Abdulrahman Mahmoud, a postdoctoral fellow at Harvard University, discusses software-directed tools and techniques for processor design and reliability enhancement in ML systems. He emphasizes the need for a nuanced approach to numerical data formats supported by robust hardware. He advocates for integrating reliability as a foundational element in the design process. Why it matters: This research addresses the critical challenge of hardware reliability in AI processors, particularly relevant as the field moves towards hardware-software co-design for sustained growth.
NYU Abu Dhabi hosted a talk by Prof. Debdeep Mukhopadhyay on the intersection of machine learning and hardware security. The talk covered using ML/DL for side-channel attacks, leakage assessment in crypto-devices, and threats to hardware security primitives. Prof. Mukhopadhyay is a visiting professor at NYU Abu Dhabi and Institute Chair Professor at IIT Kharagpur. Why it matters: The talk highlights the growing importance of hardware security in modern systems and the role of machine learning in both attacking and defending hardware vulnerabilities.
The UAE has announced the development of a new AI chip during a tech summit in Dubai. Details about the chip's specifications, capabilities, or intended applications were not disclosed in the provided article. Why it matters: This announcement signals the UAE's ambition to develop indigenous AI hardware capabilities, reducing reliance on foreign suppliers.
KAUST researchers have designed an integrated circuit logic lock to protect electronic devices from cyberattacks. The protective logic locks are based on spintronics and can be incorporated into electronic chips. The lock uses a magnetic tunnel junction (MTJ) where the keys are stored in tamper-proof memory, ensuring hardware security. Why it matters: This hardware-based security feature could significantly increase confidence in globalized integrated circuit manufacturing, protecting against counterfeiting and malicious modifications.
The Technology Innovation Institute (TII) in Abu Dhabi has launched a cloud service providing access to its in-house Quantum Processing Units (QPUs), which range from 5 to 25 qubits. Initially available to TII partners, the service allows users to run quantum workloads on TII’s quantum hardware via the cloud, using the open-source Qibo framework as the software layer. These QPUs feature in-house fabricated chips and demonstrate quantum coherence times up to ten times longer than TII's first-generation prototypes. Why it matters: This launch provides a platform for experimentation and development of hybrid quantum-classical workflows on locally developed infrastructure, accelerating quantum research in the region.