Ingenuity demonstrates the ways in which the field of Artificial Intelligence has already improved in leaps and bounds. It should be noted that the helicopter is not entirely free from human intervention; it can be controlled when it needs to be. Ingenuity does have a sequenced engine and responds to commands from Earth, but for the duration that the robot is actively carrying out those commands, it runs autonomously. Whilst travelling from A to B, Ingenuity's guidance software takes over and will execute the helicopter's take off, landing and course without the need of human intervention, modelling a'half-house' AI system. The helicopter's semi-autonomous nature allows the device to be more pragmatic and address any threats or challenges that may arise: blockades and restraints to name but two.

Sonoma County is adding artificial intelligence to its wildfire-fighting arsenal. The county has entered into an agreement with the South Korean firm Alchera to outfit its network of fire-spotting cameras with software that detects wildfire activity and then alerts authorities. But emergency workers will first have to "teach" the system to differentiate between images that show fire smoke, and others that might show clouds, fog, or vapor from geothermal geysers. The software will use feedback from humans to refine its algorithm and will eventually be able to detect fires on its own -- or at least that's what county officials hope. "It's kind of like learning how to read," Godley said.

Over the last decade, AI, once a far-fetched theme of old-school science-fiction movies, has quickly grown into one of the most prolific emerging technologies – and, by association, one of the most recognizable buzzwords out there. Virtually every industry today – healthcare, transportation, manufacturing, agriculture, banking, retail, finance – has either implemented or is planning to implement AI in some way. And we know this because many of the companies operating in these spaces publicize their use of AI. What looks more cutting-edge than announcing your new AI-powered initiative? It's as much a marketing gimmick as it is about product and service efficiency. As with other fields, AI has surfaced as an accelerator of cybersecurity innovation over the past five years.

NASA's Mars Ingenuity helicopter could finally take to the Martian skies next month after spending its first month strapped to the Perseverance rover while it charges. The US space agency confirmed that the 30 days'test flight window' for the rotorcraft will begin'no earlier than the first week of April'. Ingenuity arrived on Mars strapped to the underside of the NASA Perseverance rover on February 18, following a hair raising '7 minutes of terror' journey to the surface. Before it can make the first flight of an aircraft on the Red Planet, Perseverance needs to'drop it off' in a clear, safe area - likely to happen in the coming weeks. The team behind the 4lb chopper are narrowing down on a launch site, that will become the first'airfield on another world' when Ingenuity makes its maiden flight. The space agency confirmed that the 30 days'test flight window' for the rotorcraft will begin'no earlier than the first week of April' NASA is set to fly where no one has flown before – Mars' atmosphere.

This tutorial is part of the 2021 Call for Code Global Challenge. In this tutorial, learn how to pull 40 years' of global satellite-based soil moisture data from the European Commission, then train a model to compute moisture trends to identify regions that have a high probability to dry out and have droughts. Copernicus is the European Union's Earth observation program, looking at our planet and its environment. It offers information services that draw from satellite Earth observation and in-situ (non-space) data. Vast amounts of global data from satellites and ground-based, airborne, and seaborne measurement systems provide information.

Coral reefs are some of the most diverse ecosystems in the world, sometimes called the "rainforests of the sea." Coral reefs are a vital part of marine life, act as guards against forces of nature and are a key source of livelihood for millions. Unfortunately, these diverse habitats are rapidly degrading. Here's how advanced technologies such as Artificial Intelligence (AI) are being used to save coral reefs. Coral reefs are natural protectors for coastlines against storms and erosion.

GMRES is a powerful numerical solver used to find solutions to extremely large systems of linear equations. These systems of equations appear in many applications in science and engineering. Here we demonstrate a real-time machine learning algorithm that can be used to accelerate the time-to-solution for GMRES. Our framework is novel in that is integrates the deep learning algorithm in an in situ fashion: the AI-accelerator gradually learns how to optimizes the time to solution without requiring user input (such as a pre-trained data set). We describe how our algorithm collects data and optimizes GMRES. We demonstrate our algorithm by implementing an accelerated (MLGMRES) solver in Python. We then use MLGMRES to accelerate a solver for the Poisson equation -- a class of linear problems that appears in may applications. Informed by the properties of formal solutions to the Poisson equation, we test the performance of different neural networks. Our key takeaway is that networks which are capable of learning non-local relationships perform well, without needing to be scaled with the input problem size, making them good candidates for the extremely large problems encountered in high-performance computing. For the inputs studied, our method provides a roughly 2$\times$ acceleration.

The ability to perceive and reason about social interactions in the context of physical environments is core to human social intelligence and human-machine cooperation. However, no prior dataset or benchmark has systematically evaluated physically grounded perception of complex social interactions that go beyond short actions, such as high-fiving, or simple group activities, such as gathering. In this work, we create a dataset of physically-grounded abstract social events, PHASE, that resemble a wide range of real-life social interactions by including social concepts such as helping another agent. PHASE consists of 2D animations of pairs of agents moving in a continuous space generated procedurally using a physics engine and a hierarchical planner. Agents have a limited field of view, and can interact with multiple objects, in an environment that has multiple landmarks and obstacles. Using PHASE, we design a social recognition task and a social prediction task. PHASE is validated with human experiments demonstrating that humans perceive rich interactions in the social events, and that the simulated agents behave similarly to humans. As a baseline model, we introduce a Bayesian inverse planning approach, SIMPLE (SIMulation, Planning and Local Estimation), which outperforms state-of-the-art feed-forward neural networks. We hope that PHASE can serve as a difficult new challenge for developing new models that can recognize complex social interactions.

In this work, we explore adversarial attacks on the Variational Autoencoders (VAE). We show how to modify data point to obtain a prescribed latent code (supervised attack) or just get a drastically different code (unsupervised attack). We examine the influence of model modifications ($\beta$-VAE, NVAE) on the robustness of VAEs and suggest metrics to quantify it.

Large-scale pre-trained language models have demonstrated strong capabilities of generating realistic text. However, it remains challenging to control the generation results. Previous approaches such as prompting are far from sufficient, which limits the usage of language models. To tackle this challenge, we propose an innovative method, inverse prompting, to better control text generation. The core idea of inverse prompting is to use generated text to inversely predict the prompt during beam search, which enhances the relevance between the prompt and the generated text and provides better controllability. Empirically, we pre-train a large-scale Chinese language model to perform a systematic study using human evaluation on the tasks of open-domain poem generation and open-domain long-form question answering. Our results show that our proposed method substantially outperforms the baselines and that our generation quality is close to human performance on some of the tasks. Narrators can try our poem generation demo at https://pretrain.aminer.cn/apps/poetry.html, while our QA demo can be found at https://pretrain.aminer.cn/app/qa. For researchers, the code is provided in https://github.com/THUDM/InversePrompting.