Foto di copertina di DHIRIA S.r.l.
DHIRIA S.r.l.

DHIRIA S.r.l.

Servizi dati per sistemi IT

Milan, Lombardy 135 follower

Dhiria is the first company in the world to provide machine and deep learning services able to operate on encrypted data

Chi siamo

What about processing directly the encrypted data? What about receiving back the results of the processing still encrypted? Processing directly the encrypted data and providing back encrypted results guarantee that nobody can look at your data. What, that could be MAGIC. No! That's Dhiria! Dhiria, first in the world, developed a technology, avaible in as-a-service manner, allowing the processing by means of machine and deep learning algorithms of encrypted data. All the companies and public administrations relying on Big Data processing Cloud services based on Artificial Intelligence will be able to access privacy-preserving services thanks to the as-a-service platform provided by Dhiria. Why should you be interested in our technology? Because you are certainly not willing to provide your identity and your data to those who are not allowed to access them.

Settore
Servizi dati per sistemi IT
Dimensioni dell’azienda
2-10 dipendenti
Sede principale
Milan, Lombardy
Tipo
Società privata non quotata
Data di fondazione
2021
Settori di competenza
Cloud, Privacy, Big Data, Machine Learning, Artificial Intelligence e IT technology

Località

Dipendenti presso DHIRIA S.r.l.

Aggiornamenti

  • Visualizza la pagina dell’organizzazione DHIRIA S.r.l.

    135 follower

    🎙️ Today, Manuel Roveri will be speaking at the Euro-Mediterranean Festival of Economy (FeuroMed) about artificial intelligence and its impact on businesses and society. As a co-founder of Dhiria, a Politecnico di Milano spin-off specializing in privacy-preserving machine learning, he will discuss the challenges and opportunities of AI for companies, exploring how businesses can leverage AI to extract value from enterprise data. AI is redefining the business landscape, from data management to process optimization and new business models. But how can we balance innovation and data accessibility? These key topics will be addressed today at FeuroMed. More details about the festival: Check the first comment for the link. #AI #PrivacyPreservingAI #FeuroMed #MachineLearning #Business #Innovation #ArtificialIntelligence

  • From January 4 to 8, 2025, our researcher Aurora AF Colombo attended the 30th edition of the Pacific Symposium on Biocomputing (PSB) on the Big Island of Hawaii. PSB is an international, multidisciplinary conference that fosters collaboration among biologists, computer scientists, physicians, and experts interested in computational methods applied to biology. During this prestigious event, Aurora AF Colombo presented a study titled “Enhancing Privacy-Preserving Cancer Classification with Convolutional Neural Networks”, conducted in collaboration with Luca Colombo, Alessandro Falcetta, and Manuel Roveri. The research introduces an innovative approach based on convolutional neural networks to identify the primary tumor site in oncology patients, while integrating 𝗲𝗻𝗰𝗿𝘆𝗽𝘁𝗲𝗱 𝗱𝗮𝘁𝗮 𝗽𝗿𝗼𝗰𝗲𝘀𝘀𝗶𝗻𝗴 𝘁𝗵𝗿𝗼𝘂𝗴𝗵 𝗵𝗼𝗺𝗼𝗺𝗼𝗿𝗽𝗵𝗶𝗰 𝗲𝗻𝗰𝗿𝘆𝗽𝘁𝗶𝗼𝗻. This enables data to be processed without decryption, ensuring the highest levels of security and privacy for patients. We are proud of Aurora AF Colombo’s contribution and our team’s continued efforts to advance scientific knowledge. At Dhiria, we remain committed to investing in Privacy-Preserving Computation Technology powered by Artificial Intelligence and Homomorphic Encryption, further expanding the range of its applications. Check the first comment for the link to the PSB paper and the official conference website.

