Juan M. Lavista Ferres

Available 4/9, our book from @WileyGlobal shares how applying AI helps improve community resiliency, health outcomes, information integrity, and conserve Earth's resources. aka.ms/AIforGoodBook Proceeds benefit @RedCross. #AIforGoodBook

By Juan Lavista Ferres (@BDataScientist), Elliot K. Fishman (@ctisus), Ed Catmull, Bert Vogelstein, and Joshua T. Vogelstein (@neuro_data) at @HopkinsMedicine @Microsoft AI for Good Lab @pixar

The Curator: a special edition from the #WEF24 Annual Meeting in Davos. Featuring Mário Centeno, the governor of the Bank of Portugal, Juan Lavista, Microsoft’s vice-president, Pritzker Prize winner Francis Kére and more monocle.com/radio/shows/th…

"AI has transformed industries, but it still has significant limitations, including a lack of true understanding, common sense reasoning, and independent thought." Check out my Q&A with Juan Lavista Ferres from Microsoft's AI For Good Lab. hub.jhu.edu/magazine/2025/…

🎥 New clip: How can data help people see, fight disease, and protect human rights? Juan Lavista Ferres, Microsoft’s Chief Data Scientist, shows us how.

We’re pleased to announce the judging panel for the #AIforGood Impact Awards 2025, organized by the @ITU in partnership with @TechToTheRescue! Following a global call that attracted over 320 applications, expert judges are now reviewing submissions across the categories of AI for People, AI for Planet, AI for Prosperity, and Pro Bono Collaboration, recognizing innovative and impactful #AI solutions that contribute to global progress. General Awards Panel: 🔹 @jamoussi, Deputy to the Director, Chief of the Study Groups and Policy Department, TSB, International Telecommunication Union 🔹 Abby Daniell, Director Worldwide Public Sector, @AWS 🔹 Lan Xue, Distinguished Professor and Dean, Schwarzman College, @Tsinghua_Uni 🔹 @BDataScientist, Corporate Vice President & Chief Data Scientist, AI for Good Lab, @Microsoft Pro Bono Collaboration Panel: 🔹 Najwa Aaraj, CEO, @TIIuae 🔹 @chrisfabian, Co-Founder, Co-lead, @Gigaglobal, @UNICEF 🔹 @maryajacques, Executive Director, Global ESG and Regulatory Compliance, @Lenovo 🔹 @siadkowski, Co-Founder and CEO, Tech To The Rescue 📣Finalists will be announced on 10 June and invited to the AI for Good Global Summit in Geneva for the Awards Ceremony on 9 July, where the winners in each category will be revealed. Don't miss out on the announcement! Join us for free or go VIP! 🌐 loom.ly/3lYtLDg 📅 8–11 July 2025 📍 @Palexpo, Geneva & online

Another reason to mark your calendar for the #AIforGood Global Summit 2025! 🗣 @BDataScientist, Corporate Vice President & Chief Data Scientist of the AI for Good Lab at @Microsoft, will join us to share their insights on the topic of AI for Good. Join us for free or go VIP! 🌐 loom.ly/3lYtLDg 📅 8–11 July 2025 📍 @Palexpo, Geneva, and online

AI is not just helpful in healthcare. It’s the only scalable solution in some cases, says @BDataScientist. During a #GlobalStage conversation, he explains how AI is being used to diagnose a leading cause of childhood blindness in newborns.

.@WithSecure's @rsiilasmaa participated in a panel discussion at the 10th Multi-stakeholder Forum on Science, Technology and Innovation, co-chaired by Finland, together with Professor Joyeeta Gupta, @BDataScientist, Keywan Riahi and @mleptin. #STIForum10 #Tech4SDGs #GlobalGoals

How are technological advancements shaping global power dynamics? Join us live from the UN Science, Technology, and Innovation Forum. May 7 at 9 am ET: Tune in for a #GlobalStage panel discussion with @jjding99 & @BDataScientist. @MSFTIssues gzeromedia.com/globalstage

Bill Gates and Paul Allen foresaw Moore's law reducing computing costs, leading to the digital revolution. @BDataScientist's new report shows #AI following a similar path—making it available to every person on the planet. @StanfordHAI

I'm eternally grateful to Bill and Steve and to Paul for what you've meant to me personally and the vision you had to build this extraordinary company of ours that I've had the privilege to be a part of. Thank You for your leadership, your passion, and for building the Microsoft that we know today. A company that has truly changed the world.

Check out my latest article: Global Renewables Watch: A New Era of Energy Insights linkedin.com/pulse/global-r… via @LinkedIn

⁦@CopagranUy⁩

¿Sabías que la tecnología puede proteger la biodiversidad en los rincones más remotos del planeta? 🔸Hablamos sobre ello con Juan M. Lavista (@BDataScientist), vicepresidente corporativo de @Microsoft y científico jefe del AI For Good Lab En Geek5D⤵️ rtve.es/a/16492053

Check out my latest article: Unlocking the Secrets of Proteins with AI  linkedin.com/pulse/unlockin…

Rapid and accurate prediction of protein homo-oligomer symmetry using Seq2Symm @NatureComms 1/ This study presents Seq2Symm, a deep learning model fine-tuned from ESM2 for predicting the symmetry of homo-oligomeric protein assemblies, outperforming template-based and prior deep learning methods in accuracy and speed. 2/ Homo-oligomers, protein complexes composed of identical subunits, play essential roles in stability, function, and molecular recognition. Predicting their symmetry is crucial for structural biology, but existing methods rely on homology templates or computationally expensive structural modeling. 3/ Seq2Symm uses only a protein's sequence as input and predicts its homo-oligomeric symmetry at a rate of approximately 80,000 proteins per hour, making it scalable for proteome-wide analyses. 4/ Compared to template-based methods like HHSearch and prior deep learning approaches, Seq2Symm achieves significantly higher AUC-PR scores, improving prediction accuracy across cyclic, dihedral, and helical symmetry classes. 5/ The model was trained on a large dataset derived from the Protein Data Bank (PDB), covering diverse symmetry groups. It was evaluated on three independent test sets, demonstrating superior performance in predicting complex symmetries. 6/ Seq2Symm can be integrated with structure prediction tools like AlphaFold2-multimer and RoseTTAFold2 to guide accurate modeling of homo-oligomeric protein assemblies, reducing computational costs compared to exhaustive brute-force searches. 7/ Large-scale application of Seq2Symm to five proteomes and 3.5 million unlabeled protein sequences from metagenomic sources revealed patterns in symmetry distribution across different biological kingdoms, highlighting its potential for evolutionary studies. 8/ The study suggests that fine-tuning protein language models specifically for symmetry prediction provides significant advantages over general-purpose sequence embeddings, paving the way for further advances in deep learning-driven structural biology. @meghanaksagar @Artur_Mell @IanRHum @samsledzieski @lab_berger @BDataScientist @minkbaek 💻Code: github.com/microsoft/seq2… 📜Paper: nature.com/articles/s4146… #ProteinSymmetry #DeepLearning #StructuralBiology #Bioinformatics #AlphaFold #ProteinComplexes #ComputationalBiology

An expanded version of our TorchGeo paper has been accepted to ACM TSAS @TheOfficialACM In addition to even more benchmarks on remote sensing datasets, we provide a thorough survey of open-source libraries for deep learning for geospatial data @calebrob6 @BDataScientist

The humanitarian sector needs innovation and ideas. Excellent brainstorming meetings with @BDataScientist @Microsoft & @Googleorg VP Maggie Johnson on mapping needs, anticipating future crises and making it easier for individuals to help us save lives. #Davos2025

Our partnership with @Microsoft has enhanced IOM’s ability to support people on the move worldwide, harnessing AI and tech to deliver innovative solutions. Great to meet with @KateBehncken and @BDataScientist to explore our next steps together! #WEF2025

Are you in Davos? Join us tomorrow at the Ice Village Dome C at 15.45 and hear from innovators and disrupters how we can tap into #data and #AI to revolutionise urban #health - with @worldheartfed @NovartisFDN with speakers from Microsoft @BDataScientist @SwissRe_PS