion matrix. Among all models, MSAnet based tool showed best performance on the training and validation dataset, and had high efficacy on testing dataset. The performance was age-affected. With assistance of the AI tool, the diagnosis time was significantly shortened by 26.8% and 32.7% for junior and senior radiologists, respectively. The area under curve (AUC) in diagnosis of AP was improved from 0.91 to 0.96 for junior radiologist and 0.98 to 0.99 for senior radiologist. In AP&PDAC diagnosis, AUC was increased from 0.85 to 0.92 for junior and 0.97 to 0.99 for senior. MSAnet based tools showed good pancreas segmentation and measurement performance, which help radiologists improve diagnosis efficacy and workflow in both AP and AP with PDAC conditions. This study developed an AI tool with automated pancreas segmentation and measurement and provided evidence for AI tool assistance in improving the workflow and accuracy of AP diagnosis.

TeenyTinyLlama: Open-source tiny language models trained in Brazilian Portuguese

Large language models (LLMs) have significantly advanced natural language processing, but their progress has yet to be equal across languages. While most LLMs are trained in high-resource languages like English, multilingual models generally underperform monolingual ones. Additionally, aspects of their multilingual foundation sometimes restrict the byproducts they produce, like computational demands and licensing regimes. In this study, we document the development of open-foundation models tailored for use in low-resource settings, their limitations, and their benefits. This is the TeenyTinyLlama pair: two compact models for Brazilian Portuguese text generation. We release them under the permissive Apache 2.0 license on GitHub and Hugging Face for community use and further development.

The application of natural language processing for the extraction of mechanistic information in toxicology.

To study the ways in which compounds can induce adverse effects, toxicologists have been constructing Adverse Outcome Pathways (AOPs). An AOP can be considered as a pragmatic tool to capture and visualize mechanisms underlying different types of toxicity inflicted by any kind of stressor, and describes the interactions between key entities that lead to the adverse outcome on multiple biological levels of organization. The construction or optimization of an AOP is a labor intensive process, which currently depends on the manual search, collection, reviewing and synthesis of available scientific literature. This process could however be largely facilitated using Natural Language Processing (NLP) to extract information contained in scientific literature in a systematic, objective, and rapid manner that would lead to greater accuracy and reproducibility. This would support researchers to invest their expertise in the substantive assessment of the AOPs by replacing the time spent on evidence gathering by a critical review of the data extracted by NLP. As case examples, we selected two frequent adversities observed in the liver: namely, cholestasis and steatosis denoting accumulation of bile and lipid, respectively. We used deep learning language models to recognize entities of interest in text and establish causal relationships between them. We demonstrate how an NLP pipeline combining Named Entity Recognition and a simple rules-based relationship extraction model helps screen compounds related to liver adversities in the literature, but also extract mechanistic information for how such adversities develop, from the molecular to the organismal level. Finally, we provide some perspectives opened by the recent progress in Large Language Models and how these could be used in the future. We propose this work brings two main contributions: 1) a proof-of-concept that NLP can support the extraction of information from text for modern toxicology and 2) a template open-source model for recognition of toxicological entities and extraction of their relationships. All resources are openly accessible via GitHub (https://github.com/ontox-project/en-tox).