John Snow Labs’ Post

Because there is still an 80%(!) failure rate of clinical trials, and additionally similar staggering statistics from an enrolment, accrual rate, and retention perspective, biopharma is still focused on becoming more efficient and effective. The culprit? Unstructured and isolated data. Manual efforts to carefully sift through scientific research, epidemiology, patient demographics, real-world evidence, regulatory guidelines, ... compound the issue, often leading to repeated one-off efforts. The good thing is that there is a way to overcome this cumbersome process, and the combination of Natural Language Processing and Knowledge Graphs gives us that opportunity. Imagine a scenario where you can turn critical information hiding in a free text document into searchable information, and then combine this with other information in a compiled view. That is the reality the combination of these technologies offers. This creates possibilities for patient recruitment, protocol development, virtual cohorts - as explained by colleague Martin Robbins in the video - and other use cases in the clinical domain. Want to learn how ONTOFORCE can create a fit-for-purpose solution for your Clinical teams? Click the link in my bio for a free demo, or have a look at the webinar that is linked in the comments. 👇👇 #NLP #textmining #knowledgegraphs #lifesciences #ontology

Confluence of Epidemiological Modeling and Deep Learning!! {Epidemiological Deep Neural Networks} Excited to share my latest project where I implemented a paper on EPI-DNNs, published in "Computers in Biology and Medicine" (IF = 7.7), within 20 days! A stage to present this was provided as a part of the Mathematical Modeling course lectured by Dr. Anushaya Mohapatra. Inspired by a brief after-class discussion with Dr. Abhishek Pandey from Yale University during his "Introduction to Epidemiological Modeling" lecture series, I explored the potential of ML in Compartmental Modeling. I improvised and adapted the original EPI-DNNs paper to Indian COVID-19 data using a completely novel and different mathematical model. My neural network predicts key epidemiological parameters over 810 days, integrating deep learning with traditional methods for improved accuracy and computing efficiency. Then the predicted time-varying parameters are used to solve ODEs using the Classical Runge Kutta Method and the results are verified, shown through different plots. These predicted parameters can further be used to forecast the growth, various types of interventions required, and many more! 🔗 Explore the Project Check out my project on GitHub: https://lnkd.in/gJ-HKt7t The Research Paper: (https://lnkd.in/gerPZPTk) #Epidemiology #DeepLearning #MachineLearning #DataScience #PublicHealth #AI #COVID19

The publication "Natural language processing analysis of the theories of people with multiple sclerosis about causes of their disease" is now available in Nature Portfolio's #communicationsmedicine. The findings emphasize the importance of communication between healthcare professionals and persons with MS about the pathogenesis of MS, the scientific evidence base and mental health. ➡ Read the full paper here: https://lnkd.in/ePNGwY6t Christina Haag, Nina Steinemann, Benjamin Victor Ineichen, Mina Stanikić, Paola Daniore, Sabin Ammann, Pasquale Calabrese, Jürg Kesselring, Jens Kuhle, Milo Puhan, Viktor von Wyl University of Zurich, Epidemiology, Biostatistics and Prevention Institute, UZH Digital Society Initiative #MultipleSclerosis #MSResearch #HealthcareCommunication #DigitalHealth #MentalHealth #NLP #uzh

How do persons with multiple sclerosis (#MS) explain the onset of their own disease? Read about the theories of persons living with MS in our new paper from the University of Zurich Digital & Mobile Health group and the Swiss Multiple Sclerosis Registry. Don't miss the accompanying blog post by the first and last authors, Christina Haag and Viktor von Wyl! https://lnkd.in/g5FhSDzy

The emergence of new digital technologies has enabled a new way of doing research, including active collaboration with the public. Text data is a powerful way to allow individuals to freely express their experiences in their own words. This is especially powerful when conducting large-scale studies involving and collaborating with the public.   In "Blending citizen science with natural language processing and machine learning: Understanding the experience of living with multiple sclerosis", the authors show how text analysis can be implemented using the example of the Swiss Multiple Sclerosis Registry, a longitudinal patient-centered project. Persons with multiple sclerosis (MS) were asked to tell the story of their life with MS in their own words, guided by open-ended questions. The authors found that important life events for persons with MS go beyond typical clinical milestones such as diagnosis and changes in symptoms or treatment.   Full article ⬇ Christina Haag, Nina Steinemann, Deborah Chiavi, Christian P. Kamm, Chloé Sieber, Zina-Mary Manjaly, Gábor Horváth, Vladeta Ajdacic-Gross, Milo Puhan, Viktor von Wyl University of Zurich, Epidemiology, Biostatistics and Prevention Institute Schulthess Klinik #ehealth #healthtech #MedicalKnowledge #implementationscience #digital #digitalhealth #uzh

This is a highly revealing discussion pointing to fundamental limitations in applications such as ChatGPT. The importance of causal reasoning for successful AI applications continues to be grossly understated. The focus of AI seems to be more so on predictions and less on causal analysis which is a critical flaw. There appears to be a systemic gap in the understanding of causal reasoning in the field of data science. There should be a stronger emphasis in the educational system for data scientists to acquire skills in causal reasoning especially from other domain experts from fields such as epidemiology, risk analysis and social sciences.

