AI-Driven Innovations: LLMs in Lifelong and Self-Supervised Learning

Abstract

This special session will bring together cutting-edge research on two rapidly advancing areas within AI: the integration of Large Language Models (LLMs) in lifelong learning environments, and the application of self-supervised learning and object detection techniques to automated sorting and intelligent manufacturing systems.

The growing capabilities of LLMs are revolutionizing the way humans interact with educational systems. These models enable adaptive, personalized, and continuous learning across disciplines, aligning with the global vision for inclusive and scalable lifelong learning. At the same time, self-supervised learning techniques are pushing the boundaries of industrial automation by significantly reducing dependency on labeled data. Combined with robust object detection methods, these advances hold enormous potential for deployment in complex manufacturing workflows, including dynamic sorting, anomaly detection, and autonomous decision-making.


This session focuses on practical and theoretical advances in applying machine learning, deep learning and computer vision techniques to real world image analysis tasks. Innovative methodologies, deployment strategies and current challenges to image analysis are topics of
interest in this special session. Emphasis is placed on approaches that demonstrate robustness, explainability and effectiveness in real scenarios in which images with high resolution are needed. Relevant topics include the entire pipeline of intelligent image processing, from data acquisition and annotation to model training, validation, deployment and post-hoc interpretation.


This session provides a chance to present both theoretical and applied research that extends the boundaries of what intelligent systems can achieve with visual data in several areas of knowledge such as healthcare, agriculture and manufacturing.

Topics

The session will emphasize:

  • Theoretical and applied advancements in LLMs for education and training
  • Self-supervised and contrastive learning for object recognition
  • Multi-modal AI systems in real-world industrial applications
  • Bridging LLMs with visual perception systems
  • Deployment challenges and case studies in manufacturing

Deep learning architectures for image segmentation, classification and detection
● Explainability and trustworthiness in image-based AI systems
● Analysis of images with high resolution for medical applications
● Analysis of images for precision agriculture
● Analysis of images for manufacturing processes
● Progressive learning
● Multimodal image analysis
● Generative AI in image synthesis and augmentation
● Edge AI and real-time image processing

This session aligns with the IDEAL conference’s focus on intelligent data engineering, machine learning, and their real-world applications. By fostering interdisciplinary discussion, it offers a platform for researchers and practitioners to explore how foundational AI technologies can be harnessed for both human-centered learning and industry 4.0 solutions, addressing both societal and economic needs.

This session is sponsored by the Horizon Europe iBot4CRMs project

Organizers

Dr. Youcef Djenouri, Norwegian Research Center, Norway (yodj@norceresearch.no)
Prof. Gautam Srivastava, Brandon University, Canada
Dr. Tomasz Michalak, Warsaw University, Poland

Submission

See submission instructions for the conference at the call for papers. At the beginning of the submission, please choose the track “AI-Driven Innovations: LLMs in Lifelong and Self-Supervised Learning”.

Submission link: https://easychair.org/conferences/?conf=ideal2025

Special Session Papers Submission Deadline: June 15, 2025 July 15, 2025