Panagiotis Zervas’ Post

From our COO Panagiotis Zervas: the story behind GARDIAN and what building a global research discovery platform actually looks like in practice. 👇

New from the Frass Institute: Insect Frass Research Is Growing: A 2021–2026 Snapshot We reviewed recent research on black soldier fly, mealworm, and cricket frass to better understand what the field is studying, where the evidence is growing, and where shared learning can help move the field forward. The snapshot draws from 135 scholarly outputs published or released from 2021 through May 2026. The big takeaway: frass research is growing quickly, and the questions are becoming more specific. Less “can frass work?” More “how, where, under what conditions, and for what outcomes?” Read the full snapshot here: https://lnkd.in/eMtiqqV7 #Frass #SoilHealth #AgResearch #InsectAgriculture

The Frass Institute is a gathering place of research and soil focused folks around insect derived byproducts. I would definitely go check out this article, check out the website, and reach out to the institute if soil health is important to you. I am currently working to use frass here in California to boost vineyard health and vegetable production. Let’s chat if this sounds interesting! 🐛

I am delighted to share that our research paper, 𝗟𝗲𝘃𝗲𝗿𝗮𝗴𝗶𝗻𝗴 𝗮 𝗖𝘂𝘀𝘁𝗼𝗺 𝗦𝗰𝗮𝗹𝗮𝗯𝗹𝗲 𝗖𝗡𝗡 𝗙𝗿𝗮𝗺𝗲𝘄𝗼𝗿𝗸 𝗳𝗼𝗿 𝗘𝗳𝗳𝗶𝗰𝗶𝗲𝗻𝘁 𝗣𝗹𝗮𝗻𝘁 𝗗𝗶𝘀𝗲𝗮𝘀𝗲 𝗖𝗹𝗮𝘀𝘀𝗶𝗳𝗶𝗰𝗮𝘁𝗶𝗼𝗻, has been published in 𝗦𝗽𝗿𝗶𝗻𝗴𝗲𝗿 𝗡𝗮𝘁𝘂𝗿𝗲. Plant disease detection plays a crucial role in protecting agricultural productivity and supporting timely intervention. In this work, we explored a deep learning-based approach for identifying diseases across multiple crop species, including 𝗰𝗼𝗿𝗻, 𝗽𝗼𝘁𝗮𝘁𝗼, 𝗿𝗶𝗰𝗲, 𝘁𝗲𝗮, 𝗮𝗻𝗱 𝘁𝗼𝗺𝗮𝘁𝗼. Our study combines a fine-tuned 𝗥𝗲𝘀𝗡𝗲𝘁𝟱𝟬 for strong classification performance with a lightweight custom CNN designed for efficient deployment on resource-constrained systems. Using more than 𝟮𝟳,𝟬𝟬𝟬 labeled images across five datasets, our framework achieved highly promising results, including up to 𝟵𝟵.𝟯𝟯% validation accuracy on rice disease classification. The study further highlights the potential of deep learning in enabling accurate, scalable, and real-time plant disease diagnosis. I am sincerely grateful to my co-authors Subhrajit Ghosh, Aditri Chaudhuri, Debangshu Roy, Prasenjit Datta, and Soumyadip Dutta, and deeply thankful to our mentors Dr. Bidhan Barai and Supriya Sarkar for their valuable guidance and support. Special thanks to Pawan Kumar Singh Sir as well. 🔗 Paper Link: https://lnkd.in/gS6FBXQF #ResearchPaper #SpringerNature #PublishedPaper #DeepLearning #MachineLearning #ArtificialIntelligence #CNN #ComputerVision #PlantDiseaseDetection #SmartAgriculture #AgricultureAI #AcademicResearch #AIinAgriculture

