I’m a PhD student in Artificial Intelligence at MICC, University of Florence, working under the guidance of Prof. Andrew D. Bagdanov and Prof. Marco Bertini. With a background in Computer Engineering and AI, my research focuses on pushing the boundaries of Multimodal Vision-Language Models (like CLIP) and their real-world applications.

This expertise is demonstrated through my first-author publications in top-tier venues, including ECCV (main conference), NeurIPS (workshop), ICLR (main conference). These works reflect my dedication to solving challenging problems and advancing the field of AI.

Currently seeking a 2025 internship to contribute to innovative teams and apply my expertise to real-world challenges! For more information, feel free to visit my website: marcomistretta.github.io

• Cross the Gap: Inter-modal CLIP Representations Are Superior for Intra-modal Tasks
  ICLR 2025 (main paper)
  Authors: Marco Mistretta*, Alberto Baldrati*, Lorenzo Agnolucci*, Marco Bertini, Andrew D. Bagdanov

• Improving Zero-shot Generalization of Learned Prompts via Unsupervised Knowledge Distillation
  ECCV 2024 (main paper)
  Authors: Marco Mistretta*, Alberto Baldrati*, Marco Bertini, Andrew D. Bagdanov
  Code: GitHub Repository

• RE-tune: Incremental Fine Tuning of Biomedical Vision-Language Models for Multi-label Chest X-ray Classification
  NeurIPS 2023, Medical Imaging meets NeurIPS Workshop
  Authors: Marco Mistretta, Andrew D. Bagdanov

I'm really into:

• 🧠 Multimodal Learning: Combining visual and language data to get a richer understanding of the world.
• 💬 Natural Language Processing (NLP): Teaching machines to understand and communicate in human language.
• 🖼️ Contrastive Self-Supervised Learning: Finding patterns in data without the need for human labels.
• ♻️ Incremental Learning: Allowing AI models to keep learning from new information without forgetting the old ones.
• 🎯 Few-Shot Adaptation: Quickly adapting AI to a diverse data distribution with minimal examples.
• 📝 Prompt Learning: Tuning only a few learnable parameters, so-called "prompts", to maximize VLMs performance.
• 🚀 Test-Time Adaptation: Letting models adjust during inference to handle unseen data on the fly.

• Programming Languages: Python, Java, C++, MATLAB, R
• Frameworks & Tools: PyTorch, TensorFlow, Hugging Face, OpenCV
• Research Areas: Vision-Language Models, Self-Supervised Learning, Few-Shot Learning, Prompt Learning, Incremental Learning

I’d love to connect! Feel free to reach out on: