On the afternoon of December 21, Beijing time, Loup Ventures, a venture capital firm, tested mainstream digital voice assistants through smart speakers, according to VentureBeat, the US technology media. Although Google Home is still ahead of all its competitors, other products are still ahead of each other's competitors, according to the US technology media VentureBeat.
As early as July, Loup Ventures released the annual Digital Assistant IQ Test Report, comparing Google Assistant with Apple Siri, Amazon Alexa and Microsoft Corpana, ranking the four AI systems for their performance in answering 800 real questions. Now, the company has released its annual Intelligent Speaker IQ Test Report, which focuses on the performance of digital assistants through smart speakers like Google Home, Apple HomePod and Amazon Echo. The results are interesting.
As before, Google Assistant on, Home Mini has a 100% understanding rate and a correct answer rate of 87.9%
Problem Understanding Rate and Correct Answer Rate of Different Intelligent Voice Assistants
Apple's Siri on the HomePod grew strongly from mid-February, but still lags behind when it passed the iPhone test in July. On the HomePod smart speaker, Siri correctly answered 74.6% of the questions, up from 52.3% in February, with an understanding rate of 99.6%. In July, the accuracy of the iPhone test was 78.5%. Interestingly, Loup noticed,
Amazon's Alexa and Microsoft's Corana rank third and fourth again, but both have improved their accuracy through smart speakers. Alexa on Echo had a correct response rate of 72.5%, compared with 64% in February and 61.4% in July. This progress makes it almost comparable to the expensive HomePod. Cortana got 63.4% accuracy through Hamankaton Invoke speakers, 57% in February and 52.4% in July.
Loup Ventures says proper nouns are still the biggest pain in smart speakers, otherwise they almost understand what users say to them. But apart from that, all the assistants are working.
The Correct Rate of Intelligent Speaker Answering Questions in Different Areas in Different Time Periods
While Loup is expected to update its smartphone digital assistant test results again in July, it does not expect artificial intelligence to answer everything they ask correctly