Abstract:While AI is extensively transforming Software Engineering (SE) fields, SE is still in need of a framework to overall consider all phases to facilitate Automated Software Evolution (ASEv), particularly for intelligent applications that are context-rich, instead of conquering each division independently. Its complexity comes from the intricacy of the intelligent applications, the heterogeneity of the data sources, and the constant changes in the context. This study proposes a conceptual framework for achieving automated software evolution, emphasizing the importance of multimodality learning. A Selective Sequential Scope Model (3S) model is developed based on the conceptual framework, and it can be used to categorize existing and future research when it covers different SE phases and multimodal learning tasks. This research is a preliminary step toward the blueprint of a higher-level ASEv. The proposed conceptual framework can act as a practical guideline for practitioners to prepare themselves for diving into this area. Although the study is about intelligent applications, the framework and analysis methods may be adapted for other types of software as AI brings more intelligence into their life cycles.

What problem does this paper attempt to address?

The problem that this paper attempts to solve is: **How to provide an Automated Software Evolution (ASEv) framework that can cover all stages of software engineering for Context - Rich Intelligent Applications (CRIA) to address the limitations of current independent handling at each stage**. Specifically, the paper focuses on the following aspects: 1. **Complexity and Diversity**: The complexity of intelligent applications, the heterogeneity of data sources, and the ever - changing context make it difficult for traditional software engineering methods to respond effectively. 2. **Importance of Multimodal Learning**: Intelligent applications usually need to learn from multiple types of data, such as text, image, audio, etc. Therefore, multimodal learning is crucial in automated software evolution. 3. **Limitations of Existing Research**: Most current research focuses on a specific stage or task in software engineering and lacks a comprehensive framework to guide the automation of the entire software life cycle. To solve these problems, the paper proposes a **Multimodal Conceptual Framework** and develops a Selective Sequential Scope Model (3S) based on this framework. This framework and model aim to: - **Integrate Multimodal Learning**: By combining data from different sources and formats, improve the understanding and decision - making ability of intelligent applications. - **Cover All Stages of Software Engineering**: Including requirements analysis, design, development, testing, maintenance, etc., to ensure that each stage can benefit from automation. - **Adapt to Context Changes**: Automatically generate new requirements and carry out related design and development not only at the traditional situation adaptation (such as decision - making) level, but also at the system evolution level. Finally, this research provides a practical guide for researchers and practitioners to help them better understand and implement the automated software evolution of context - rich intelligent applications. ### Formulas Involved This article does not involve specific mathematical formulas, but mentions some techniques related to machine learning and multimodal learning, such as: - **Deep Neural Network (DNN)** - **Transformer Architectures** - **Knowledge Graph - enhanced Large - scale Language Model (KG - enhanced LLM)** The specific formulas of these techniques can be found in relevant literature.

A Data-to-Product Multimodal Conceptual Framework to Achieve Automated Software Evolution for Context-rich Intelligent Applications

A Survey And Extension On Software Evolution And Computer-Aided Prototyping

An Ide For Evolutionary Software Development

Towards AI-Native Software Engineering (SE 3.0): A Vision and a Challenge Roadmap

Research on the technology of creative product evolutionary design based on semantics feature

Self-Organized Semantic Feature Evolution for Axiomatic Design

Research on Evolutionary Design Technology for Creative Product on Integrated Platform

A Framework to Study Human-AI Collaborative Design Space Exploration

Towards Enhanced Context Awareness with Vision-based Multimodal Interfaces

Supporting context — Aware service evolution with a process management requirements model

SSE: Multimodal Semantic Data Selection and Enrichment for Industrial-scale Data Assimilation

CSM-H-R: A Context Modeling Framework in Supporting Reasoning Automation for Interoperable Intelligent Systems and Privacy Protection

An Enhanced Ontology Based Context Model and Fusion Mechanism.

Rethinking Software Engineering in the Foundation Model Era: From Task-Driven AI Copilots to Goal-Driven AI Pair Programmers

Evolving Testing Scenario Generation Method and Intelligence Evaluation Framework for Automated Vehicles

High-confidence Software Evolution

A conceptual framework for context-driven self-adaptive intelligent user interface based on Android

A Service-oriented Framework for Context-aware Semantic Media Applications

Building Smart Systems to Satisfice User Intention in Situ: A Look Back to the Future*

Towards a Programming Framework for Activity-Oriented Context-Aware Applications.

Bridging Multimedia Modalities: Enhanced Multimodal AI Understanding and Intelligent Agents