Researchers at Microsoft claimed to
have developed the first technology that recognizes the words in a conversation
as well as humans do. A team in Microsoft Artificial Intelligence and Research
created a speech recognition system that makes the same or fewer errors than
professional transcriptionists. The system had a word error rate (WER) of 5.9%
- the lowest ever recorded against the industry standard Switchboard speech
recognition task. The research milestone does not mean the computer recognized
every word perfectly. In fact, humans do not do that, either. Instead, it means
that the error rate – or the rate at which the computer misheard a word like “have”
for “is” or “a” for “the” – is the same as you would expect from a person
hearing the same conversation. The milestone means that, for the first time, a
computer can recognize the words in a conversation as well as a person would. The
milestone comes after decades of research in speech recognition, beginning in
the early 1970s with DARPA, the US agency tasked with making technology
breakthroughs. This accomplishment is the culmination of over twenty years of
effort. The milestone will have broad implications for consumer and business
products that can be significantly augmented by speech recognition. That includes
entertainment devices like the Xbox, accessibility tools such as instant speech
to text transcription and personal digital assistants such as Cortana.
Thursday, 3 November 2016
This Lab In A Smartphone Detects Cancer Instantly
Scientists at Washington State
University have developed a new low-cost, portable laboratory on a smartphone
that can analyse several samples at once to catch a cancer biomarker with 99%
accuracy, producing lab-quality results. At a time when patients and medical
professionals expect faster results, researchers are trying to translate bio-detection
technologies used in laboratories to the field and clinics, so patients can get
instant diagnoses in a physician’s office, an ambulance or the emergency room. The
researchers created an eight-channel smartphone spectrometer that can detect
human interleukin-6 (IL-6), a known biomarker for lung, prostate, liver, beast
and epithelial cancers. A spectrometer analyses the amount and type of
chemicals in a sample by measuring the light spectrum. Although smartphone
spectrometer exist, they only monitor or measure a single sample at a time,
making them inefficient for real world applications. The new multichannel
spectrometer can measure up to eight different samples at once using a common
test called Elisa, or colorimetric test enzyme linked immunosorbent assay, that
identifies antibodies and color change as disease markers. Although researchers
only used the smartphone spectrometer with standard lab-controlled samples,
their device has been 99% accurate. The researchers are now using the portable
spectrometer in real world situations. With this eight channel spectrometer,
researchers can put eight different samples to do the same test, or one sample
in eight different wells to do eight different tests; this increases this device’s
efficiency. The spectrometer would be especially useful in clinics and
hospitals that have a large number of samples without onsite labs, or for
doctors who practice abroad or in remote areas. Those can’t carry a whole lab
with them. They need a portable and efficient device.
Now, A Tattoo To Detect Alcohol Level In Real Time
Scientists have developed a
wearable skin tattoo that detects alcohol levels in sweat and transmits the
information t a smartphone, allowing users to monitor their drinking in real
time. The device could help reduce unsafe drinking that can lead to vehicle
crashes, violence and the degeneration of the health of heavy drinkers. It resembles
a temporary tattoo, but is actually a biosensor patch that is embedded with
several flexible wireless components. One component releases a chemical that
stimulated perspiration on the skin below the patch. Another component senses
changes in the electrical current flowing through the generated sweat, which
measures alcohol levels and sends them to the user’s cell phone. About 88,000
people in the US die from alcohol related causes including driving fatalities, which
accounted for nearly 10,000 deaths in 2014. This significant problem has been
addressed by the use of blood tests or breathalyzers by law enforcement. The new
wearable monitor has the advantage of being non-invasive and unseen by others.
A Smartwatch That Identifies Gestures And Objects
Scientists have developed a
technology that enables smart watches to recognize gestures such as taps and
flicks as well as identify objects held in the user’s hand, making possible new
types of interactions with wearable devices. Researchers at Carnegie Mellon
University in the US discovered that a software upgrade that repurposes the
smartwatch’s existing accelerometer enables the device to detect and
distinguish a variety of taps, flicks and scratches by the hands and fingers. This
function makes possible new applications that use common gestures to control
the smartwatch and other objects connected through the internet of things. By
monitoring vibrations that occur when people hold objects or use tools, the
smartwatch would be capable of recognizing objects and activities. It could
even be used to help tune a guitar, with the smartwatch displaying the note
transmitted. The technology has been developed by PhD students at the Human
Computer Interaction Institute (HCII).
Subscribe to:
Posts (Atom)