From The Lab: Information Technology
From the world of information technology, here are the latest publications, experiments, and breakthroughs, and what they mean.
Context: Optical fibers cross cities and oceans and form the backbone of much of the world’s high-speed data communications network. But the throughput of optical fiber is limited by how fast data can be switched across networks. In a conventional router, light from fibers must be converted into an electrical signal, switched to an appropriate cable, then converted back to light again; this process can slow the speed of information transfer by a factor of ten. A router that did not need to use an electrical signal would be inherently faster. But the electrical signals are necessary to hold data intact until the optical switch to the next cable is ready. A team of scientists from the Ecole Polytechnique Federale de Lausanne in Switzerland has demonstrated a novel way to slow down light in an optical fiber, so that switching from cable to cable can be coordinated with light signals instead of electricity.
Methods and Results: The speed at which light travels through different physical media is not constant; light interacts with the matter it passes through, and this interaction can slow it down ever so slightly. Kwang Yong Song and colleagues fired an intense laser down a standard telecommunications fiber, causing atoms in the fiber to vibrate. If an optical signal is sent from the fiber’s other end, the light interacts with these moving atoms more than it would with unperturbed ones, and it slows down by tens of nanoseconds along several kilometers’ length of fiber. Critically, the light’s speed in the fiber can be easily modulated: the more intense the laser, the more the atoms move, and the more the optical signal is delayed.
Why it Matters: The Swiss team is not the first to slow down light, but its experiments are notable for how they might affect the telecommunications industry. The new technique makes all-light routers feasible and could boost throughput from existing optical networks severalfold. That might enable scattered computers to link together seamlessly into a networked supercomputer, for instance. The method also works in standard optical fibers, making it compatible with existing telecommunications networks.
Context: Microphones placed around a meeting room tend to yield recordings where voices overlap and are hard to distinguish. When there are at least as many microphones present as people talking, computer algorithms have been able to isolate the audio of each speaker. But if fewer microphones are used, these methods don’t work, and problems of voice overlapping can persist. Alternative methods require creating a profile of each speaker’s voice from previous recordings or making certain assumptions about the audio signals. Now Francis Bach and Michael Jordan of the University of California, Berkeley, have developed an algorithm that separates the voices of multiple speakers in recordings made with just one microphone, without requiring strong prior assumptions or speaker profiles.
Methods and Results: Bach and Jordan’s algorithm homes in on the voice characteristics that are most likely to vary among people. The recorded sounds are laid out in a spectrogram, which shows the intensity of sound of various frequencies over time in a two-dimensional graph. Bach and Jordan’s algorithm automatically divides up the spectrogram among the speakers; it assumes that parts of the spectrogram are likely to be from the same speaker if they are near each other on the graph, vary similarly over time, or are alike in pitch and timbre. The algorithm is trained on samples in which separately recorded voices have been mixed; based on the training, the algorithm assigns a relative importance to each characteristic – say, timbre or tempo. Then the algorithm applies this training to new recordings. So far, the authors have been able to separate the overlapping voices in several recordings of pairs of speakers. Although the separation is not perfect, both speakers are more intelligible.
Why it Matters: Historians, journalists, lawyers, and other professionals rely on recorded conversations. These recordings are often made using a single microphone but feature multiple voices. By making babble more comprehensible, Bach and Jordan’s algorithm promises to make such recordings more useful and easier to analyze. Hence, users may no longer be forced to haul around bulky, expensive equipment when recording important conversations and events.
Be there when AI pioneers take center stage at EmTech Digital 2019.Register now