About The Event
The impact of Foundation/Large Models (FM) on the Software Industry and on Software Engineering (SE) is breathtaking, especially given the rapid and impactful progress in this nascent field. FMs promise to dramatically change and improve how we develop software (FM4SE) as well as the actual definition of software and the software engineering needs for this new nascent field itself (SE4FM).
As academics and practitioners alike need to prepare for the FM revolution, it is of critical importance that we kickstart a global dialogue on important strategic future directions and challenges, on important community-wide initiatives and on the shape of Software and SE in the next few years. Such a dialogue requires close interactions between industry and academia.
Furthermore, young professionals and researchers require training in state-of-the-art and state-of-the-future techniques and insights that will guide the future of FM4SE and SE4FM.
On behalf of the organizing committee of FM Vision 2030, we announce two co-located tracks:
- The FM+SE School: The School will be held on Nov 16-17, 2023 in Mexico City. The School will provide lectures and panels by leading industrial and academic experts.
- The FM+SE Vision 2030 Meeting: The Vision Meeting will be held from Nov 16-19, 2023 in Mexico City. The Meeting includes free-form discussions by world leading industrial and academic experts from many top institutions, like Google, Meta, ServiceNow, Microsoft, University of Waterloo, Queen’s University, University of California, etc.
The key goals of FM+SE Vision 2030 Week are to provide:
- A discussion-oriented setting (similar to a Dagstuhl) for free-form, focused discussions on FM topics and the future of FMs, especially in the context of SE and software. The topics, the directions, and the meeting output will be defined by the world experts in attendance.
- A networking platform for researchers, students, leading researchers, and practitioners in the nascent field of FM4SE and SE4FM.
- A platform to kick-start global industrial and academic initiatives around FM4SE and SE4FM.
- A training avenue (FM+SE School) for key FM+SE techniques and results.
FM Vision 2030 is an in-person event, i.e., none of the breakout discussions will be recorded, nor broadcast, in order to ensure lively participation by all the attendees. Given the large number of participating companies, we remind everyone to avoid using the event to advertise their own company’s products and/or achievements, instead we encourage everyone to adopt a collaborative-attitude while ensuring that their company secrets and products are kept out of the discussions. We have a unique opportunity to define the future of this area. The organizers will eventually prepare material to share with the SE community as a whole to help define the key research agendas. We encourage other participants to work with us on these post-event reports.
The FM+SE Vision 2030 week will be held at the Hilton Mexico City Santa Fe. Vision Meeting participants will be staying in the Hilton, while School participants are expected to arrange their own accommodations. There are many international hotels around the Hilton in addition to the Hilton itself, which has ample room availability at this moment.
Unfortunately, we are not able to provide invitation letters for Mexico visas at this time. However, please check the Mexico embassy in your country as nationals from many countries might not require one. Check here for more details.
Program
INTRO
[room Nogal 1/2]
FM+SE Vision Opening
Demos
Panel A: Impact of Foundation Models on Software
- Zhen Ming (Jack) Jiang, York University
- Rob DeLine, Microsoft Research
- Charles Sutton, Google
- Ipek Ozkaya, SEI
- Prem Devanbu, UC Davis
Chair: Ahmed E. Hassan, Queen’s University
Panel B: Impact of Foundation Models on Developer Productivity
- Edward Aftandilian, GitHub
- Peggy Storey, University of Victoria
- Foutse Khomh, Polytechnique Montreal
- Hironori Washizaki, Waseda University
- Priyanshu Gupta, Microsoft Research
Chair: Tom Zimmermann, Microsoft Research
Break
BREAK-OUT SESSION 1
Formation of break-out groups, followed by break-out discussions on the following topics:
- Leveraging FMs to improve the productivity of developers (e.g., FM-powered coding and FM-powered reviewing)
[room Maple A; Tom Zimmermann/Yuan Tian] - Leveraging FMs to improve the quality of current software (e.g., FM-powered Quality Assurance like testing)
[room Maple B; Lingming Zhang/Ahmed Abdellatif] - Impact of FMs on collaboration practices in software development and software copyright/licenses for both classical
and FM-powered software (with-in
team, inner source and open source collaborations)
[room Maple C; Bram Adams/Cor-Paul Bezemer]
Lunch in lobby
BREAK-OUT SESSION 2
Formation of break-out groups, followed by break-out discussions on the following topics:
- How will FMs shape the definition of future software (e.g., how will future FM-powered software look
like)?
[room Maple A; Zhen Ming (Jack) Jiang/Shaowei Wang] - New SE challenges for the developers of FM-powered software (FM/FM-powered agents)
[room Maple B; Edward Aftandilian/Dayi Lin] - Performance optimization and DevOps for FM-powered software (debugging, monitoring, performance optimization,
release engineering)
[room Maple C; Weiyi Shang/Mohammed Sayagh]
Break
BREAK-OUT SUMMARIES 1
[room Maple A]
Presentations of each break-out group discussion of the day, followed by meta-discussions.
