Deix Plenary 2023

**2023: A Year of Expansion and Innovation**

2023 has been a memorable year for us at Deix, marked by significant expansion in our client portfolio and entry into new sectors. This year has seen us not only consolidating our existing partnerships but also exploring new collaborations that have enriched our professional journey.

**Revolutionary Innovations: Varda-AI and Jinn**

One of the highlights of 2023 was the continued investment in creating vertical products, leading to the birth of two revolutionary innovations: [Varda-AI](/en/solutions/products/varda-ai) and [Jinn](/en/solutions/products/jinn). These solutions, a result of our commitment to research and development, represent a significant step towards the future of technology and automation.

**A Moment of Unity and Celebration**

This year's plenary meeting was the perfect occasion to take stock of our internal and external activities and to reflect on future prospects. But it wasn't just a moment of work: we also had the opportunity to celebrate together, with a lunch that strengthened the team spirit and got us into the festive mood. It was a time to celebrate the successes achieved and to recognize the work of every member of the Deix team.

**Looking to the Future**

As we close a chapter full of successes, we look forward with enthusiasm and confidence. 2024 promises to be a year of further challenges and opportunities, and we are ready to make the most of them, carrying forward our mission of innovation and excellence.

Thank you all for being part of this incredible journey, and let's get ready for another year of triumphs and achievements.

A Year of growth, new beginnings, and team spirit

Deix was among the 10 startups participating in the event

Open Innovation Day CTE.COBO

We attended the Open Innovation Day event at the House of Emerging Technologies of the Municipality of Bologna, and came back rich with new connections and inspirations.

Along with 9 other selected startups, we had the opportunity to interact with several corporations. It was an exceptional occasion to showcase Deix's distinctive competencies in Artificial Intelligence and Optimization, and to explore potential future collaborations.

We thank ART-ER, who organized the event, for this fantastic opportunity and everyone who participated and shared their vision and experiences with us.

Borsa della Ricerca 2023

We are excited to share the wonderful experience we had at the 14th National Research Exchange Forum in Catania, held at the enchanting Monastery of the Benedictines. The Monastery provided an incredibly picturesque setting for an event that lived up to expectations and then some.

But what truly made this edition of the Forum unforgettable was the opportunity to meet and connect with many exceptional companies and professionals. The carefully planned one-to-one meetings before the event served as a springboard for deep and meaningful discussions. We were able to share our experiences, discover new perspectives, and create connections that could open incredible doors for the future. It was truly inspiring to see the innovative potential in our country, and the Forum was the fertile ground to nurture it.

A warm thank you to the University of Catania and the entire Forum team for creating such a valuable experience.

An unforgettable experience in Catania at the Research Exchange Forum

Two new technical books from our scientific board

We are pleased to announce the release of two books authored by members of our scientific committee.

These books cover topics that form the foundation of many activities we engage in at Deix, and it is gratifying to see them featured in dedicated publications.

**Introduction to Methods for Nonlinear Optimization**

*L. Grippo, M. Sciandrone* - *Springer, 2023*

This book has two main objectives :
- to provide a concise introduction to nonlinear optimization methods, which can be used as a textbook at a graduate or upper undergraduate level;
- to collect and organize some important topics on optimization algorithms, not easily found in textbooks, which can provide material for advanced courses or can serve as a reference text for self-study and research.

**Complementi di Matematica**

*G. Liuzzi, M. Sciandrone - Hoepli, 2023*

Designed for undergraduate students in engineering programs, "Complements of Mathematics" introduces fundamental mathematical concepts for the study of properties of sets in n-dimensional Euclidean space and functions of multiple real variables.

The approach taken diverges significantly from the traditional one, allowing students to engage with algorithmic mathematics geared towards modeling and numerically solving real-world problems.

AI will save the world

Daring? Absolutely, but behind this futuristic vision lies a tangible reality. While the title might appear provocative, by exploring the potential of Green AI, we can uncover how artificial intelligence is indeed playing a significant role in the pursuit of sustainable solutions to address global environmental challenges.
However, caution is warranted; it's essential to view this reality as a two-faced Janus entity. In fact, AI can - at the same time - both save and endanger the environment we inhabit.

From a positive point of view, AI can play (and is already doing so!) a vital role in searching for solutions to support sustainability and the environment. For example, by leveraging data collected from machinery and IoT devices, it is possible to reduce energy consumption or material waste. In the context of smart cities, historical mobility and behavior prediction data provide a fundamental contribution to making our cities more livable.

