Moussa Doumbouya

Title: The Innovative Mind of Moussa Doumbouya

Introduction

Moussa Doumbouya is a prominent inventor based in Santa Clara, CA. He has made significant contributions to the field of deep learning and artificial intelligence. His work focuses on enhancing the interpretability of deep neural networks, which is crucial for advancing AI technologies.

Latest Patents

Moussa Doumbouya holds a patent titled "Semantic coherence analysis of deep neural networks." This patent discloses methods and apparatus for interpreting a deep neural network (DNN) using a Semantic Coherence Analysis (SCA)-based interpretation technique. The invention analyzes a multi-layered DNN trained for one task to select a layer that produces salient features for another task. The DNN layers are tested with samples labeled with a set of concept labels, and the output features are gathered and analyzed according to these concepts. The output is scored with a semantic coherence score, indicating how well the layer separates the concepts. A layer is selected based on its semantic coherence score, and in some embodiments, a support vector machine (SVM) or additional neural network may be added to the selected layer for generating classification results.

Career Highlights

Moussa Doumbouya is currently employed at Apple Inc., where he applies his expertise in deep learning and neural networks. His innovative approach to semantic coherence analysis has the potential to revolutionize how deep learning models are interpreted and utilized in various applications.

Collaborations

Moussa has collaborated with notable colleagues, including Xavier Suau Cuadros and Luca Zappella. Their combined efforts contribute to the advancement of technology in the field of artificial intelligence.

Conclusion

Moussa Doumbouya's work exemplifies the intersection of innovation and technology in the realm of deep learning. His contributions are paving the way for more interpretable and effective AI systems.