    • Nessuna descrizione alternativa per questa immagine
  • We are pleased to announce that Aurora AF Colombo has joined the Dhiria team as a new researcher! 🎓👩🔬 Aurora is pursuing a doctoral program funded by Dhiria at the Politecnico di Milano, focusing on the application of Homomorphic Encryption technologies to process encrypted medical data for precision medicine. 🧬🔒 This research will be closely aligned with Dhiria’s activities, and we will support her with a dedicated scholarship. 🎓💰 Already an active member of our team, Aurora represented Dhiria at the Hackathon AI x Oncology on December 3, 2024. 🧠🤖 This event allowed us to forge new collaborations with medical experts and share the advancements of our research in a still largely unexplored field. By adopting homomorphic encryption, Dhiria ensures complete confidentiality of patient data in precision medicine, enabling direct processing of encrypted information without ever compromising security. 🔐🩺 This technology represents a breakthrough in safeguarding patients’ rights and strengthening trust in the use of Artificial Intelligence in personalized medical processes. 🚀💡 With Aurora’s arrival, Dhiria will accelerate the development of innovative solutions to improve diagnostic and treatment processes, maintaining the highest level of sensitive data protection. 🏥📊 We continue to invest and innovate, making our Privacy Preserving Computation technologies available to all sectors that, like the medical field, require the secure and confidential handling of sensitive data. 💻🔒 Welcome, Aurora! 👋🎉 (Link to Aurora's post in the first comment! 🔗💬)

  • 🚀 We are thrilled to share some exciting news! 🚀 Our Luca Colombo participated in the prestigious “7th HomomorphicEncryption.org Standards Meeting” in Salt Lake City, USA! This meeting, organized by HomomorphicEncryption.org, brings together the top experts in the field: industry leaders, government representatives, and academics. We’re talking about an exclusive circle of specialists in homomorphic encryption — a revolutionary technology that is key to Dhiria’s products. 📍 The event was an unmissable opportunity to discuss: - Standardization of homomorphic encryption. - Software tools for developing and testing innovative solutions. - New approaches to improve efficiency. - Emerging industrial applications. 💡 A rich space for exchange and collaboration with world-class experts, where ideas and innovations come to life. Our participation reaffirms Dhiria’s commitment to staying at the forefront of the development and adoption of this groundbreaking technology. 👉 Stay tuned to see how we continue to push the boundaries of innovation in homomorphic encryption!

    • Nessuna descrizione alternativa per questa immagine
  • Visualizza la pagina dell’organizzazione DHIRIA S.r.l.

    135 follower

    Dhiria had the privilege of visiting the European Space Agency to discuss the importance of privacy-preserving computation in the processing of Earth Observation data. Dhiria’s solutions have shown great results, both on the EuroSAT benchmark of Sentinel-2 and in the EO4SEC project, aimed at monitoring illegal activities in the Mekong region. Dhiria’s focus in this field includes identifying changes and anomalies in encrypted image sequences and privacy-preserving object detection in encrypted images. Looking ahead, Dhiria plans to enable not only inference but also fully encrypted training, where models are generated directly from encrypted data. Manuel Roveri

    • Nessuna descrizione alternativa per questa immagine
  • 🌍 Dhiria at the forefront of future mobility! 🚗💡 We are proud to share that our colleague Riccardo Pazzi participated in an event organized by Openstage on the theme of urban mobility, in collaboration with the science communication association On Air - Voce alla scienza. The event brought together emerging voices in scientific outreach alongside established communicators, creating a space for discussions on crucial topics related to mobility. 🚀 During the event, in addition to talks from three members of On Air - Voce alla scienza, a "call for ideas" was held to evaluate innovative proposals for the future of mobility. In his presentation, Riccardo Pazzi addressed how privacy-preserving technologies can be used to improve mobility. Since personal mobility data contains sensitive information, it is crucial to process it securely. Riccardo explained how these technological solutions can not only ensure privacy but also contribute to making mobility safer and more efficient. We are proud of Riccardo Pazzi’s contribution to the event, bringing forward his ideas on how technology can enhance urban mobility in a secure and efficient way.

