معرفی
Marco Braga is a researcher focused on information retrieval, large language models, and parameter-efficient fine-tuning techniques. His work spans personalized community question answering systems, synthetic data generation, and adapter-based model optimization.
Recent research interests include:
- Zero-shot learning and task arithmetic for LLMs
- Context-aware personalization in retrieval systems
- Hybrid neural-symbolic approaches for structured reasoning
- Telecommunications domain modeling
Key article trends show expertise in:
- Adapter modules and low-rank parameter optimization
- Community question answering personalization
- Synthetic dataset creation for training LLMs
- Model efficiency in resource-constrained settings
Collaborations include Gabriella Pasi, Pranav Kasela, and Alessandro Raganato. His work appears in SIGIR, COLING, WI-IAT, and IIR conferences, with notable Zenodo dataset contributions.
۰مقاله منتشرشده
