In the News
Face-mask recognition has arrived—for better or worse
How to scan a face mask Today’s facial recognition software studies the features around the eye, nose, mouth, and ears to identify an individual whose picture is already supplied, either by the individual or in a criminal database.
Sponsor
451 Research Report on Orca Security’s Light, Agentless Approach to Cloud Security
Orca’s combination of SaaS delivery, SideScanning technology, and access to cloud configuration APIs provides security visibility and context into different aspects of cloud security with less friction than agent-based approaches.
In The News
A Pandemic AI Engine Without Borders
Researchers worldwide build a decentralized learning model to improve COVID-19 diagnosis.
Is the Internet Conscious? If It Were, How Would We Know?
There’s a lot of discussion about artificial consciousness and the possibility of machines gaining self-awareness once they become sufficiently complex.
Applied use cases
VisionLabs face biometrics and body temperature scanning deployed by McDonald’s Russia
McDonald’s Russia has installed facial recognition and body temperature scanning technology from VisionLabs to scan customers to prevent the spread of COVID-19
It’s Time for a Reckoning About This Foundational Piece of Police Technology
This article is part of the Policing and Technology Project, a collaboration between Future Tense and the Tech, Law, & Security Program at American University Washington College of Law that examines the relationship between law enforcement, police reform, and technology.
Approximately Correct Machine Intelligence (ACMI) Lab
Building intelligent systems applicable in the real world requires more than prediction. Driving decisions requires causal insights. Reliability requires models that are provably robust under clear assumptions.
Ethics
AI ethics groups are repeating one of society’s classic mistakes
If organizations working on global AI ethics fail to acknowledge this, they risk developing standards that are, at best, meaningless and ineffective across all the world’s regions.
Andrew Critch on AI Research Considerations for Human Existential Safety
In this episode of the AI Alignment Podcast, Andrew Critch joins us to discuss a recent paper he co-authored with David Krueger titled AI Research Considerations for Human Existential Safety.
Repeating society's classic mistakes in AI ethics, contesting ML benchmarks, cognitive science of fake news, insurance for the gig economy, and more ...
In featured work from our staff this week, Abhishek and Victoria published a piece for the MIT Technology Review on how we might be repeating some of society’s classic mistakes in the domain of AI ethics, our report on publication norms for responsible AI prepared for Partnership on AI, a podcast...
Robotics
Robot takes contact-free measurements of patients’ vital signs
“We thought it should be possible for us to use a robot to remove the health care worker from the risk of directly exposing themselves to the patient.” Using four cameras mounted on a dog-like robot developed by Boston Dynamics, the researchers have shown that they can measure skin temperature,...
Boston Dynamics CEO talks profitability and the company’s next robots
When we interviewed Boston Dynamics founder and former CEO Marc Raibert in November 2019, we discussed the company’s customers, potential applications, AI, simulation, and those viral videos.
Global Supply Chain Disruptions Driving Robotics Adoption
Over the past few decades, many corporations in advanced economies took advantage of low trade barriers and globally integrated supply chains to outsource manual tasks to workers in countries with lower labor costs.
Research
Optimizing with constraints: reparametrization and geometry.
In particular, we show that mirror descent is equivalent to gradient descent on a reparametrized objective with straight-through gradients: replacing a constrained variable x with some squishing function σ(u), but treating σ as if it were the identity in the backward pass.
Aligning AI With Shared Human Values
This requires connecting physical and social world knowledge to value judgements, a capability that may enable us to filter out needlessly inflammatory chatbot outputs or eventually regularize open-ended reinforcement learning agents.
ELLIS inaugurates 30 research units at leading institutions across Europe
Tübingen, September 15, 2020 – The European Laboratory for Learning and Intelligent Systems (ELLIS) today inaugurated 30 research units at a virtual ceremony.