    • Nessuna descrizione alternativa per questa immagine
  • DHIRIA S.r.l. ha diffuso questo post

    I'm thrilled to announce that I had the opportunity to present the research paper, "Training Encrypted Neural Networks on Encrypted Data with Fully Homomorphic Encryption", at the 12th Workshop on Encrypted Computing & Applied Homomorphic Cryptography (WAHC '24), part of the ACM CCS conference held in Salt Lake City, USA! 🎉 This work, developed with Alessandro Falcetta and Manuel Roveri, represents a significant step towards building end-to-end privacy-preserving deep learning as-a-service solutions. By leveraging Fully Homomorphic Encryption, we demonstrate the feasibility of training machine learning models directly on encrypted data, ensuring that sensitive information is never exposed during the process. This approach paves the way for future AI applications in sectors where privacy is paramount, such as healthcare, finance, and critical infrastructure. It was an amazing experience sharing this work and exchanging ideas with the HomomorphicEncryption.org community. Thank you to everyone who joined and contributed to such insightful discussions. 🚀 AI-Tech Research Lab DHIRIA S.r.l. #AI #DeepLearning #PrivacyPreservingAI #HomomorphicEncryption #FHE #WAHC24

    • Nessuna descrizione alternativa per questa immagine
    • Nessuna descrizione alternativa per questa immagine
    • Nessuna descrizione alternativa per questa immagine
    • Nessuna descrizione alternativa per questa immagine
  • 🌍 Dhiria at WAHC 2024! 🔐🤖 We are excited to announce that our colleague Luca Colombo presented the paper "Training Encrypted Neural Networks on Encrypted Data with Fully Homomorphic Encryption" at the WAHC 2024 - 12th Workshop on Encrypted Computing & Applied Homomorphic Cryptography, part of the ACM Conference on Computer and Communications Security in Salt Lake City. 🚀 In this work, Luca Colombo, Manuel Roveri e Alessandro Falcetta explored a crucial topic for the future of cryptography and artificial intelligence: how to train encrypted neural networks on encrypted data using fully homomorphic encryption. This cutting-edge solution paves the way for new possibilities in AI, ensuring the security and privacy of sensitive data throughout the processing phase. We are incredibly proud of Luca’s contribution and the Dhiria team’s involvement in such an important workshop for the future of applied cryptography! #AI #DeepLearning #PrivacyPreservingAI #HomomorphicEncryption #FHE #WAHC24

    • Nessuna descrizione alternativa per questa immagine
  • 🌍 Incredible experiences at the AI-ML Systems Conference 🎓In addition to our own presentations, the Dhiria team had the chance to attend high-level keynotes, workshops, and tutorials that enriched our knowledge in key areas of AI. 🧠 Keynote by Dan Roth We attended an insightful session by Prof. Dan Roth on the reasoning capabilities of Large Language Models (LLMs)—such as ChatGPT and Google Gemini. Prof. Roth highlighted a critical issue: the trust users place in these tools despite their current inability to reason logically. While these models provide answers that seem plausible, true logic and reasoning remain uniquely human qualities. There’s still much work ahead! 🔍 🚀 “AI in Space” Workshop with Prof. Brandon Lucia Prof. Brandon Lucia presented a space-edge AI chip designed to operate in the harsh conditions of space, addressing challenges like energy consumption and extreme environments. This chip accelerates AI operations on satellites with unprecedented efficiency. What impressed us most is its support for advanced encryption algorithms, possibly including homomorphic encryption algorithms like the ones we use at Dhiria. A groundbreaking advancement for AI in space!" 📊 Graph Deep Learning Tutorial with Prof. Davide Bacciu Thanks to Prof. Davide Bacciu’s tutorial, we delved into the world of Graph Deep Learning, a technology crucial for working with graph data structures. This has powerful applications in advanced fields such as medicine, social network analysis, and physics. We are excited to explore how this innovative research area could open new paths for Dhiria. Participating in these events is always valuable for engaging with other organizations within the scientific community and the industry, in order to enrich one's knowledge base. 🎯 These experiences have inspired our entire team, reaffirming the importance of staying at the forefront of artificial intelligence advancements. We can’t wait to apply this new knowledge to our projects! Manuel Roveri Alessandro Falcetta Luca Colombo Francesco Puoti

Pagine simili