In this work, we introduce a deep learning-based polar code construction algorithm. The core idea is to represent the information/frozen bit indices of a polar code as a binary vector which can be interpreted as trainable weights of a neural network (NN). For this, we demonstrate how this binary vector can be relaxed to a soft-valued vector, facilitating the learning process through gradient descent and enabling an efficient code construction. We further show how different polar code design constraints (e.g., code rate) can be taken into account by means of careful binary-to-soft and soft-to-binary conversions, along with rate-adjustment after each learning iteration. Besides its conceptual simplicity, this approach benefits from having the "decoder-in-the-loop", i.e., the nature of the decoder is inherently taken into consideration while learning (designing) the polar code. We show results for belief propagation (BP) decoding over both AWGN and Rayleigh fading channels with considerable performance gains over state-of-the-art construction schemes.

A cutting-edge, £1 million robot ship will cross the Atlantic Ocean unmanned next year to commemorate 400 years since the maiden voyage of the Mayflower to the USA. The Mayflower Autonomous Ship (MAS) will set off on its pioneering, 2,750-mile trip in September 2020, following in the trail of its namesake 400 years earlier. The 15-metre long, catamaran-style ship will be powered by state-of-the-art renewable energy. It will be unmanned but will have marine AI on board, and will be steered from a control room in Plymouth, Devon - where the original Mayflower set off from. It will carry three research pods, containing sensors and other equipment, which scientists hope will pave the way for ground-breaking research into ocean conditions for autonmous navigation.

A hundred and thirty three years--that's how long Virginia-based Newport News Shipbuilding has been in the business of manufacturing ships. As the sole developer of U.S. Navy aircraft carriers, the company has constructed more than 30 warships, including the world's first and largest nuclear-powered carrier, which weighs 100,000 tons and comprises 300 million parts. Traditionally, building a ship like this might require 30 million to 40 million man-hours. Digital transformation has upended the ship building business, said Bharat Amin, VP and chief information officer at Newport News Shipbuilding, on stage at Siemens' Spotlight on Innovation, an annual technology conference held recently in Orlando. "You hear about smart cities, we want to be a smart shipyard," Amin told the audience.

The American Bureau of Shipping recently collaborated with Google Cloud and software engineers SoftServe to use artificial intelligence (AI) models to detect levels of corrosion and marine coatings breakdowns on brown- and bluewater vessels. The pilot project is aimed at developing image recognition software tools that can examine early signs of degradation in hull structures, to avoid unsafe working conditions, unscheduled maintenance and resulting operational downtime. The effort demonstrated how AI can support early detection of structural anomalies that are usually found through traditional, visual inspections. The project was focused on corrosion and coatings failures, but ABS engineers believe the new tools could also be used to detect stress fractures and larger hull deformations. These AI techniques -- in tandem with advanced data algorithms -- could be used to analyze images over time to understand the trends in corrosion and asset fatigue that would support a transition to more efficient class and maintenance regimes for everything from workboats to offshore structures.

We propose a new framework for constructing polar codes (i.e., selecting the frozen bit positions) for arbitrary channels, and tailored to a given decoding algorithm, rather than based on the (not necessarily optimal) assumption of successive cancellation (SC) decoding. The proposed framework is based on the Genetic Algorithm (GenAlg), where populations (i.e., collections) of information sets evolve successively via evolutionary transformations based on their individual error-rate performance. These populations converge towards an information set that fits both the decoding behavior and the defined channel. Using our proposed algorithm over the additive white Gaussian noise (AWGN) channel, we construct a polar code of length 2048 with code rate 0.5, without the CRC-aid, tailored to plain successive cancellation list (SCL) decoding, achieving the same error-rate performance as the CRC-aided SCL decoding, and leading to a coding gain of 1 dB at BER of $10^{-6}$. Further, a belief propagation (BP)-tailored construction approaches the SCL error-rate performance without any modifications in the decoding algorithm itself. The performance gains can be attributed to the significant reduction in the total number of low-weight codewords. To demonstrate the flexibility, coding gains for the Rayleigh channel are shown under SCL and BP decoding. Besides improvements in error-rate performance, we show that, when required, the GenAlg can be also set up to reduce the decoding complexity, e.g., the SCL list size or the number of BP iterations can be reduced, while maintaining the same error-rate performance.

