Eden Moran, M.Sc. Intelligent Systems

The project develops a system for implicit entity recognition in text, identifying people, organizations, locations, and other entities that are referred to indirectly rather than named explicitly. By combining contextual reasoning, semantic inference, and discourse-level cues, the model infers hidden or implied entities from descriptions, roles, actions, or relationships. This enables deeper text understanding in domains where entities are frequently mentioned obliquely, improving downstream tasks such as information extraction, summarization, and knowledge base construction. 

Omri Sasson, M.Sc. Intelligent Systems

The project develops a framework for generating imperfect student simulations using LLMs, creating virtual learners that exhibit realistic misunderstandings, partial knowledge, inconsistent reasoning, and common conceptual errors. By modeling not only correct responses but also typical mistakes and learning gaps, the system produces richer training data for tutoring platforms, assessment tools, and classroom-support systems. This enables more robust evaluation of educational technologies and provides instructors with realistic scenarios for testing feedback strategies and adaptive teaching methods. 

Niv Cohen, M.Sc. Intelligent Systems

Diagnostic expertise depends on asking the question that most reduces uncertainty given incomplete patient information. This project develops a Large Language Model that learns to generate high-value diagnostic questions by modeling questioning as a sequential information acquisition process. Trained on clinical records, dialogues, and simulated patient interactions, the system estimates the expected diagnostic gain of candidate questions and selects the optimal next query. The objective is to achieve clinician-level efficiency by reaching accurate diagnoses with fewer, more targeted questions.

Yigal Meshulam, M.Sc. Intelligent Systems

The project develops a decentralized multi-robot task-allocation framework operating under a zero-knowledge assumption, in which robots have no prior information about task characteristics or peer capabilities. Instead, each robot independently selects tasks and learns solely by observing its own outcomes. Using a collaborative-filtering-inspired approach, robots share minimal performance signals to infer which task–robot pairings are most effective, gradually converging toward efficient global allocation without central coordination or explicit communication of internal models. This results in scalable, resilient task assignment suitable for environments with uncertainty and dynamism. 

Matan Chazanovitz, M.Sc. Intelligent Systems

This project develops an adaptive super-resolution approach that selectively enhances images, focusing only on regions that exhibit noticeable quality degradation while preserving areas that are already clear. Instead of applying a uniform upscaling across the entire image, the method dynamically adjusts the restoration intensity based on local image conditions. This selective strategy improves overall visual fidelity, prevents unnecessary artifacts in clean regions, and enables more efficient use of computational resources.