Lifeboat Foundation

If you’ve ever experienced jet lag, you are familiar with your circadian rhythm, which manages nearly all aspects of metabolism, from sleep-wake cycles to body temperature to digestion. Every cell in the body has a circadian clock, but researchers were unclear about how networks of cells connect with each other over time and how those time-varying connections impact network functions.

In research published Aug. 27 in PNAS, researchers at Washington University in St. Louis and collaborating institutions developed a unified, data-driven computational approach to infer and reveal these connections in biological and chemical oscillatory networks, known as the topology of these complex networks, based on their time-series data. Once they establish the topology, they can infer how the agents, or cells, in the network work together in synchrony, an important state for the brain. Abnormal synchrony has been linked to a variety of brain disorders, such as epilepsy, Alzheimer’s disease and Parkinson’s disease.

Jr-Shin Li, professor of systems science & mathematics and an applied mathematician in the School of Engineering & Applied Science, developed an algorithm, called the ICON (infer connections of networks) method, that shows for the first time the strength of these connections over time. Previously, researchers could only determine whether a connection existed between networks.

The algorithm is saving about $10 million as part of an effort to replace the city’s water infrastructure.

To catch you up: In 2014, Flint began getting water from Flint River rather than the Detroit water system. Mistreatment of the new water supply, combined with old lead pipes, created contaminated water for residents.

Solving the problem: Records that could be used to figure out which houses might be affected by corroded old pipes were missing or incomplete. So the city turned to AI. Using 71 different pieces of information—like the age or value of the home—Georgia Tech researchers developed an algorithm that predicted whether or not a home was connected to lead pipes.

Continue reading “AI is helping find lead pipes in Flint, Michigan” »

One of the most significant AI milestones in history was quietly ushered into being this summer. We speak of the quest for Artificial General Intelligence (AGI), probably the most sought-after goal in the entire field of computer science. With the introduction of the Impala architecture, DeepMind, the company behind AlphaGo and AlphaZero, would seem to finally have AGI firmly in its sights.

Let’s define AGI, since it’s been used by different people to mean different things. AGI is a single intelligence or algorithm that can learn multiple tasks and exhibits positive transfer when doing so, sometimes called meta-learning. During meta-learning, the acquisition of one skill enables the learner to pick up another new skill faster because it applies some of its previous “know-how” to the new task. In other words, one learns how to learn — and can generalize that to acquiring new skills, the way humans do. This has been the holy grail of AI for a long time.

Continue reading “Artificial General Intelligence Is Here, and Impala Is Its Name” »

Level 4 – Awareness + World model: Systems that have a modeling system complex enough to create a world model: a sense of other, without a sense of self – e.g., dogs. Level 4 capabilities include static behaviors and rudimentary learned behavior.

Level 5 – Awareness + World model + Primarily subconscious self model = Sapient or Lucid: Lucidity means to be meta-aware – that is, to be aware of one’s own awareness, aware of abstractions, aware of one’s self, and therefore able to actively analyze each of these phenomena. If a given animal is meta-aware to any extent, it can therefore make lucid decisions. Level 5 capabilities include the following: The “sense of self”; Complex learned behavior; Ability to predict the future emotional states of the self (to some degree); The ability to make motivational tradeoffs.

Level 6 – Awareness + World model + Dynamic self model + Effective control of subconscious: The dynamic sense of self can expand from “the small self” (directed consciousness) to the big self (“social group dynamics”). The “self” can include features that cross barriers between biological and non-biological – e.g., features resulting from cybernetic additions, like smartphones.

Continue reading “The Transhumanist Bill of Rights version 2.0” »

Quantum computers aren’t yet practical, but Microsoft has already developed a programming language for them. Q# works inside Visual Studio, just like most other languages, and could offer aspiring programmers a chance to learn the basics of quantum physics through trial-and-error.

Engineers at Caltech have developed a new control algorithm that enables a single drone to herd an entire flock of birds away from the airspace of an airport. The algorithm is presented in a study in IEEE Transactions on Robotics.

The project was inspired by the 2009 “Miracle on the Hudson,” when US Airways Flight 1549 struck a flock of geese shortly after takeoff and pilots Chesley Sullenberger and Jeffrey Skiles were forced to land in the Hudson River off Manhattan.

“The passengers on Flight 1549 were only saved because the pilots were so skilled,” says Soon-Jo Chung, an associate professor of aerospace and Bren Scholar in the Division of Engineering and Applied Science as well as a JPL research scientist, and the principal investigator on the drone herding project. “It made me think that next time might not have such a happy ending. So I started looking into ways to protect airspace from birds by leveraging my research areas in autonomy and robotics.”

Continue reading “Engineers teach a drone to herd birds away from airports autonomously” »

The state of artificial intelligence (AI) in smart homes nowadays might be likened to a smart but moody teenager: It’s starting to hit its stride and discover its talents, but it doesn’t really feel like answering any questions about what it’s up to and would really rather be left alone, OK?

William Yeoh, assistant professor of computer science and engineering in the School of Engineering & Applied Science at Washington University in St. Louis, is working to help smart-home AI to grow up.

The National Science Foundation (NSF) awarded Yeoh a $300,000 grant to assist in developing smart-home AI algorithms that can determine what a user wants by both asking questions and making smart guesses, and then plan and schedule accordingly. Beyond being smart, the system needs to be able to communicate and to explain why it is proposing the schedule it proposed to the user.

There is. Our engagement with AI will transform us. Technology always does, even while we are busy using it to reinvent our world. The introduction of the machine gun by Richard Gatling during America’s Civil War, and its massive role in World War I, obliterated our ideas of military gallantry and chivalry and emblazoned in our minds Wilfred Owen’s imagery of young men who “die as Cattle.” The computer revolution beginning after World War II ushered in a way of understanding and talking about the mind in terms of hardware, wiring and rewiring that still dominates neurology. How will AI change us? How has it changed us already? For example, what does reliance on navigational aids like Waze do to our sense of adventure? What happens to our ability to make everyday practical judgments when so many of these judgments—in areas as diverse as credit worthiness, human resources, sentencing, police force allocation—are outsourced to algorithms? If our ability to make good moral judgments depends on actually making them—on developing, through practice and habit, what Aristotle called “practical wisdom”—what happens when we lose the habit? What becomes of our capacity for patience when more and more of our trivial interests and requests are predicted and immediately met by artificially intelligent assistants like Siri and Alexa? Does a child who interacts imperiously with these assistants take that habit of imperious interaction to other aspects of her life? It’s hard to know how exactly AI will alter us. Our concerns about the fairness and safety of the technology are more concrete and easier to grasp. But the abstract, philosophical question of how AI will impact what it means to be human is more fundamental and cannot be overlooked. The engineers are right to worry. But the stakes are higher than they think.

In Machine Learning one of the biggest problem faced by the practitioners in the process is choosing the correct set of hyper-parameters. And it takes a lot of time in tuning them accordingly, to stretch the accuracy numbers.

For instance lets take, SVC from well known library Scikit-Learn, class implements the Support Vector Machine algorithm for classification which contains more than 10 hyperparameters, now adjusting all ten to minimize the loss is very difficult just by using hit and trial. Though Scikit-Learn provides Grid Search and Random Search, but the algorithms are brute force and exhaustive, however hyperopt implements distributed asynchronous algorithm for hyperparameter optimization.