Adversarial training is one of the most effective approaches to improve model robustness against adversarial examples. However, previous works mainly focus on the overall robustness of the model, and the in-depth analysis on the role of each class involved in adversarial training is still missing. In this paper, we propose to analyze the class-wise robustness in adversarial training. First, we provide a detailed diagnosis of adversarial training on six benchmark datasets, i.e., MNIST, CIFAR-10, CIFAR-100, SVHN, STL-10 and ImageNet. Surprisingly, we find that there are remarkable robustness discrepancies among classes, leading to unbalance/unfair class-wise robustness in the robust models. Furthermore, we keep investigating the relations between classes and find that the unbalanced class-wise robustness is pretty consistent among different attack and defense methods. Moreover, we observe that the stronger attack methods in adversarial learning achieve performance improvement mainly from a more successful attack on the vulnerable classes (i.e., classes with less robustness). Inspired by these interesting findings, we design a simple but effective attack method based on the traditional PGD attack, named Temperature-PGD attack, which proposes to enlarge the robustness disparity among classes with a temperature factor on the confidence distribution of each image. Experiments demonstrate our method can achieve a higher attack rate than the PGD attack. Furthermore, from the defense perspective, we also make some modifications in the training and inference phases to improve the robustness of the most vulnerable class, so as to mitigate the large difference in class-wise robustness. We believe our work can contribute to a more comprehensive understanding of adversarial training as well as rethinking the class-wise properties in robust models.

NASA's Perseverance celebrated its 100th Martian day on Tuesday since the rover put its massive wheels in the dusty landscape of the Red Planet on February 18. Perseverance, nicknamed Perky, has since hit a number of milestones that could not only help NASA find life, but also pave the way for humans to one-day walk on Mars. These achievements include recording sounds on Mars, making oxygen using carbon dioxide in the atmosphere and sending back more than 75,000 pictures of the Martian world. The car-sized vehicle also helped the US space agency fly the first powered drone, Ingenuity, on another world and is currently on a mission, exploring the Jezero Crater, to find signs of ancient microbial life. World's first all-electric AI speedboat appears to float Dominique Samuels claims it's'psychotic' to unfriend an anti-vaxxer NASA's Perseverance celebrated its 100th Martian day on Tuesday since the rover put its massive wheels in the dusty landscape of the Red Planet on February 18 Perseverance embarked on its 239 million-mile journey to Mars on July 30, 2020 from Florida's Space Coast facility. Strapped atop an Atlas V-541 rocket, the rover and its travel companion, Ingenuity, took off from Launch Complex 41 at 4:50am ET. NASA held a live briefing on February 18, as the world waited to learn if the rover and helicopter survived the'seven minutes of terror.'

Scientists track over 23,000 objects in low-Earth orbit to avoid collisions with satellites and the International Space Station.

Moving along with Kaggle's push into reinforcement learning (RL), I've been diving more into RL algorithms and at the same time working out which can realistically be trained (we don't all have an army of GPUs!). In this article I discuss how I trained an agent to play connect four at human-level in under 4 hours using Google's online notebook system Collaboratory. The article is fully replicable, you can find the notebook here, train your own agent and see for yourself just how amazing the power of AlphaZero Colab Julia is! Before diving into it, first a bit of background on the key components, if you're not interested in any, please skip ahead to the next section. AlphaZero is a reinforcement learning program first shared in 2017 by Deepmind. It is able to play Go, Shogi, and Chess, three uniquely complex strategy games.

A company-provided photo of a Kargu Rotary Wing Attack Drone Loitering Munition System manufactured by the STM defense company of Turkey. A U.N. report says the weapons system was used in Libya in March 2020. A company-provided photo of a Kargu Rotary Wing Attack Drone Loitering Munition System manufactured by the STM defense company of Turkey. A U.N. report says the weapons system was used in Libya in March 2020. Military-grade autonomous drones can fly themselves to a specific location, pick their own targets and kill without the assistance of a remote human operator.

At issue is a type of AI task called conditional image generation, in which AI systems create images that meet a specific set of conditions. For example, a system could be trained to create original images of cats or dogs, depending on which animal the user requested. More recent techniques have built on this to incorporate conditions regarding an image layout. This allows users to specify which types of objects they want to appear in particular places on the screen. For example, the sky might go in one box, a tree might be in another box, a stream might be in a separate box, and so on.

The Royal Navy is using artificial intelligence (AI) at sea for the first time to test against supersonic missile threats. The trial is part of Exercise Formidable Shield, which is currently taking place off the coast of Scotland until June 3 and is led by Naval Striking and Support Forces NATO on behalf of the US Sixth Fleet. Research, led by Defence Science and Technology Laboratory (DSTL) scientists, finds that AI accelerates engagement timelines, improves early detection of lethal threat, and provides Royal Navy Commanders with a rapid hazard assessment to select the optimum weapon or measure to counter and destroy the target. The Type 45 Destroyer (HMS Dragon) and Type 23 Frigate (HMS Lancaster) are testing two AI applications, Startle and Sycoiea. It is noted that the Startle system is designed to provide live recommendations, ease the load on sailors monitoring the'Air Picture' in the Operations Room, while Sycoiea system helps in identifying the nearest threat and how best to deal with it. These AI-based applications are being trailed to ensure that they work along with the existing radar and combat management systems.

HALIFAX, NS, June 1, 2021 /CNW/ - Global Spatial Technology Solutions ("GSTS" or "the Company") an Artificial Intelligence (AI) and Maritime Analytics company, announced today that it has been selected by Defence Research and Development Canada (DRDC) to provide advanced Maritime Risk Detection and Assessment capabilities in support of maritime border security and surveillance. The solution will identify ships in an area of interest and using the cutting-edge techniques of artificial intelligence and machine learning, consolidate a ship's identity, movement history, and risk status with information collected from multiple sensors. Fusing the intelligence into a single operating picture, GSTS's solution enables users to improve Maritime Domain Awareness. The total contract is funded under the Canadian Safety and Security Program. This powerful solution will leverage OCIANA, an AI-based platform developed by GSTS that rapidly processes data from multiple sensor sources to provide intelligence in near real-time.

Germany is the first country in the world to legalize fully autonomous vehicles on public roads. The landmark legislation passed both the lower and upper chambers of Germany's parliament last week with a comfortable majority. The legislation will allow Level 4 autonomous vehicles to operate on public roads in Germany without drivers behind the wheel and without obtaining special permits. The new legal framework mandates that autonomous vehicles must be manufactured and maintained in accordance with new, yet-to-be-crafted technical requirements. In addition, the law calls for technical command centers where live supervisors will oversee each fleet of self-driving cars and have the ability to control and deactivate them remotely when problems arise.

The difficulty and expense of collecting river water samples in remote areas has led to significant -- and in some cases, decades-long -- gaps in available water chemistry data, according to a Penn State-led team of researchers. The team is using artificial intelligence (AI) to predict water quality and fill the gaps in the data. Their efforts could lead to an improved understanding of how rivers react to human disturbances and climate change. The researchers developed a model that forecasts dissolved oxygen (DO), a key indicator of water's capability to support aquatic life, in lightly monitored watersheds across the United States. They published their results in Environmental Science & Technology.