Hassan Alsayhah

تخصص المعلوماتية الحيوية · دفعة 2024

GANLA: GPU Accelerated Network Layout Algorithm

Network layout algorithms are used to visualize networks in a human friendly manner and are pivotal for data exploration. Currently, the Kamada-Kawai algorithm is commonly used for network visualization. However, Kamada-Kawai does not scale to large networks, such as those derived from human functional genomics data. We have developed a fast, scalable version of Kamada-Kawai which takes advantage of graphics processing unit (GPU) hardware acceleration, called GPU Accelerated Network Layout Algorithm (GANLA). GANLA computes the layout positions for all nodes in the network in parallel, whereas Kamada-Kawai computes these positions sequentially. Moreover, GANLA utilizes a mini-batch technique which enables the scalable layout of networks with many nodes, with a fixed memory footprint. In conclusion, GANLA scales to larger networks and should be faster than Kamada-Kawai algorithm, enabling the exploitation of larger and more complex biological networks relevant to human functional genomics.

اقرأ التقرير النهائي (PDF)

اطّلع عليه في تقرير البرنامج (PDF)

تخصص الذكاء الاصطناعي · دفعة 2023

Cultural Enrichment: Through AI-Generated Arabic Comics

This study introduces an innovative approach to crafting captivating Arabic stories for children, merging the rich tradition of storytelling with cutting-edge AI technologies to create engaging comics. Our primary emphasis lies in a finely tuned main model for text generation, complemented by a range of specialized models that amplify the storytelling experience. This comprehensive framework adeptly navigates the intricacies of Arabic language processing, integrating dedicated models to generate compelling titles, vibrant images, voice-over and even voice narration. This approach holds great promise in delivering culturally immersive narratives that not only captivate the attention of young readers but also provide an exhilarating multi-sensory storytelling adventure through the world of comics.

اطّلع عليه في تقرير البرنامج (PDF)