The use of AI, however, can pose **risks** to the environment. For instance, the implementation of AI infrastructure requires a significant amount of energy and resources, especially for data processing in large data centers. Additionally, widespread use of AI could lead to an increase in the production and disposal of electronic devices, contributing to the rise in electronic waste.

- **High Energy Consumption** in Data Centers: training and running artificial intelligence models require a significant amount of energy, especially in data centers. This can have a substantial impact on the environment, contributing to carbon emissions and the use of non-renewable energy resources. It is important to explore solutions to reduce energy consumption, such as model optimization, the adoption of energy-efficient hardware, and the use of renewable energy sources to power data centers.
- Training machine learning models requires a **vast amount of data**, and maintaining a historical record of this data over time can consume a significant amount of resources, including storage space, energy, and processing capacity. This can have an environmental impact, as data centers hosting these resources consume energy and generate waste. In the development of new models and algorithms, it is important to question which and how much information is actually valuable: when it comes to data, it's easy to accumulate too much of low-added-value data.
- **Electronic Waste Generation**: artificial intelligence relies on sophisticated hardware, such as servers, storage devices, and sensors, which can contribute to the growing production of electronic waste. It is essential to explore strategies to address this issue, such as responsible recycling of electronic components, extending the useful life of devices, and adopting modular design to facilitate the replacement and upgrading of obsolete components.
- Use of **Unsustainable Materials**: Some components of the hardware used in artificial intelligence may require the use of expensive, rare, or unsustainable materials. This raises concerns about resource availability and the environmental impact associated with the extraction and production of such materials.

Are there **solutions** to these problems?

Green AI focuses on creating energy-efficient machine learning models, optimizing data collection processes, and utilizing adaptation and model evolution techniques that are environmentally responsible in terms of resource management.

The traditional lifecycle of a machine learning algorithm involves starting from scratch whenever you want to update the model or train it for a new task. Typically, in this process, you either expand the training dataset with new data or modify the model's structure, allowing it to evolve over time. This process presents two major challenges. On one hand, it's necessary to maintain a memory of all the data over time, leading to resource usage and privacy risks. On the other hand, starting the training process from scratch consumes a significant amount of energy, especially for more advanced models.

Large technology companies like Google and Microsoft are working to optimize the energy efficiency of their data centers, which house the servers used for training and executing artificial intelligence models. This includes using more efficient cooling systems, optimizing infrastructure, and adopting low-energy consumption technologies.

Some companies are developing specialized hardware for artificial intelligence, such as high-performance graphics processing units (GPUs) or artificial intelligence accelerator processing units. These devices are designed to perform AI operations more energy-efficiently compared to general-purpose computing solutions.

The most important aspect, and potentially of greater impact, is to directly intervene in the development of innovative algorithms and training paradigms that require fewer computational resources. In this regard, the most promising methodologies belong to the so-called Continual Learning. These techniques allow models to evolve over time without starting from scratch each time, but by retaining memory of what has already been learned, thus avoiding the need for lengthy retraining or maintaining overly deep data history.

In summary, awareness of the environmental issues caused by AI urges us to act responsibly, promoting the research and development of Green AI. As Luciano Floridi states, "Green and blue are two colors that will save the world." This phrase emphasizes the importance of combining the environment (represented by green) and digital technologies (represented by blue) to address the challenges of the 21st century. Green AI, which merges artificial intelligence and environmental sustainability, will be crucial for humanity. By harnessing the potential of digital technologies in a responsible and sustainable manner, we can work towards a future where AI is an ally in solving environmental problems and improving people's lives.

Only if it becomes green

Pitch Your Innovation Industry 4.0 Edition

We were in Bologna for the event "Pitch Your Innovation Industry 4.0 Edition", organized by the Startup Community of CRIF and G-Factor, the incubator-accelerator of Fondazione Golinelli.

We had the pleasure of introducing Deix S.r.l. to the attending companies and getting to know other startups that are about to take off.

Deix becomes a member of Assintel

This month Deix has become a member of Assintel, the national association for ICT and Digital companies of Confcommercio.

The association will allow us to expand our network, connect with new realities, and have our say on the role of digital companies in the Italian landscape.

Demo Day in TERNA

Today our own Jacopo Corno is at TERNAIDEASpark by Terna SpA in Turin for the Demo Day of the Data Science for Resilience call!

We have been selected as one of the top 10 finalist startups for our proposal on asset management and optimization of maintenance interventions on the Terna grid. 🏅

Our ability to combine data analysis and prediction algorithms with optimization models has been particularly appreciated, enabling us to provide prescriptive suggestions to define the best strategies. 🚀

Articles and news