Country
Counterexample-Driven Synthesis for Probabilistic Program Sketches
Češka, Milan, Hensel, Christian, Junges, Sebastian, Katoen, Joost-Pieter
Probabilistic programs are key to deal with uncertainty in e.g. controller synthesis. They are typically small but intricate. Their development is complex and error prone requiring quantitative reasoning over a myriad of alternative designs. To mitigate this complexity, we adopt counterexample-guided inductive synthesis (CEGIS) to automatically synthesise finite-state probabilistic programs. Our approach leverages efficient model checking, modern SMT solving, and counterexample generation at program level. Experiments on practically relevant case studies show that design spaces with millions of candidate designs can be fully explored using a few thousand verification queries.
RL-GAN-Net: A Reinforcement Learning Agent Controlled GAN Network for Real-Time Point Cloud Shape Completion
Sarmad, Muhammad, Lee, Hyunjoo Jenny, Kim, Young Min
We present RL-GAN-Net, where a reinforcement learning (RL) agent provides fast and robust control of a generative adversarial network (GAN). Our framework is applied to point cloud shape completion that converts noisy, partial point cloud data into a high-fidelity completed shape by controlling the GAN. While a GAN is unstable and hard to train, we circumvent the problem by (1) training the GAN on the latent space representation whose dimension is reduced compared to the raw point cloud input and (2) using an RL agent to find the correct input to the GAN to generate the latent space representation of the shape that best fits the current input of incomplete point cloud. The suggested pipeline robustly completes point cloud with large missing regions. To the best of our knowledge, this is the first attempt to train an RL agent to control the GAN, which effectively learns the highly nonlinear mapping from the input noise of the GAN to the latent space of point cloud. The RL agent replaces the need for complex optimization and consequently makes our technique real time. Additionally, we demonstrate that our pipelines can be used to enhance the classification accuracy of point cloud with missing data.
The Shot That Stopped Basketball
The nature of basketball is such that its most cathartic moment--when the ball goes decisively and irretrievably through the hoop--is the same every time. The ball piercing the basket is both a discrete event and a continuous waterfall of motion that, for active players, is constant throughout their careers. They shoot in practice, they shoot in the game, they shoot and shoot and shoot. The motion becomes so ingrained in their muscle memory that the gesture requires only its activation; everything else--the elevation, the aiming at the basket, the cocking of the elbow and the follow-through of the hand--is programmed. I found myself thinking about the waterfall of shots in the wake of one of the more dramatic ones in recent N.B.A. history: Damian Lillard, of the Portland Trail Blazers, hitting the game-winning, series-ending shot against the Oklahoma City Thunder in Game Five.
Immigration Services Agency to toughen Japanese-language school standards
The Immigration Services Agency plans to strengthen its eligibility standards for Japanese-language schools, it was learned Saturday. The move comes as Japanese-language schools have been under fire for accepting many foreign students whose purpose is to work in Japan. The number of Japanese-language schools recognized by the government grew 1.6 times over the past five years to 749 as of April 2. The government late last year outlined plans to improve the quality of Japanese-language schools as part of efforts to bring in more foreign workers to the country. Under the agency's plan, the requirement for the average student attendance rate would be revised from the current 50 percent or more in a month to 70 percent or more in a period of seven months. Schools failing to meet the requirement would not be allowed to accept foreign students.
Trains, planes and automobiles to celebrate Golden Week
If you're looking for some mundane distractions to get you through the holiday period, Shukan Taishu (May 6-13) has got just the thing. Its "Reiwa Commemorative Edition" introduces unusual rides. Not to be outdone, a park in Tochigi has camels for the same purpose, as does another in Chiba offering elephant rides. At Hakkeijima Sea Paradise in Yokohama, visitors from age 10 (who can prove they can swim for a distance of 25 meters) may emulate the "boy on a dolphin" theme and ride atop a friendly beluga whale. Two-wheeled Segway personal transporters are available for inexpensive rental at Showa Memorial Park in the city of Tachikawa.
In 1983, This Bell Labs Computer Was the First Machine to Become a Chess Master
Chess is a complicated game. It's a game of strategy between two opponents, but with no hidden information and all of the potential moves known by both players at the outset. With each turn, players communicate their intent and try to anticipate the possible countermoves. The ability to envision several moves in advance is a recipe for victory, and one that mathematicians and logicians have long found intriguing. Despite some early mechanical chess-playing machines--and at least one chess-playing hoax--mechanized chess play remained hypothetical until the advent of digital computing.
Headspace: How the meditation app turns your stressful phone into a source of calm
Meditation and mindfulness have been around for thousands of years. But the advent of smartphones and computers led to a new phenomenon: the mindfulness app. There are a few to choose from, including the punchy, assertive 10% Happier, the elegant and placid Calm and the first app that really brought mindfulness to our phones, Headspace. Andy Puddicombe, a former Buddhist monk who went on to run a meditation clinic in London, met a new business partner, Richard Pierson, and launched Headspace in 2010. The company began as an events organisation and led to the now-ubiquitous app in 2012.
Alphabet's smart neighborhood could have shape-shifting 'superblocks'
Too many cities are built around cars rather than people. Sidewalk Labs, an offshoot of Google's parent company Alphabet, wants its smart neighborhood in Toronto to be different. It's considering a so-called superblock concept, modeled after Barcelona's, that bundles smaller streets together and limits vehicles to the perimeter. The smaller lanes inside each superblock would then become safer, quieter spaces for pedestrians and cyclists. Sidewalk Labs wants to go a step further, though, with real-time traffic monitoring and movable street furniture.
Hitting the Books: When better living through technology isn't enough
Welcome to Engadget's newest series, Hitting the Books. With less than one in five Americans reading just for fun these days, we've done the hard work for you by scouring the internet for the most interesting, thought provoking books on science and technology we can find and delivering an easily digestible nugget of their stories. Modern tech culture has long been enamored with the mythos of the lone genius achieving superhuman status (a la The Matrix). Whether it's Jack Dorsey's self flagellating dietary restrictions, Peter Thiel's obsession with "young blood" transfusions, or Tim Ferris' outright maniacal 4-hour self improvement regimens, if you're a wealthy white guy in Silicon Valley and not trying to live forever, you're doing it wrong. But for all the Bond villain-esque grifters selling the promise of eternal youth in 12 easy steps, a dedicated cadre of technologists have spent years investigating how we might actually achieve Ray Kurtweil's predicted singularity.
Video Friday: Massive Solar-Powered Drone, and More
Video Friday is your weekly selection of awesome robotics videos, collected by your Automaton bloggers. We'll also be posting a weekly calendar of upcoming robotics events for the next few months; here's what we have so far (send us your events!): Let us know if you have suggestions for next week, and enjoy today's videos. Soft-bubble, "a highly compliant dense geometry tactile sensor for robot manipulation," is a sort of combined gripper and 3D camera that uses a soft membrane to grasp and image objects at the same time. HAPS Mobile, a SoftBank-backed company, is developing a high-altitude pseudo satellite: a massive, solar-powered, long endurance drone that acts like a much cheaper and more versatile satellite over a smaller area.