Lifeboat Foundation

A quintillion calculations a second. That’s one with 18 zeros after it. It’s the speed at which an exascale supercomputer will process information. The Department of Energy (DOE) is preparing for the first exascale computer to be deployed in 2021. Two more will follow soon after. Yet quantum computers may be able to complete more complex calculations even faster than these up-and-coming exascale computers. But these technologies complement each other much more than they compete.

It’s going to be a while before quantum computers are ready to tackle major scientific research questions. While quantum researchers and scientists in other areas are collaborating to design quantum computers to be as effective as possible once they’re ready, that’s still a long way off. Scientists are figuring out how to build qubits for quantum computers, the very foundation of the technology. They’re establishing the most fundamental quantum algorithms that they need to do simple calculations. The hardware and algorithms need to be far enough along for coders to develop operating systems and software to do scientific research. Currently, we’re at the same point in quantum computing that scientists in the 1950s were with computers that ran on vacuum tubes. Most of us regularly carry computers in our pockets now, but it took decades to get to this level of accessibility.

In contrast, exascale computers will be ready next year. When they launch, they’ll already be five times faster than our fastest computer – Summit, at Oak Ridge National Laboratory’s Leadership Computing Facility, a DOE Office of Science user facility. Right away, they’ll be able to tackle major challenges in modeling Earth systems, analyzing genes, tracking barriers to fusion, and more. These powerful machines will allow scientists to include more variables in their equations and improve models’ accuracy. As long as we can find new ways to improve conventional computers, we’ll do it.

The context: Studies show that when people and AI systems work together, they can outperform either one acting alone. Medical diagnostic systems are often checked over by human doctors, and content moderation systems filter what they can before requiring human assistance. But algorithms are rarely designed to optimize for this AI-to-human handover. If they were, the AI system would only defer to its human counterpart if the person could actually make a better decision.

The research: Researchers at MIT’s Computer Science and AI Laboratory (CSAIL) have now developed an AI system to do this kind of optimization based on strengths and weaknesses of the human collaborator. It uses two separate machine-learning models; one makes the actual decision, whether that’s diagnosing a patient or removing a social media post, and one predicts whether the AI or human is the better decision maker.

The latter model, which the researchers call “the rejector,” iteratively improves its predictions based on each decision maker’s track record over time. It can also take into account factors beyond performance, including a person’s time constraints or a doctor’s access to sensitive patient information not available to the AI system.

“We are impressed with the early results demonstrated as we scale Loihi to create more powerful neuromorphic systems. Pohoiki Beach will now be available to more than 60 ecosystem partners, who will use this specialized system to solve complex, compute-intensive problems.” –Rich Uhlig, managing director of Intel Labs

Why It’s Important: With the introduction of Pohoiki Beach, researchers can now efficiently scale up novel neural-inspired algorithms — such as sparse coding, simultaneous localization and mapping (SLAM), and path planning — that can learn and adapt based on data inputs. Pohoiki Beach represents a major milestone in Intel’s neuromorphic research, laying the foundation for Intel Labs to scale the architecture to 100 million neurons later this year.

Deepfakes are the most concerning use of AI for crime and terrorism, according to a new report from University College London.

The research team first identified 20 different ways AI could be used by criminals over the next 15 years. They then asked 31 AI experts to rank them by risk, based on their potential for harm, the money they could make, their ease of use, and how hard they are to stop.

Deepfakes — AI-generated videos of real people doing and saying fictional things — earned the top spot for two major reasons. Firstly, they’re hard to identify and prevent. Automated detection methods remain unreliable and deepfakes also getting better at fooling human eyes. A recent Facebook competition to detect them with algorithms led researchers to admit it’s “very much an unsolved problem.”

Swarm intelligence (SI) is concerned with the collective behaviour that emerges from decentralised self-organising systems, whilst swarm robotics (SR) is an approach to the self-coordination of large numbers of simple robots which emerged as the application of SI to multi-robot systems. Given the increasing severity and frequency of occurrence of wildfires and the hazardous nature of fighting their propagation, the use of disposable inexpensive robots in place of humans is of special interest. This paper demonstrates the feasibility and potential of employing SR to fight fires autonomously, with a focus on the self-coordination mechanisms for the desired firefighting behaviour to emerge. Thus, an efficient physics-based model of fire propagation and a self-organisation algorithm for swarms of firefighting drones are developed and coupled, with the collaborative behaviour based on a particle swarm algorithm adapted to individuals operating within physical dynamic environments of high severity and frequency of change. Numerical experiments demonstrate that the proposed self-organising system is effective, scalable and fault-tolerant, comprising a promising approach to dealing with the suppression of wildfires – one of the world’s most pressing challenges of our time.

