In Stanford Reading Test humans get beaten by Alibaba Neural Network. A deep neural network model have been developed by the Institute of Data Science of Technologies, –the research arm of Alibaba— outdid human participants in the Stanford Question Answering Dataset abbreviated as Squad.
Alibaba’s AI learning model can read from paragraphs to sentences to words. And this model was based on the Hierarchal Attention Network, making it much similar to the natural human language.
According to Si Lou, chief scientist of natural language at IDST, first time it has happened that a machine has outperformed humans on such a test. Human participants in the reading test were able to score 82.304 based on the January 5 tests.
A similar kind of technology from Microsoft Research Asia scored 82.640, also beating the humans in January 3 tests.
Must Read: A right way to get dressed could relieve aching back in winters
The giant Alibaba’s result was officially registered by Squad on January 11, a day ahead that of Microsoft’s, technically making Alibaba the first to beat human scores.
Andrew Pickup, a spokesman for Microsoft opines that these kinds of tests are useful benchmarks for how far along the AI journey the companies may be. He also adds “However, the real benefit of AI is when it is used in harmony with humans.”
Stanford University AI experts managed to develope the 100,000 questions test to get measured the ability of machines to process a large amount of information and answer them accurately. The questions have been generated based on hundreds of Wikipedia articles. That’s way, the AI model can supply specific and precise answers only.
Squad is regarded to be the most comprehensive and authoritative machine-reading gauge. Tech companies like Google, Facebook, Microsoft, IBM, and Samsung do also use it to test if the machine learning models they built could answer questions from the data set.
Squad developer namely Pranav Rajpurkar of the Stanford Machine Learning Program describes Alibaba’s feat in the reading test as a strong start to the year 2018 in terms of AI development.
Using the said neural network that can answer questions that offer clear-cut answers, humans intervention & input can be lessened in answering the queries that are related to customer service. The learning model can identify questions that are asked by customers and could look for the most relevant answers from a set of already prepared materials.
Chinese tech giant Alibaba tested this AI technology on chatbot Ali Xiaomi that have been used by retailers on the company’s virtual stores, Taobao & Tmall. A similar tech that is also based on the network was used to answer voluminous inbound queries from customers during the month of November 2017 Singles Day sale in China. Alibaba could rake in $25 billion in the said online shopping festival alone.
Natural language processing technology had been the most dominant form of AI in consumer technology products with virtual assistant and voice activation features that have been previewed during the CES 2018.
