Welcome to the November 8, 2017 edition of ACM TechNews, providing timely information for IT professionals three times a week.
ACM TechNews mobile apps are available for Android phones and tablets (click here) and for iPhones (click here) and iPads (click here).
To view "Headlines At A Glance," hit the link labeled "Click here to view this online" found at the top of the page in the html version.
The online version now has a button at the top labeled "Show Headlines."
|
|
|
Killer Robots: Australia's AI Leaders Urge PM to Support a Ban on Lethal Autonomous Weapons
UNSW Newsroom
Wendy Frew
November 7, 2017
University of New South Wales professor Toby Walsh has organized a letter from 122 members of the Australian artificial intelligence (AI) research community to Australian Prime Minister Malcolm Turnbull, urging support for a ban on lethal autonomous weapons at a upcoming United Nations conference. "Australia should also commit to working with other states to conclude a new international agreement that achieves this objective," the letter says. The letter was issued at the same time as one signed by many Canadian AI experts calling on Canadian Prime Minister Justin Trudeau to support such a prohibition. Both letters follow a call to the UN by more than 100 founders of robotics and AI firms warning against weaponized AI. Walsh notes although AI "can be used for immense good," there is a sinister side to the technology, "especially if we let machines decide who lives and who dies. I very much hope we make the right choice here."
|
Touch-Sensitive Avatar-Robotic Arm Based on Real-Time Haptics
Keio University (Japan)
November 7, 2017
Researchers at Keio University in Japan say they have developed a "real-time-avatar-robotic arm" that transmits sound, vision, and tactile sensations to remotely located users. "This 'real haptics' is an integral part of the Internet of Actions technology, having applications in manufacturing, agriculture, medicine, and nursing care," says Keio's Takahiro Nozaki. He says the robotic arm is the world's first high-precision tactile-force transmission technology that remembers human movements, edits them, and reproduces them. In addition, the researchers note the robotic arm does not use conventional touch sensors, making it more affordable, compact, and robust in terms of malfunction and noise. The system uses high-precision motors integrated in the avatar arm and algorithms to drive them. The researchers say they plan to improve the system's ability to recognize the shape, material composition, and position of an object, and manipulate it according to real-time instructions from a user located at a distance, where the arm acts as a real-time avatar.
|
China's AI Ambitions Revealed by List of Most Cited Research Papers
Nikkei Asian Review
Shigenori Arai
November 2, 2017
Chinese universities claimed two of the top 10 spots on a list showing sources of the most frequently cited research papers in the field of artificial intelligence (AI) between 2012 and 2016, which was compiled via an analysis of scholarly publications by the Nikkei Asian Review and Elsevier. In addition, 15 Chinese organizations were among the top 100 on the list. A key factor in China's ascendancy in the AI sector is its ability to attract Chinese researchers who have studied and worked in the U.S. back to their homeland via financial incentives and a prosperous environment. The findings suggest China's strengthening of its research base is feeding into leading domestic universities as well as lower-tier schools and institutes. Singapore also scored high on the list, with three institutions in the top 100, and U.S. companies such as Microsoft and Google fared well, while Japan lagged far behind China.
|
Distinguishing Between Humans and Computers in the Game of Go
Phys.org
Lisa Zyga
November 6, 2017
Researchers at the universities of Toulouse and Paris-Saclay in France found they could differentiate between human and computerized Go players by analyzing the statistical characteristics of thousands of games played by people and algorithms. The researchers built databases of 8,000 games played by amateur humans, 8,000 played by the software Gnugo, 8,000 played by the software Fuego, and 50 games played by the software AlphaGo. Their analysis found network-based software forms more "communities"--signs the algorithms are creating varied and diverse strategies--than humans. The team also found statistical differences between the computer- and human-generated networks are much larger than the variability within each network. They suggest these differences could form the basis of a new type of Turing test. "We think our work indicates a path towards a better characterization and understanding of the differences between human and computer decision-making processes, which could be applied in many different areas," says Paris-Saclay's Olivier Giraud.