The presence of outliers in a classification or regression dataset can result in a poor fit and lower predictive modeling performance.

Identifying and removing outliers is challenging with simple statistical methods for most machine learning datasets given the large number of input variables. Instead, automatic outlier detection methods can be used in the modeling pipeline and compared, just like other data preparation transforms that may be applied to the dataset.

In this tutorial, you will discover how to use automatic outlier detection and removal to improve machine learning predictive modeling performance.

A quintillion calculations a second. That’s one with 18 zeros after it. It’s the speed at which an exascale supercomputer will process information. The Department of Energy (DOE) is preparing for the first exascale computer to be deployed in 2021. Two more will follow soon after. Yet quantum computers may be able to complete more complex calculations even faster than these up-and-coming exascale computers. But these technologies complement each other much more than they compete.

It’s going to be a while before quantum computers are ready to tackle major scientific research questions. While quantum researchers and scientists in other areas are collaborating to design quantum computers to be as effective as possible once they’re ready, that’s still a long way off. Scientists are figuring out how to build qubits for quantum computers, the very foundation of the technology. They’re establishing the most fundamental quantum algorithms that they need to do simple calculations. The hardware and algorithms need to be far enough along for coders to develop operating systems and software to do scientific research. Currently, we’re at the same point in quantum computing that scientists in the 1950s were with computers that ran on vacuum tubes. Most of us regularly carry computers in our pockets now, but it took decades to get to this level of accessibility.

In contrast, exascale computers will be ready next year. When they launch, they’ll already be five times faster than our fastest computer —Summit, at Oak Ridge National Laboratory’s Leadership Computing Facility, a DOE Office of Science user facility. Right away, they’ll be able to tackle major challenges in modeling Earth systems, analyzing genes, tracking barriers to fusion, and more. These powerful machines will allow scientists to include more variables in their equations and improve models’ accuracy. As long as we can find new ways to improve conventional computers, we’ll do it.

Researchers from the Institute of Industrial Science at The University of Tokyo designed and built specialized computer hardware consisting of stacks of memory modules arranged in a 3D-spiral for artificial intelligence (AI) applications. This research may open the way for the next generation of energy-efficient AI devices.

Machine learning is a type of AI that allows computers to be trained by example data to make predictions for new instances. For example, a smart speaker algorithm like Alexa can learn to understand your voice commands, so it can understand you even when you ask for something for the first time. However, AI tends to require a great deal of electrical energy to train, which raises concerns about adding to climate change.

Now, scientists from the Institute of Industrial Science at The University of Tokyo have developed a novel design for stacking resistive random-access memory modules with oxide semiconductor (IGZO) access transistor in a three-dimensional spiral. Having on-chip nonvolatile memory placed close to the processors makes the machine learning training process much faster and more energy-efficient. This is because electrical signals have a much shorter distance to travel compared with conventional computer hardware. Stacking multiple layers of circuits is a natural step, since training the algorithm often requires many operations to be run in parallel at the same time.

A drone has successfully inspected a 19.4 meter high oil tank onboard a Floating Production, Storage and Offloading vessel. The video shot by the drone was interpreted in real-time by an algorithm to detect cracks in the structure.

Scout Drone Inspection and class society DNV GL have been working together to develop an autonomous drone system to overcome the common challenges of tank inspections. For the customer, costs can run into hundreds of thousands of dollars as the tank is taken out of service for days to ventilate and construct scaffolding. The tanks are also tough work environments, with surveyors often having to climb or raft into hard to reach corners. Using a drone in combination with an algorithm to gather and analyse video footage can significantly reduce survey times and staging costs, while at the same time improving surveyor safety.

Continue reading “Artificial intelligence and algorithms bring drone inspection breakthrough” »