Search for: All records

According to the market research firm Tractica, the global artificial intelligence software market is forecast to grow to 126 billion by 2025. Additionally, the Gartner group predicts that during the same time as much as 80% of the routine work ,  which represents the bulk of human hours spent in today's project management (PM) activities, can be eliminated because of collaboration between humans and smart machines. Today's PM practices rely heavily on human input. However, that is not the optimum use of the human project manager's intuitive, innovative, and creative abilities. Many aspects of a project manager's work could be managed by machines that utilize AI/ML approaches to address nonroutine and predictive tasks. This paper describes IT project management (ITPM) processes and associated tasks and identifies the AI/ML approaches that can support them. 

Homomorphic encryption (HE) is the ultimate tool for performing secure computations even in untrusted environments. Application of HE for deep learning (DL) inference is an active area of research, given the fact that DL models are often deployed in untrusted environments (e.g., third-party servers) yet inferring on private data. However, existing HE libraries [somewhat (SWHE), leveled (LHE) or fully homomorphic (FHE)] suffer from extensive computational and memory overhead. Few performance optimized high-speed homomorphic libraries are either suffering from certain approximation issues leading to decryption errors or proven to be insecure according to recent published attacks. In this article, we propose architectural tricks to achieve performance speedup for encrypted DL inference developed with exact HE schemes without any approximation or decryption error in homomorphic computations. The main idea is to apply quantization and suitable data packing in the form of bitslicing to reduce the costly noise handling operation, Bootstrapping while achieving a functionally correct and highly parallel DL pipeline with a moderate memory footprint. Experimental evaluation on the MNIST dataset shows a significant ( 37× ) speedup over the nonbitsliced versions of the same architecture. Low memory bandwidths (700 MB) of our design pipelines further highlight their promise toward scaling over larger gamut of Edge-AI analytics use cases.