Meet Sirius: the Free Software Inspired by Siri
With the invention of Intelligent Personal Assistant (IPA) within mobile devices, Apple, Microsoft and Google revolutionized the industry by allowing people to initiate multiple basic applications like setting alarms, meetings, and calls just by speaking. As the next level to this, the engineering researchers from Michigan University introduces Sirius, an open-source software featuring same functionalities like Apple’s Siri application with customization facility at free of cost.
Sirius woks on various technologies including voice recognition, image matching, natural language processing and question-and answer system that accepts queries in the form of voice and images and responds in natural language.
The new application, supported with cloud, combines speech recognition, Image Matching (IMM) service, Natural Language Processing (NLP) and a question-and-answer system into one device. It deals with various types of queries from users including voice, image accompanied with speech and action-based processing.
With the help of Sirius, the mobile device works as a talking personal assistance where user can ask factual questions for answers generated from static version of the site. The developers can customize this application to receive knowledge from particular domain such as medicine, sports, automotive and food, reports Nicole Casal Moore for Michigan News.
In their study, researchers also determine the best acceleration platform to handle a complex workload. They have found that Field-Programmable Gate Array (FPGA) accelerated servers achieve an average of sixteen times reduction on query latency and Graphic Processing Unit (GPU) servers facilitate ten times query latency reduction translating to 2.6 times Total Cost of Ownership (TCO) reduction.