Lifeboat Foundation

As developers unlock new AI tools, the risk for perpetuating harmful biases becomes increasingly high — especially on the heels of a year like 2020, which reimagined many of our social and cultural norms upon which AI algorithms have long been trained.

A handful of foundational models are emerging that rely upon a magnitude of training data that makes them inherently powerful, but it’s not without risk of harmful biases — and we need to collectively acknowledge that fact.

Recognition in itself is easy. Understanding is much harder, as is mitigation against future risks. Which is to say that we must first take steps to ensure that we understand the roots of these biases in an effort to better understand the risks involved with developing AI models.

Dowsett’s algorithm was recently published in npj Breast Cancer, a Nature Partner Journal supported by the Breast Cancer Research Foundation. It is intended to help physicians triage postmenopausal women with ER+ HER2–breast cancers, which represent around 70% of breast cancer cases.1 During the pandemic, many within this patient group were prescribed neoadjuvant endocrine therapy (NeoET), rather than surgery, as a disease management strategy.

Analysis of biomarkers in biopsies helps identify breast cancer patients in need of urgent surgery or chemotherapy during COVID-19 pandemic.

Researchers at Skolkovo Institute of Science and Technology (Skoltech) in Russia have recently developed an innovative system for human-swarm interactions that allows users to directly control the movements of a team of drones in complex environments. This system, presented in a paper pre-published on arXiv is based on an interface that recognizes human gestures and adapts the drones’ trajectories accordingly.

Quadcopters, drones with four rotors that can fly for long periods of time, could have numerous valuable applications. For instance, they could be used to capture images or videos in natural or remote environments, can aid search-and–rescue missions and help to deliver goods to specific locations.

Continue reading “DronePaint: A human-swarm interaction system for environment exploration and artistic painting” »

New chip eliminates the need for specific decoding hardware, could boost efficiency of gaming systems, 5G networks, the internet of things, and more.

A new silicon chip can decode any error-correcting code through the use of a novel algorithm known as Guessing Random Additive Noise Decoding (GRAND). The work was led by Muriel Médard, an engineering professor in the MIT Research Laboratory of Electronics.

Is an academic doctor and medical technology entrepreneur, working in the field of the computational biology of aging.

Dr. Radenkovic is also a Partner at the SALT Bio-Fund, and a co-founder of Hooke, an elite longevity research clinic in London.

Continue reading “Dr. Dina Radenkovic, MD — Longevity Physician, Med-Tech Entrepreneur, Thought Leader, Financier” »

Circa 2021

A crime is a deliberate act that can cause physical or psychological harm, as well as property damage or loss, and can lead to punishment by a state or other authority according to the severity of the crime. The number and forms of criminal activities are increasing at an alarming rate, forcing agencies to develop efficient methods to take preventive measures. In the current scenario of rapidly increasing crime, traditional crime-solving techniques are unable to deliver results, being slow paced and less efficient. Thus, if we can come up with ways to predict crime, in detail, before it occurs, or come up with a “machine” that can assist police officers, it would lift the burden of police and help in preventing crimes. To achieve this, we suggest including machine learning (ML) and computer vision algorithms and techniques.

Learning how to learn is something most humans do well, by leveraging previous experiences to inform the learning processes for new tasks. Endowing AI systems with such abilities however remains challenging, as it requires the machine learners to learn update rules, which typically have been manually tuned for each task.

The field of meta-learning studies how to enable machine learners to learn how to learn, and is a critical research area for improving the efficiency of AI agents. One of the approaches is for learners to learn an update rule by applying it on previous steps and then evaluating the corresponding performance.

To fully unlock the potential of meta-learning, it is necessary to overcome both the meta-optimization problem and myopic meta objectives. To tackle these issues, a research team from DeepMind has proposed an algorithm designed to enable meta-learners to teach themselves.

On Oct. 7 2015, Perimeter Institute Director Neil Turok opened the 2015/16 season of the PI Public Lecture Series with a talk about the remarkable simplicity that underlies nature. Turok discussed how this simplicity at the largest and tiniest scales of the universe is pointing toward new avenues of physics research and could lead to revolutionary advances in technology.

Perimeter Institute (charitable registration number 88,981 4323 RR0001) is the world’s largest independent research hub devoted to theoretical physics, created to foster breakthroughs in the fundamental understanding of our universe, from the smallest particles to the entire cosmos. The Perimeter Institute Public Lecture Series is made possible in part by the support of donors like you. Be part of the equation: https://perimeterinstitute.ca/inspiring-and-educating-public.

Continue reading “Neil Turok Public Lecture: The Astonishing Simplicity of Everything” »

If travel to distant stars within an individual’s lifetime is going to be possible, a means of faster-than-light propulsion will have to be found. To date, even recent research about superluminal (faster-than-light) transport based on Einstein’s theory of general relativity would require vast amounts of hypothetical particles and states of matter that have “exotic” physical properties such as negative energy density. This type of matter either cannot currently be found or cannot be manufactured in viable quantities. In contrast, new research carried out at the University of Göttingen gets around this problem by constructing a new class of hyper-fast ‘solitons’ using sources with only positive energies that can enable travel at any speed. This reignites debate about the possibility of faster-than-light travel based on conventional physics. The research is published in the journal Classical and Quantum Gravity.

The author of the paper, Dr Erik Lentz, analysed existing research and discovered gaps in previous ‘warp drive’ studies. Lentz noticed that there existed yet-to-be explored configurations of space-time curvature organized into ‘solitons’ that have the potential to solve the puzzle while being physically viable. A soliton — in this context also informally referred to as a ‘warp bubble’ — is a compact wave that maintains its shape and moves at constant velocity. Lentz derived the Einstein equations for unexplored soliton configurations (where the space-time metric’s shift vector components obey a hyperbolic relation), finding that the altered space-time geometries could be formed in a way that worked even with conventional energy sources. In essence, the new method uses the very structure of space and time arranged in a soliton to provide a solution to faster-than-light travel, which — unlike other research — would only need sources with positive energy densities.