IBM researchers have set a milestone in conversational speech recognition by achieving a new industry record of a 5.5 percent word error rate, surpassing its previous record of 6.9 percent, according to the company's blog post. The researchers conducted a difficult speech recognition task to achieve this record, where they recorded conversations between humans discussing typical everyday topics like "buying a car." This recorded corpus, titled "SWITCHBOARD", has been used for over two decades to benchmark speech recognition systems. To achieve the 5.5 percent record, the researchers focused on extending the company's application of deep learning technologies by combining LSTM (Long Short Term Memory) and WaveNet language models with three strong acoustic models. The first two models were six-layer bidirectional LSTMs, with one of the models being equipped with multiple feature inputs and the other being trained with speaker-adversarial multi-task learning.

Nice and steady does it. A video analysis system uses motion tracking and machine learning to assess how good surgeons are at suturing a wound. "Different surgeons all have different styles of suturing," says Aneeq Zia at the Georgia Institute of Technology in Atlanta, whose team developed the system. The team first captured footage of 41 surgeons and nurses practising their suturing and knot-tying skills on foam boards. Participants also wore an accelerometer on each hand to capture motion data, and a clinician then watched the videos and rated each person's skills.

My wife used to love reading but since her stroke has aphasia, no speech, limited vision and limited dexterity in her left hand only. She can select TV channels on a remote but she cannot read a short news story let alone a novel, so she listens to the radio and watches a lot of TV. I thought of getting her a Kindle e-reader but they don't seem to do text to speech any more. A shop assistant suggested a tablet with a text-to-speech app. It needs a really simple interface or my wife will not be able to use it without assistance.

Didi Chuxing President Jean Liu speaks at the WSJ D Live technology conference in Laguna Beach, California, on Oct. 25, 2016. The market for Silicon Valley engineers with expertise in artificial intelligence and automated driving tech keeps getting tighter. Didi Chuxing, the ride-hailing giant backed by Apple that forced Uber to abandon its efforts to win over the Chinese market last year, just opened an R&D center in the self-driving car capital with a big "help wanted" sign in the window. Mountain View, California-based Didi Labs will focus on "AI-based security and intelligent driving technologies," the company said on March 8. It will be led by computer scientist Fengmin Gong, vice president of the Didi Research Institute and a co-founder of Palo Alto Networks.

How do you ensure a smooth hand-off between data science and engineering on ML projects? How do you ensure a smooth hand-off between data science and engineering on ML projects? Excellent question, and one that I've been thinking a lot. At Quora, my data science team works closely with machine learning engineers. Joe Isaacson already provided a great answer to this question, and I will add some of my learnings.

Self-driving cars, medical research, financial market predictions: The applications of machine learning technologies seem limitless. Data processing isn't a small task: Countless human work hours are spent collating and developing data, seeking patterns and information from seemingly unrelated tidbits. And there is a constantly growing number of data sources: Excel spreadsheets or SQL (or NoSQL) databases have long-since replaced the old-school filing cabinets of records, not to mention the growth in tracked website usage, which can improve web experience for users, as well as advertising options for companies. With all of these potential areas of growth, which industries can you expect to be most affected by advances in machine learning? Having personal exposure to emerging health-care tech, I can assure you we're in for some exciting times.

In 2011, IBM sent its supercomputer Watson onto the popular American TV quiz show Jeopardy where it succeeded in matching wits with and beating two of the TV show's most successful players. That was over five years ago, but if you ask members of the public to describe IBM Watson, those in the know will say that it's a huge great black mainframe computer that's incredibly smart. Yet according to IBM, that's where you'd be wrong – the computing giant is adamant that the future of artificial intelligence will not be one big scary digital brain, and technology is definitely not going to kill us off one day. "When IBM Watson first came out, we used to think about it as a giant brain in a jar, but it's not that," John Cohn, an IBM Fellow in the IBM Watson Internet of Things (IoT) division tells IBTimes UK while showing us around IBM's new global IoT headquarters in Munich, Germany. "It's a bunch of tools that you can use to compose systems that interact naturally with humans, learns from their situation, adapting and then applying that knowledge.

Physicists have developed a quantum machine learning algorithm that can handle infinite dimensions--that is, it works with continuous variables (which have an infinite number of possible values on a closed interval) instead of the typically used discrete variables (which have only a finite number of values). The researchers, Hoi-Kwan Lau et al., have published a paper on generalizing quantum machine learning to infinite dimensions in a recent issue of Physical Review Letters. As the physicists explain, quantum machine learning is a new subfield within the field of quantum information that combines the speed of quantum computing with the ability to learn and adapt, as offered by machine learning. One of the biggest advantages of having a quantum machine learning algorithm for continuous variables is that it can theoretically operate much faster than classical algorithms. Since many science and engineering models involve continuous variables, applying quantum machine learning to these problems could potentially have far-reaching applications.

When Amazon purchased Kiva Systems in 2012, the interest in Autonomous Mobile Robots (AMRs) for the warehouse soared. For a while, Kiva, now rebranded Amazon Robotics, continued to sell robots to other companies. But, after piloting the robots in some warehouses, and figuring out the optimal way to deploy them, Amazon stopped selling robots to other companies and took everything their robotic division could produce for their own distribution centers.

IBM is teaming up with Salesforce to make it easier for Salesforce customers to use data from IBM's Watson artificial intelligence platform. As part of the partnership, IBM has signed a deal to deploy the Salesforce Service Cloud for internal use there. The value of the deal and proposed pricing of the joint products were not disclosed. The deal has several parts. Both companies, like many others in the tech industry, are making investments in artificial intelligence and machine learning, by which computer programs attempt to connect data in new ways to provide new kinds of insights and assistance to users, and both frequently cite the importance of AI in their sales pitches to customers.