Files
logseq/pages/SWEN Retreat 2025.md
T

4.3 KiB

tags:: #todoist-task, SWEN, meeting date:: 13-07-2025 - 16:57 link:: https://univaq-my.sharepoint.com/:w:/g/personal/juri_dirocco_univaq_it/EQ-K23f6XrJPvu0NMEBJbwMB2mOKj0PTLNoPBrgOsUEeGQ?e=fSHI0E progress:: {{renderer :todomaster}}

- ### Tasks
	-
- ### Notes
	- ## First day [[15-07-2025]]
	  collapsed:: true
		- ### Points for discussion
			- Exciting and scaring times
			- GenAI / Agentic systems:
				- Hype? What will remain?
			- Vision of what we are doing and what the others are doing
			- It's not a matter of visibility!
			- Understand what it is needed and how to cooperate
			- Can we come up with some cross-group collaboration
			- [SwissLLM - Generative-KI-Lösungen Made in Switzerland](https://www.swissllm.ch/)
			- **Simplify** - [Progetti Finanziati - Apprendo](https://apprendo.tech/progetti-finanziati/)
			- Paper on object-oriented programming published on 1984 that seems to be similar to what is happening nowadays with generative AI
			- ### Topics
				- Generative AI in software development lifecycle
					- Tools that provide users/delelopers with additional automation capabilities that need to be mastered. Many times we read that we should not accept AI as it i but somehow we have to master it
					- I'm convinced that such tools represent on the one hand a multiplier of our capabilities and on the other hand the subject of investigations
					- What's the impact of this technology for our bachelor and master programs
				- Ethical Aspects
					- Can be addressed in different ways depending on the application domains.
				- Empirical Studies
					- Empirical studies vs the empirical validation of the approach you are proposing
						- I can agree that requests on the empirical validation is getting tremendously on the one hand critical and on the other hand very demanding
					- The availability of data like GitHub etc. initiated the field of mining software repositories field, which is similar what happens since decades in other discplines (like biology, etc.)
				- Software sustainability
					- What's the meaning?
				- More
					- How do we use LLMs?
						- for our daily tasks
						- if we use it for software development
					- Who should develop tools including LLMs?
					- Training and fine-tuning existing models
						- Discrimination of universities depending on their possibility of training/fine-tuning
				- Alfonso was suggesting to use Wikidata for generating domain models from textual descriptions
	- ## Second day [[16-07-2025]]
		- Topics of the groups
			- Green SW Engineering
				- Sustainabiolity
			- Simlation in federated digital twins
			- Cultural Heritage
				- RecSys
			- Multi-view Modeling
			- Reuse of LLM Agents for Software Engineering
			- Performance Analysis
				- Model
				- Code
				- NF Analysis for SW
				- AI Ops
				- Code execution and LLMs
			- Ethic for SW
			- MDE techniques + JJodel
		- CONCEPTS
			- System views
				- Pre deployment
					- Statics
						- Forward vs reverse engineering
					- Dynamics
						- Forward vs reverse engineering
					- Analysis
						- Predictive
						- Prescriptive
					- Simulation
						- It requires semantics
				- Post deployment
					- System Monitoring
					- System Adaptation
			- Supporting tools
				- Modeling environments
				- LLMs
					- To move from statics to dynamics
					- To support deployment (e.g., reconstruct the microservice architecture)
					- Configurations
						- Single agent
						- MAS workflows
							- e.g, design patterns for software architecture perspectives
			- WPs
				- **WP1: Pre-deployment design and anlysis of software systems**
					- T1.1: Multi-dimensional non-functional analysis
					- T1.2: Multi-dimensional software architectures (e.s. green) (deps: T1.1)
					- T1.3: Code execution and LLMs (deps: T3.1)
				- **WP2: Post-deployment system monitoring and adaptations**
					- T2.1: Reconstruction of microservice architectures
					- T2.2: System monitoring
					- T2.3: System adaptation
				- **WP3: Modeling languages and supporting tools**
					- T3:1 Reuse of LLM-based agents
					- T3.2 JJodel
				- **WP4: Fairness and Ethics of systems**
					-
				- **WP5: Impact**
					- Use case scenarios (Cultural heritage, Territorial monitoring)
					- Education
				-
				-
		-