We propose a new polar code construction framework (i.e., selecting the frozen bit positions) for the additive white Gaussian noise (AWGN) channel, tailored to a given decoding algorithm, rather than based on the (not necessarily optimal) assumption of successive cancellation (SC) decoding. The proposed framework is based on the Genetic Algorithm (GenAlg), where populations (i.e., collections) of information sets evolve successively via evolutionary transformations based on their individual error-rate performance. These populations converge towards an information set that fits the decoding behavior. Using our proposed algorithm, we construct a polar code of length 2048 with code rate 0.5, without the CRC-aid, tailored to plain successive cancellation list (SCL) decoding, achieving the same error-rate performance as the CRC-aided SCL decoding, and leading to a coding gain of 1 dB at BER of $10^{-6}$. Further, a belief propagation (BP)-tailored polar code approaches the SCL error-rate performance without any modifications in the decoding algorithm itself.

Designing and modifying complex hull forms for optimal vessel performances have been a major challenge for naval architects. In the present study, Principal Component Analysis (PCA) is introduced to compress the geometric representation of a group of existing vessels, and the resulting principal scores are manipulated to generate a large number of derived hull forms, which are evaluated computationally for their calm-water performances. The results are subsequently used to train a Deep Neural Network (DNN) to accurately establish the relation between different hull forms and their associated performances. Then, based on the fast, parallel DNN-based hull-form evaluation, the large-scale search for optimal hull forms is performed.

ULSAN, SOUTH KOREA – When Lee Dong-hee came to Ulsan to work for Hyundai Heavy Industries five years ago, shipyards in the city known as Hyundai Town operated day and night and workers could make triple South Korea's annual average salary. But the 52-year-old was laid off in January, joining some 27,000 workers and subcontractors who lost their jobs at Hyundai Heavy between 2015 and 2017 as ship orders plunged. To support their family, Lee's wife took a minimum wage job at a Hyundai Motor supplier. His 20-year-old daughter, who entered a Hyundai Heavy-affiliated university hoping to land a job in Ulsan, is now looking for work elsewhere. The Lee family's fortunes mirror the decline of Ulsan, which is now reeling from Chinese competition, rising labor costs and its overreliance on Hyundai -- one of the giant, family-run conglomerates, known as chaebol, that dominate South Korea.

The Robot Business Division of Hyundai Heavy Industries Holdings, the No. 1 robot company in Korea, has signed an MOU on the robot business with Naver Labs, a research and development corporation of Naver, the largest internet company in Korea. The two companies announced on May 28 that they held an MOU ceremony, with the participation of Yoon Jung-keun, head of the Robot Business Division and Song Chang-hyun, CEO of Naver Labs, at Hyundai Building in Gyeongdong in Seoul. Through this MOU, the two companies will join forces in the development and production of service robots. Hyundai Heavy Industries will take charge of production, sales, quality control, and development of service robots based on its capability to commercialize robots. Naver Labs will be in charge of technology research and development and the development of system and application software for robot production.

Finland's prime minister Juha Sipilä has said that the country's legislation must be changed so that it will allow experimenting with Artificial Intelligence technologies faster. Talking in Turku the politician sees that Finland must already think of the next hundred years, just after it's centennial anniversary in 2017. He believeves that adopting Artificial Intelligence at scale in different fields will be key in the future. Currently one of the Nordic country's strengths is it's heavy industry, like shipyards and automotive industry in the Southwest Finland. Both are being receiving continuous investment, with the Turku shipyard training programme and expansion of the Mercedes-Benz production in Uusikaupunki.