The broad influence of Artificial Intelligence (AI) and Machine Learning (ML) has led to a massive new software paradigm of Intelligent systems. The recent report of JST[1] forecasts a 60 trillion JPY market from AI applications in industry, in particular in domains such as transportation, healthcare, finance, etc., among which many sectors and applications require quality and reliability of the intelligent systems to be medium to high level. For those systems, either important parts of the system are based on AI or generated by AI (i.e., AIWare). Among the various key attributes of research and industry pursuit of governance and regulation of intelligent systems, sustainability is becoming a critical factor. In particular, the ever-growing energy consumption and carbon emissions from AI-based software play a central role in the sustainability of intelligent systems, from large-scale data centers to personal devices. Thus, improving the energy efficiency of the software is critical for reducing the energy footprint of intelligent systems and contributing to sustainable development.

The challenge of ensuring the energy efficiency of software in the era of AI is twofold. First, despite the rich knowledge of performance and energy research on traditional software, as a new paradigm of software, AI-based software, like ChatGPT, may not fit into the well-adopted engineering processes. The best practices distilled through decades of experience may not apply. Second, AI-generated software itself may not have the consideration of efficiency, leading to suboptimal software that may waste energy during runtime. The proposed research toward more energy-efficient AIWare will contribute theories, methods, and tools to support the development, testing, and auditing of energy efficiency on AIWare. The outcome not only benefits the software industry by providing more cost-effective AI solutions across industries but also protects the environment by reducing global electricity demand and CO2 emissions.

The connection between development and operations, i.e., DevOps, has become the main principle for developing and operating modern software systems, which aims to enhance the velocity of releasing high-quality software through automation and integrating development and operations responsibilities. On the one hand, the fast DevOps release cycles pose significant challenges for ensuring the energy efficiency of intelligent software and its deployment environment. On the other hand, DevOps provides great opportunities for enhancing intelligent software sustainability, since the integration between development and operations allows for continuous collection and analytics of energy consumption data, which can be leveraged as feedback for optimizing the energy efficiency of software and its deployment environment.

The goal of the meeting is to bring together software engineering, performance and AI experts from academia and industry, featuring and taking a special focus on the sustainability of intelligent systems, to discuss how to design new engineering practices, especially for intelligent systems that would allow to effectively engineer an AI-based system in a more energy efficient way. The long-term goal is to make the activities across the whole development lifecycle of intelligent systems sustainable (Green Intelligent Systems Engineering) as it is nowadays for traditional software. The community of AIWare and green software has fastly grown in the past few years with some early-stage results by researchers around the world. We envision a meeting that gathers the world-leading researchers and industry practitioners who have achieved important results over the past few years. The meeting will enable further exchange of state-of-the-art and -practice ideas and techniques, promote sustainability as an important research direction, and emphasize its potential industrial applications, therefore contributing to the currently urgent demand of sustainable intelligent systems.

The meeting will have two types of sessions:

• The first type of session will have short presentations where each participant will introduce their research work, and list some ideas/challenges/research directions that they would like to discuss during the meeting. Such presentations will be scheduled during the first two days of the meeting.

• The second type of session will consist of intensive discussions among sub-groups of participants. The topics of the discussions will be decided at the meeting among those proposed by participants during their short talks; still, organizers will prepare beforehand some possible topics of discussion. The discussions will proceed in two phases, by first exploring different topics, and then deepening on a restricted set of selected topics.

  • [Exploration phase]

    In the first instance of this type of session, the meeting will use a dedicated method to guide the discussion. The current plan is to use a variation of the world café method (http://www.theworldcafe.com/method.html): this will allow participants to move across the different sub-groups and so get to know the different initial topics.

  • Afterwards, each participant will decide which topic they would like to explore. At this stage, some topics that did not get too much interest will be dropped.

  • [Deepening phase] The remaining sessions will be dedicated to intensive discussion on the selected topics. Ultimately, each group should come up with a research agenda for its topic.

  • Afterwards, we are planning to sort out a research agenda for engineering sustainable intelligent software systems, and plan to make a book proposal through the agenda that is discussed and agreed upon by the participants. We will apply to the “Call for Book Proposals” provided by the Shonan meeting organization.

[1]: https://www.jst.go.jp/crds/pdf/2018/SP/CRDS-FY2018-SP-03.pdf

Sunday

19:00- Welcome reception

Monday

9:00-9:15 Opening

9:15-10:30 Introduction. 2 mins for each participant

10:30-11:00 Break

11:00-12:00 Presentations & Discussion (6 presentations)

12:00-13:30 Lunch

13:30-13:40 Group Photo Shooting

13:40-15:30 Presentations & Discussion (11 presentations)

15:30-16:00 Break

16:00-17:30 Presentations & Discussion (9 presentations)

17:30-17:45 Wrap-up of first day

18:00- Dinner

Night session

Tuesday

8:55-9:00 Opening

9:00-10:30 Presentations & Discussion (9 presentations)

10:30-11:00 Break

11:00-12:00 Definition of an initial set of topics for the discussion in the sub-groups

12:00-13:30 Lunch

13:30-15:00 Discussion of the initial selected topics in sub-groups (exploration phase)

15:30-16:00 Break

16:00-17:15 Discussion of the initial selected topics in sub-groups (exploration phase)

17:15-17:45 Summary of the discussions of the different sub-groups and selection of a restricted set of topics to be discussed in the deepening phase. Each participant is allocated to a specific topic

18:00- Dinner

Night session

Wednesday

9:00-10:30 Discussion of the selected topics in sub-groups (deepening phase)

10:30-11:00 Break

11:00-12:00 Discussion of the selected topics in sub-groups (deepening phase)

12:00-13:30 Lunch

13:30- Excursion (including Dinner)

Thursday

9:00-10:30 Discussion of the selected topics in sub-groups (deepening phase)

10:30-11:00 Break

11:00-12:00 Presentations of the work of the individual sub-groups. Final planning of future work and wrap-up

12:00-13:30 Lunch