Skip directly to content

Information technology

Microsoft’s neural network at par with humans in recognizing speech

Microsoft's artificial intelligence program has reached the accuracy level of a professional human transcriber in recognizing speech, achieving the same error rate of 5.9%, an improvement from its record of 6.3% just a month ago. The overall score is a measurement of how the AI and the human transcriptionists did on three indicators - how many times both wrongly substituted sounds, omitted a word, or inserted the wrong word. According to Microsoft's chief speech scientist Xuedong Huang, achieving parity with a human is a historic achievement. The company used 2,000 hours of data to train its neural network. Improved automated speech recognition systems can enhance current speech-to-text services as well as Cortana, adding features that could allow the deaf to understand spoken words.

Key Takeaway: 

Microsoft’s neural network performed as accurately as a professional human transcriptionist at recognizing speech with an error rate of 5.9%; the improvement can introduce better speech-to-text systems and artificial intelligence assistants that can aid the deaf to hear.

Publication: 
Publication Date: 
October 19, 2016

Wal-Mart adopts blockchain for first time to track pork supplies

Wal-Mart has partnered with IBM to use the latter's blockchain technology to track its pork supplies in China. This is the first time that Wal-Mart, the largest retailer in the world, is adopting blockchain technology. Tracking pork is just a start for the company. If the project is successful, it could also expand to include other commodities. Wal-Mart wants to be able to monitor where the meat goes in its supply chain using serial numbers and record systems. The retailer giant is also partnering with Tsinghua University for academic research on the project. The news is reported as a significant development for the growing blockchain as a service (BaaS) industry.

Key Takeaway: 

Retail giant Wal-Mart is adopting blockchain technology through a partnership with IBM to monitor the pork supply chain in China, which could be the beginning of a new solution for tracking commodities and food supply chains.

Transforming Business Models: 
Publication: 
Publication Date: 
October 19, 2016

Singapore will pilot autonomous buses and real-time monitoring of rail system

The Singaporean Land Transport Authority in partnership with the Nanyang Technology University will test the use of two driverless buses within the Jurong Innovation District to improve road safety, reduce vehicle congestion and pollution, and address manpower challenges. The autonomous buses will be driving along a pre-determined route and will be recharged at bus stops. Heavy focus is being placed by LTA on road safety as the project announcement came a day after a crash incident involving nuTonomy, which was testing its autonomous cars in the One North Business District. The two parties will also develop a real-time monitoring system for its national rail network to identify potential defects without disrupting operations.

Key Takeaway: 

The transport regulator of Singapore in partnership with the NTU will pilot the use of autonomous buses in the Jurong Innovation District to address transportation-related issues.

Publication: 
Publication Date: 
October 19, 2016

Tesla Model S, X and Model 3 will have self-driving hardware

Tesla Motors announced it will start adding $8,000 worth of full autonomous driving hardware to the Model S, X and Model 3 sedans in production. The new vehicles will have eight surround cameras with 360 degree visibility up to 250 meters of range and 12 updated ultrasonic sensors that can detect objects at almost twice the distance of the existing system. Its computing system will be 40 times more powerful than the initial one it introduced. However, the hardware still has to wait for advances in the company's AutoPilot software before the cars can fully drive themselves without human intervention. The system will also have to gather data from millions of miles of real-world driving before it can be commercialized.

Key Takeaway: 

Tesla will begin installing the hardware that can enable full self-driving capability in its Model S, X and Model 3 sedans but software required to run them is still being developed.

Market Disruptions: 
Publication: 
Publication Date: 
October 19, 2016

Robot babies used in Japan to encourage population growth

Decline in the population has led to a trend in Japan of developing robotic babies to encourage couples to have children. Some robots look like real children while others mimic or imitate the behavior of babies. Examples include Toyota's Kirobo Mini, which can recognize and respond to people, and Yotaro's robotic baby simulator, which projects facial emotions and expressions in response to touch, mood or illness. There are also projects in development which use artificial intelligence to enable the robots to form a relationship with humans. Studies in the US and Australia have shown that robotic babies have increased pregnancies among test subjects.

Key Takeaway: 

Faced with a declining population, roboticists and engineers in Japan are developing robots that look or resemble real babies to encourage young couples to start their own families.

Transforming Business Models: 
Publication: 
Publication Date: 
October 17, 2016

Pages