LIGHTNING TALKS
[room Nogal 1/2]
Reception in lobby
TODAY'S CHALLENGES of SE for FMs
[room Nogal 1/2]
Demos
Panel C: SE Challenges for Foundation Models
- Raymond Li, servicenow
- Petros Maniatis, Google
- Kensen Shi, Google
- Vincent Hellendoorn, Google
- Boyuan Chen, Huawei
Chair: Zhen Ming (Jack) Jiang, York University
Panel D: Foundation Models for Software Engineering
- Gustavo Soares, Microsoft
- Andreas Zeller, Saarland University
- Alberto Bacchelli, University of Zurich
- Jie Zhang, King's College London
- Yiling Lou, Fudan University
Chair: Bram Adams, Queen's University
Break
BREAK-OUT SESSION 3
Formation of break-out groups, followed by break-out discussions on the following topics:
- Synergizing classical SE analysis and FMs?
[room Maple 1; Andreas Zeller/Peter Chen] - FMs for beyond code generation, code reviewing and generating white-box tests (aka current popular uses)?
[room Maple 2; Diomidis Spinellis/Gustavo Soares] - FMs for non-geeks (software makers) – project managers, end users, managers
[room Maple 3; Rob Deline/Heng Li]
Lunch in lobby
BREAK-OUT SUMMARIES 2
[room Maple A]
Presentations of each break-out group discussion of the day, followed by meta-discussions.
NETWORKING EXCURSION
Meet in hotel lobby, last bus leaves by 3:45pm
Dinner
FMs in the FIELD
[room Nogal 1/2]
Demos
Panel E: Foundation Models and Open Source
- Diomidis Spinellis, Athens University of Economics and Business & Delft Technical University
- Daniel German, University of Victoria
- Marco Castelluccio, Mozilla
- Tapajit Dey, SEI
- Zhen Ming (Jack) Jiang, York University
Chair: Bram Adams, Queen's University
Panel F: Foundation Models and Software Trustworthiness
- Liming Zhu, CSIRO
- Ipek Ozkaya, SEI
- Lingming Zhang, University of Illinois
- Susumu Tokumoto, Fujitsu
- Dayi Lin, Huawei
Chair: Foutse Khomh, Polytechnique Montreal
Break
BREAK-OUT SESSION 4
Formation of break-out groups, followed by break-out discussions on the following topics:
- How can we make the life of the developers of FMs easier - what tools would a developer of a FM wish/dream to have?
[room Nogal 1/2; Vincent Hellendoorn/Boyuan Chen] - What opportunities do we have with FMs to improve the efficiency of SE researchers?
[room Maple 2; Charles Sutton/Jenny Liang] - The future IDE for AIWare - what capabilities should it have?
[room Maple 3; Peggy Storey/Dayi Lin]
Lunch in lobby
BREAK-OUT SESSION 5
Formation of break-out groups, followed by break-out discussions on the following topics:
- Beyond HumanEval - next generation benchmarks for coding assistance
[room Nogal 1/2; Raymond Li/Gopi Krishnan Rajbahadur] - How should we change our SE curriculum (needed courses/projects)?
[room Maple 2; Alberto Bacchelli/Foutse Khomh] - What are neat/promising applications that multi-agents + human-in-the-loop achieve together?
[room Maple 3; Petros Maniatis/Filipe Cogo]
Break
BREAK-OUT SUMMARIES 3
[room Nogal 1/2]
Presentations of each break-out group discussion of the day, followed by meta-discussions.
Dinner
BREAK-OUT SESSION 6
Selection of discussion topics from the shared spreadsheet, followed by formation of break-out groups and break-out discussions
Break
BREAK-OUT SUMMARIES 4 and CLOSING
[room Nogal 1/2]
Presentations of each break-out group discussion of the day, followed by closing discussions of the event.
Lunch in lobby
Event Venue
Hilton Mexico City Santa Fe
Antonio Dovali Jaime 70, Santa Fe,
Zedec Sta Fé, Álvaro Obregón,
01219 Ciudad de México,
CDMX, Mexico
website
Supporters
(pending)
Contact Us
If you have any questions or concerns, do not hesitate to contact us. We hope you can join us in Mexico this November to plan the FM future of the Software Industry and Software Engineering!
Ahmed E. Hassan, Fellow of
ACM/IEEE/NSERC Steacie, Queen’s University
Bram Adams, IEEE Senior Member, Queen’s
University
Foutse Khomh, IEEE Senior Member, Canada CIFAR AI Chair and
IVADO Research Chair, Polytechnique Montreal
Nachi Nagappan, Fellow of ACM/IEEE, Meta
Thomas Zimmermann, Fellow of ACM/IEEE, Microsoft
Research