|
Choreography With Computers
Asian Scientist
Sim Shuzhen
November 6, 2017
Singapore Management University professor Akshat Kumar is developing computational methods for choreographing the movements of autonomous cars and ships using multi-agent planning and automated decision-making. "One of my research aims is to develop computationally efficient techniques to realize the promise of large cooperative teams making decisions towards a common goal," Kumar says. He bases his computational models on actual data, employing multi-agent reinforcement learning to find optimization strategies. "We build simulators for different urban phenomena--for maritime ecosystems or for how taxis move in urban environments, for example," Kumar notes. "Using the data coming from these simulators, we can actually optimize agent policies." Kumar also is grappling with integrating both robotic and human decision-making systems within computational algorithms, as well as the challenge of controlling vast numbers of agents while still maintaining computational efficiency. He says there is a need to build specific capabilities into decision-making algorithms that can behave predictively under specific circumstances.
|
A 'Virtual Wall' That Improves Wireless Security and Performance
Dartmouth College
David Hirsch
November 8, 2017
Researchers at Dartmouth College have used three-dimensional (3D) printing to produce an inexpensive reflector that directs wireless signals to where users need them most, which could enhance Wi-Fi signals and strengthen wireless security. Tailoring the coverage of wireless networks inside buildings is important to improving signal reception in desired areas while weakening signals in others, and by shaping wireless signals users can increase wireless efficiency by lessening the signal-deadening impact of building materials and interior layouts. The researchers improved on previous studies with a systematic approach to optimizing reflector shapes for enabling a more developed set of signal distributions. They designed an algorithm that optimizes a reflector's 3D shape to target wireless coverage, and also developed an approach to simulating how radio signals spread and interact with objects in their environment. The researchers were scheduled to present their findings today at the ACM International Conference on Systems for Energy-Efficient Built Environments (BuildSys 2017) in Delft, the Netherlands.
|
New Research Aims to Solve the Problem of AI Bias in 'Black Box' Algorithms
Technology Review
Jackie Snow
November 7, 2017
Former Microsoft researcher Sarah Tan and colleagues suggest they may have mitigated the problem of uncovering potential bias within "black box" algorithms. They tested their solution on black-box assessment models concerning loan risks and default rates and recidivism risk predictions. The team first generated a model that emulates the black-box algorithm being examined and estimates a risk score based on an initial set of data. They next built a second model trained on real-world outcomes, using it to determine which factors from the initial dataset were vital in final outcomes. In the first case, the researchers found the lending model seemed to be disregarding a key factor, while in the second instance they found the recidivism algorithm was probably biased for certain age and racial groups. "We need to be aware this is happening, and not close our eyes to it and act like it's not happening," says University of Massachusetts, Amherst professor Brendan O'Connor.
*May Require Free Registration
|
AI Researchers Leave Elon Musk Lab to Begin Robotics Startup
The New York Times
Cade Metz
November 6, 2017
Alumni of Elon Musk's OpenAI lab and the University of California, Berkeley have launched Embodied Intelligence, a company that wants to bring advanced, machine-learning robotic automation to factories, warehouses, and households. Embodied Intelligence co-founder Pieter Abbeel says the company's specialty will be the development of complex algorithms that can enable machines to learn tasks on their own via reinforcement learning. Of particular note is the startup's focus on imitation learning, a process in which machines can learn discrete tasks demonstrated by humans. Co-founder Peter Chen says they compile data on what a person is doing using virtual reality headsets and handheld motion trackers, and "then we can train the machine to imitate the human." Many researchers see machine learning as essential to revolutionizing the field of robotics. "It is obvious that this is what you need to build flexible, agile robotics," says machine-learning pioneer Geoff Hinton.
