This article discusses retrieval augmentation in text generation, where information retrieved from an external source is used to condition predictions. It references recent work on retrieval-augmented image captioning, showing that model size can be greatly reduced when training data is available through retrieval. The author intends to continue this work focusing on the intersection of retrieval augmentation and in-context learning, and controllable image captioning for language learning materials. Why it matters: This research direction has the potential to improve transfer learning in vision-language models, which could be especially relevant for downstream applications in Arabic NLP and multimodal tasks.
Dr. David Edwards from Harvard University spoke at KAUST about creativity in innovative communities. He believes that we are at the dawn of a grassroots renaissance in the arts, sciences and engineering. Edwards highlighted the importance of learning, experimentation, and production centers in fostering innovation. Why it matters: This talk suggests KAUST is looking to foster a cross-disciplinary culture of innovation, aligning with broader trends in AI and technology development that require diverse skill sets.
This article summarizes a talk by Erci Xu on doing computer systems research, focusing on idea generation and paper writing. Xu shares experiences on developing research ideas and provides a tutorial on academic writing principles. He has published 20 papers in venues like OSDI, FAST, ATC, and Eurosys and received awards including two FAST Best Paper Awards. Why it matters: The talk and summary offer valuable guidance for researchers in the Middle East, particularly those at institutions like MBZUAI, on how to conduct impactful computer systems research and effectively communicate their findings in top-tier academic publications.
A talk at the Directed Energy Research Center (DERC) at TII will discuss rapid prototyping using laser-cutting facilities available at MakerSpace in Al Zeina. The talk will cover constructing prototypes from wood and acrylic and compare this approach to traditional 3D printing. The speakers will also describe the impact of the ‘4th Industrial Revolution’ on manufacturing in the UAE, and how makerspaces can contribute to Operation 300bn. Why it matters: This highlights the UAE's focus on advanced manufacturing and the role of makerspaces in fostering innovation and developing local capabilities.
This paper introduces two methods for creating Arabic LLM prompts at scale: translating existing English prompt datasets and creating natural language prompts from Arabic NLP datasets. Using these methods, the authors generated over 67.4 million Arabic prompts covering tasks like summarization and question answering. Fine-tuning a 7B Qwen2 model on these prompts outperforms a 70B Llama3 model in handling Arabic prompts. Why it matters: The research provides a cost-effective approach to scaling Arabic LLM training data, potentially improving the performance of smaller, more accessible models for Arabic NLP.
KAUST PhD student Amal Aboulhassan founded MaterialSolved, a startup created with support from the KAUST New Ventures Accelerator. The startup's focus area is not specified in the provided text. Why it matters: KAUST's efforts to translate research into startups highlights the increasing focus on commercializing academic innovation within the Kingdom.