Ofir Mulla

Title: Ofir Mulla: Inventor in Visual Classification and Multimedia Intelligence

Introduction

Ofir Mulla is an inventor based in Talme Elazar, Israel. His patent work includes machine-learning techniques for classifying object states in visual content and systems for translating natural-language requests into structured database queries. He is associated with Lumana Inc. and holds 3 patents.

Latest Patents

One of Mulla’s latest patents, “Techniques for Classifying Custom Object States in Visual Content,” describes a system and method for training a classifier. A first machine-learning model identifies instances of an object in visual content. Selected examples are then labeled according to the states shown, and a second model is trained to classify those states. This approach builds on object detection to distinguish different conditions or states of an object.

Another patent, “Multimedia Intelligence Agent,” concerns structured querying. The method embeds a natural-language request and compares it with examples that pair natural-language text with database-formatted queries. It uses matching examples to help generate a query, consults a language model, and then queries the database using a structured query based on the model’s outputs.

Career Highlights

Mulla’s disclosed inventions address two practical challenges: interpreting visual information at a more detailed level and making databases accessible through natural-language requests. His work combines machine learning, visual-content analysis, language models, and database query generation.

Collaborations

Mulla is associated with Lumana Inc. His listed coworkers include Aviad Zabatani and Ron Kimmel.

Conclusion

Ofir Mulla’s patent portfolio reflects an inventive focus on applying machine learning and language technologies to visual classification and structured information retrieval.