*May Require Free Registration
|
Nanoscale 'Abacus' Uses Pulses of Light Instead of Wooden Beads to Perform Calculations
University of Exeter
November 2, 2017
Researchers at the University of Exeter in the U.K. have developed a nanoscale optical "abacus" that uses light signals to perform arithmetic computation, a breakthrough that could lead to new, more powerful computers that combine computing and storage functions in one element. The researchers say the device works by counting pulses of light, similar to how beads are used to count with a conventional abacus. "This device is able to carry out all the basic functions you'd associate with the traditional abacus--addition, subtraction, multiplication, and division--but what's more it can do this using picosecond light pulses," says Exeter professor C. David Wright. The researchers have installed the optical abacus on a photonic microchip that can be easily manufactured. Thus far, the team has used the new system to perform calculations with two-digit numbers using two photonic phase-change cells, but they say the extension to large multi-digit numbers simply requires the use of more cells.
|
Researchers Develop Data Bus for Quantum Computer
University of Innsbruck
November 3, 2017
Researchers at the University of Innsbruck in Austria say they have developed a technique to transfer quantum information between systems that are encoded in different ways. Other scientists already have developed small-scale quantum processors and memories, and they have used different protocols to encode logical quantum bits (qubits). "We have developed a protocol that allows us to merge quantum systems that are encoded differently," says University of Innsbruck researcher Hendrik Poulsen Nautrup. The researchers propose a method that involves locally modifying specific elements of the encoded qubits, a process known as lattice surgery, which enables them to couple systems such as quantum processors and memories. The team says the next step is for quantum information to be teleported from the processor to the memory, and vice versa. "Similar to a data bus in a conventional computer, scientists can use this technique to connect the components of a quantum computer," Nautrup notes.
|
Computer Scientists Aid in Major Astronomical Discovery
University of California, Merced
Jason Alvarez
November 1, 2017
Researchers at the University of California, Merced (UC Merced) recently helped to make a major advance in astronomy, as scientists announced last month they observed two neutron stars colliding. The UC Merced team developed data-processing tools that enabled astronomers to determine they were indeed observing an unprecedented event. The database techniques--the array similarity joint operator and views over arrays--enable scientists to rapidly compare huge amounts of array data across multiple databases by minimizing data transfer, thus reducing network congestion and eliminating redundant processing. "Our techniques let astronomers assess far more candidate images than would otherwise be possible," says UC Merced professor Florin Rusu. The researchers note the new database techniques are broadly applicable, and could be used in a variety of fields, from astronomy to genetics. The database techniques were first described in papers presented at the 2016 and 2017 ACM Special Interest Group on Management of Data (SIGMOD) conferences.
|
Students Explore Immigration Through a Big Data Lens
Texas Advanced Computing Center
Aaron Dubrow
November 1, 2017
Graduate and undergraduate students will learn how to use advanced computing skills to investigate U.S. immigration policies next week at SC17: the International Conference for High Performance Computing, Networking, Storage, and Analysis, in Denver, CO. Organizers say participants will learn computing, data analysis, and visualization skills, and then tackle critical immigration issues in teams by analyzing large datasets to extract convincing arguments for their positions. "Data analysis and visualization provide a vehicle to get to the truth behind the rhetoric," notes the Texas Advanced Computing Center's Kelly Gaither. Students also will get mentorship, career advice, and networking opportunities with other computing professionals. "Our goal is to provide students with advanced computing skills and the ability to visualize complex data in a way that is useful to everyone, from those affected by immigration policy to those creating immigration policy," says University of Illinois professor Ruby Mendenhall.
|
3D Printing Gets a Turbo Boost From U-M Technology
University of Michigan News
James Lynch
October 31, 2017
University of Michigan (U-M) researchers say they have developed an algorithm that enables three-dimensional (3D) printers to deliver high-quality results at speeds up to two times faster than those in conventional systems, with no added hardware costs. The software monitors a 3D printer's vibrations and makes adjustments in real time to enable it to work more quickly. "Armed with knowledge of the printer's dynamic behavior, the program anticipates when the printer may vibrate excessively and adjusts its motions accordingly," says U-M professor Chinedum Okwudire. He notes the new software is like a person who realizes their voice is going to be overly amplified and acts preemptively to alter it to fit the conditions. The researchers say the software acts similarly because it knows what the behavior of the printer is going to be ahead of time, and they note it also can be used on industrial-grade machines that suffer from similar limitations due to vibrations.
|
|