🚨 Course announcement 🚨 We are excited to launch GLLVMs: Advanced multivariate analysis of ecological communities in R (7–10 July, online) with Bert van der Veen This advanced course introduces Generalized Linear Latent Variable Models (GLLVMs), a modern statistical framework for analysing multivariate ecological community data. It provides a powerful, model-based alternative to classical ordination methods such as CCA, NMDS and PCoA, integrating ideas from GLMs, GLMMs, and joint species distribution models. Across four days of lectures and hands-on R sessions, participants will learn how to fit, diagnose, and interpret GLLVMs using the gllvm package, and how to translate model outputs into ecological insight. Topics include: • Multispecies GLMs and GLMMs • Joint Species Distribution Models • Model-based ordination • Covariate and random effect structures • Model diagnostics and interpretation The course is designed for researchers working with multivariate ecological data who want a modern, statistically rigorous approach to community analysis. #Ecology #Statistics #RStats #EcologicalModelling #Bioinformatics #DataScience #GLLVM #EcologyResearch https://lnkd.in/dD8jYPU6

New paper by Prof. Gilad Ravid from our department (with Yafit Cohen, Amit Malka, Yonatan Goldwasser, Elia Scudiero, Eitan Goldshtein, Ohaliav Keisar, and Guy Lidor), published in Computers and Electronics in Agriculture! The study introduces a machine-learning framework that uses Sentinel-2 satellite imagery and historical yield records from 123 maize fields in Israel (2018-2021) to classify field performance as early as 5 days after sowing. By integrating spatial, temporal, and spatio-temporal features beyond traditional vegetation indices, the approach improved classification accuracy by an average of 5.3% (F1-score). The framework has been deployed as a free web-based decision support tool for farmers. Gilad Ravid This research was funded by the Chief Scientist of the Israeli Ministry of Agriculture and Food Security, Netafim, and the Israel Association of Field Crop Growers. Read more: https://lnkd.in/dWDjpiPs

New commentary: Debunking the myth of the quintessential resource manager: Precision in actionable science. Thanks to a great team of USGS coauthors (led by Amanda E. Cravens, Nicole Herman-Mercer, Aparna Bamzai-Dodson, and Megan Moore, Ph.D.) on this new (open-access) commentary that came out last week in Conservation Science and Practice. We argue it's important to understand managers as individuals in order to effectively create actionable science and we share strategies we've used to do so. Check it out here: https://lnkd.in/gphfqPsZ.

PEACE doesn't only exist in the absence of war. It exists in the quiet work trenches of progress. Of guided innovation. Of governmental policymaking. Of industries and practitioners that seek solutions and freedom for exploration. Read the #CoverStory interview with the award-winning Dr. Catherine Nakalembe and a powerful, dynamically journalistic cache of content in this upcoming "POWER EXPOSED" Edition! #staytuned #comingsoon #GTM39 📢 NOTE: The #GlobalTrendsMagazine is produced by the #GlobalChamberOfBusinessLeaders, which is the force behind the #GlobalCoalitionForCooperationAndPeace (GCCP). https://lnkd.in/eDxt_sYa

Spearheading the preparation of exclusive interviews with world leaders, high-profile professionals and innovative entrepreneurs... One thing that I'm adamant about is 'breaking things down' where necessary because industry jargon can be well, jarring for some readers. ⏱️ COMING SOON! In this upcoming Edition of the #GlobalTrendsMagazine (GTM), my interview with the award-winning #CatherineNakalembe presented her with the opportunity to give the face of science, data and actionable governance a makeover. The #MayIssue is mere days away! "Where Trends Are Born, Not Followed." #GTM39 #globaltrendsmagazine2026 #africa #usa #datasets #agriculture #chiefeditorlife

We are anticipating sharing this #CoverStory and 39th Edition of our #GlobalTrendsMagazine (GTM) with you. Stay tuned! See relevant details in the original post. #comingsoon #GTM39 #globaltrendsmagazine2026 #globalchamberofbusinessleaders #digitalmagazine #pressreader https://lnkd.in/eDxt_sYa