INFORMATION SCIENCE · AI · MULTIMODAL RETRIEVAL

Academic work and research

Reza Javidi’s research connects information science, artificial intelligence, artistic experience and practical information problems.

ACADEMIC STATUS

PhD student in Information Science and Knowledge Studies

University of Tehran · Focused on information retrieval, knowledge organization and applications of artificial intelligence in specialist settings.

University of Tehran website ↗

Research Lens

From image and text to meaning, time and context: the central question is how to find meaningful similarity in complex data.

Research areas

Information retrieval

Neural retrieval, similarity modelling, semantic search and the design of responsive information systems.

Multimodal retrieval

Combining images, text and contextual signals to retrieve results beyond visual similarity alone.

Language and vision-language models

LLMs and VLMs, embeddings, transformers, and their use in knowledge organization and representation.

Medical information

Applications of artificial intelligence in specialist settings and the relationship between imaging data and clinical text.

Retrieval of similar clinical cases

A central focus of his doctoral research is retrieving similar clinical cases using medical images and clinical text together. The working premise is that clinical similarity is not necessarily the same as visual similarity: a system needs to combine visual, textual and temporal evidence within one model.

This perspective connects experience with image and performance to questions of meaning, context and memory in information systems.

Earlier research

The influence of online environments on students’ searches for scientific information, the effects of media on professional audiences, and the impact of knowledge management on organizational productivity.

Reference project

Developing a Comprehensive Dictionary of Information Science covering information retrieval, the semantic web, data, machine learning, neural networks and large language models.