🔰 Early Detection of Plant Viral Disease Using Hyperspectral Imaging and Deep Learning ✳ We are excited to share news of a groundbreaking paper titled "Early Detection of Plant Viral Disease Using Hyperspectral Imaging and Deep Learning." The paper explores the use of hyperspectral imaging and machine learning to detect the grapevine vein-clearing virus (GVCV) in Chardonel grapevines. This research is particularly important given the threat GVCV poses to grapevine sustainability in the Midwest region of the U.S.A. ✳ Traditional methods for detecting plant disease are time-consuming and not ideal for large-scale intervention. However, hyperspectral imaging shows great potential as a noninvasive and nondestructive tool for monitoring plant biotic and abiotic stress. By capturing and storing an object's spectroscopy information in a spectral cube, hyperspectral imaging offers many opportunities for early sensing of plant diseases. 📑 The paper explores several approaches to processing hyperspectral imagery data, including pixel-wise methods that extract only spectral information and image-wise methods that extract joint spatial-spectral information. The authors also tested traditional machine learning algorithms, such as support vector machines and random forests, as well as deep learning approaches, such as convolutional neural networks. 📝 The results of this research are promising, demonstrating the potential for hyperspectral imaging and machine learning to accurately detect GVCV infections in grapevines at an early stage. This technology has the potential to revolutionize the way we monitor plant health, providing previsual indicators via subtle changes in spectral reflectance due to absorption or reflectance. ⬇ 📑 Paper Can be accessed from following link: https://lnkd.in/dGR5uk3p ______________________________________HYSPIM___________________________________  💻 For more information visit our website: 🌐HYSPIM.COM ______________________________________HYSPIM__________________________________ #HyperspectralImaging #DeepLearning  #PlantDiseaseDetection #PrecisionFarming  #Agriculture #GrapevineVeinClearingVirus  #ClimateChange #RemoteSensing  #MachineLearning

🚨Calling all graduate students in #environmentalscience and #epidemiology in London🚨 The UK's national scientific academy is investigating how AI tools like ChatGPT might generate, or help spread, scientific misinformation. The Royal Society's motto is nullius in verba: take nobody's word for it. That applies to ChatGPT as well.* We still have some spaces left to join our big experiment on 25 October! Come hang out, have afternoon tea, and help us test these technologies. No tech background required--just a laptop (with Chrome), your science expertise, and plenty of critical thinking. Sign up below! 👇🏻 And please do share 🙏🏻 https://lnkd.in/geF3WanQ #sciencepolicy #misinformation #environment #climatechange *we'll actually be using LLaMA for this

Worthy 5 min read on the serious ramifications of Bad (biased, racist, sexist, etc.) data in = Bad (biased, racist, sexist, etc.) data out, when it comes to AI

I am thrilled to share my latest paper, "Deep Learning Guided Prediction Modeling of Dengue Virus Evolving Serotype," that has just been published in ScienceDirect. As a first-author from my PhD thesis co-authored with Muhammad Zubair Yousaf , Dr. Rashid Saif and zubia rashid. This research marks a significant stride in understanding how viruses like the Dengue Virus (DENV) evolve and adapt, impacting public health globally. 🧬 Key Insights: - Our study highlights the relentless evolution of viruses and their ability to escape immune responses, complicating diagnostics and vaccination efforts. - We introduced the DL-DVE model, which utilizes Long Short-Term Memory (LSTM) and Feedforward Neural Networks (FNN) trained on complete genome sequences of DENV. This model not only learns complex patterns but also predicts emerging DENV serotypes with high precision. 📊 Achievements: - The generative model demonstrated an impressive accuracy of 93%. - Our classification model identified specific serotypes with an ROC-AUC value of 0.818 and maintained nearly 99% accuracy, precision, recall, and F1 score. This paper underscores the critical role of deep learning in advancing our understanding of life-threatening viruses and enhancing our predictive capabilities in public health. Feel free to read more about our findings and methodologies here: https://lnkd.in/dVKVtuhG Explore our model and code on GitHub: https://lnkd.in/d4tH8BuK Let's continue pushing the boundaries of science and technology to safeguard global health! #Virology #DeepLearning #PublicHealth #DengueVirus #AI #ScienceResearch