interactive learning

Practitioners face significant challenges in effectively transforming customer feedback data into actionable software improvements. Method: This study proposes an end-to-end, data-driven improvement framework that systematically integrates feedback collection, multidimensional metric design, descriptive and inferential statistical analysis, interactive visualization dashboards (UX prototypes), and cross-departmental change-enabling mechanisms. Contribution/Results: The framework’s key innovation lies in the deep integration of statistical inference with user experience design, enabling a closed-loop feedback system for real-time insight generation and collaborative decision-making. Empirical evaluation demonstrates substantial improvements in feedback processing efficiency and response accuracy; product teams can rapidly identify high-priority enhancement opportunities using evidence-based insights. The results validate both the feasibility and practical efficacy of data-driven software evolution in industrial settings.

Large global software enterprises (e.g., Dell Technologies) face persistent challenges in continuously integrating customer feedback within Agile and DevOps practices, resulting in requirement misalignment and delivery deviations. To address this, this study proposes and empirically validates a customer feedback closed-loop integration model grounded in action research. For the first time, the model systematically connects demand validation, feedback analysis, and release decision-making across an organization exceeding 10,000 employees, integrating Agile/DevOps principles with multidimensional metrics—including defect escape rate, deployment frequency, product adoption rate, and customer satisfaction (CSAT). Empirical evaluation demonstrates a 40% reduction in customer feedback response cycle time, a 35% increase in critical product deployment frequency, and statistically significant improvements in defect interception rate and CSAT. This work contributes a scalable, reusable methodology and empirical evidence for customer-centric continuous delivery in large-scale software organizations.

This study investigates how human-in-the-loop (HITL) feedback influences users’ perceptions of system accuracy and trust, highlighting the critical moderating role of task subjectivity. Through three controlled user experiments that systematically differentiate between objective and subjective task contexts, the research analyzes behavioral measures to assess the effects of feedback interaction. Findings reveal that in objective tasks, providing feedback significantly diminishes users’ trust in and perceived accuracy of the system, whereas this negative effect vanishes in subjective tasks. These results underscore task type as a pivotal factor shaping human–AI trust dynamics and offer important theoretical grounding and practical guidance for the design of HITL systems.

This work proposes a context-aware, interactive feedback approach to address the ambiguity and inefficiency of mobile application user feedback caused by insufficient contextual information, which often increases developers’ clarification costs. By leveraging multimodal large language models (MLLMs), the method dynamically generates adaptive follow-up questions in real time based on contextual cues—such as user-submitted screenshots—to collaboratively construct structured, high-quality bug reports or feature requests. Integrated into an iOS framework and evaluated within a real-world fitness application, the approach enables users to provide feedback more effortlessly and effectively. Expert assessment of 54 generated reports demonstrates significantly higher completeness compared to traditional form-based submissions, particularly in defect descriptions and feature requests.

This work addresses the insufficient collaboration and cumulative bias arising from disjoint modeling of recommender and user agents in recommendation systems. We propose the first collaborative optimization framework explicitly designed for a dual-agent closed-loop feedback paradigm. Methodologically, we establish a bidirectional iterative feedback mechanism: the recommender agent generates recommendations and observes responses from the user agent, while the user agent dynamically refines its preference representation based on feedback; both agents co-evolve via an LLM-driven, interpretable interaction protocol. Our key contribution is the first formalization of the recommender–user dual-agent closed-loop feedback process, jointly optimizing recommendation quality and mitigating bias—without exacerbating popularity or positional biases. On three benchmark datasets, our approach achieves average improvements of 11.52% in recommendation accuracy over a recommender-only baseline and 21.12% over a user-only baseline, significantly enhancing fidelity in user behavior modeling.

Existing approaches to requirement prioritization often overlook the semantic interdependencies among requirements, thereby compromising prioritization effectiveness. This work addresses this limitation by introducing requirement interconnectedness into user feedback–driven prioritization for the first time, proposing a dependency-aware search-based optimization framework. The method first applies natural language processing to cluster app store feedback into semantically coherent requirement groups and then automatically infers “requires”-type dependencies among these groups. These dependencies are explicitly integrated into a search algorithm to guide the optimization of requirement priorities. Evaluated on 94 real-world instances across four software systems, the proposed approach significantly outperforms ReFeed, demonstrating that explicitly modeling requirement interconnections effectively enhances both prioritization accuracy and release planning quality.

This work proposes a localized multi-agent feedback system based on small open-source large language models (LLMs), addressing the limitations of conventional feedback systems that treat feedback as static and unidirectional, thereby failing to support personalized, interactive formative guidance. The system uniquely implements fully interactive feedback through a multi-agent architecture featuring pedagogically oriented feedback generation, an LLM-as-a-judge regeneration mechanism guided by human-aligned criteria, and a context-aware reflexive tool-calling agent enabling students to ask follow-up questions and engage in dynamic dialogue. Experimental results demonstrate that criterion-guided regeneration significantly enhances feedback quality, and the interactive agents achieve performance comparable to state-of-the-art closed-source models in both efficiency and quality. Deployment in real classroom settings further validates the system’s effectiveness in guiding student inquiry.

This work addresses the lack of evaluation frameworks for assessing how retrieval-augmented generation (RAG) systems adapt following user or expert feedback. It introduces, for the first time, a “feedback adaptation” problem setting, quantifying adaptation speed and reliability through two metrics: correction latency and post-feedback performance. To enable real-time feedback integration without retraining during inference, the authors propose PatchRAG, which combines semantic relevance analysis with behavioral change detection to achieve zero-latency corrections and cross-query semantic generalization. Experimental results demonstrate that PatchRAG significantly outperforms baseline methods, maintaining immediate responsiveness while exhibiting strong generalization capabilities after receiving feedback.

This work proposes TRAFA, a novel interactive assistance system that introduces predictive feedback into procedural task guidance—addressing the limitation of traditional systems that only provide reactive corrections after errors occur. TRAFA implements a Track-Forecast-Act framework: it continuously tracks hand and object states in real time, leverages scene context to forecast user intent, and proactively triggers interventions before task constraints are violated. By shifting from post-hoc recovery to preemptive error prevention, TRAFA enhances both accuracy and efficiency. Experimental results in an assembly task demonstrate that TRAFA significantly improves task performance while maintaining a feedback frequency comparable to conventional reactive systems.

This work proposes a human-feedback-driven dual-RAG architecture designed to continuously enhance the accuracy, relevance, and overall quality of Retrieval-Augmented Generation (RAG) systems through a human-in-the-loop mechanism. The approach introduces an auxiliary feedback RAG module that automatically collects, categorizes, and integrates user feedback into the primary RAG inference pipeline, enabling autonomous iterative refinement without requiring explicit supervisory signals. Leveraging an LLM-as-a-Judge evaluation strategy, the method demonstrates significant improvements in response quality across three benchmark datasets encompassing both general and domain-specific knowledge, thereby advancing RAG systems toward self-optimizing capabilities.