Recently Browsing 0 members
No registered users viewing this page.
By Ather Fawaz
New research shows that near-term quantum computers can learn to reason
by Ather Fawaz
The applications and development of quantum computers have steadily picked up pace in the last few years. We've seen researchers applying this novel method of computation in a variety of domains including quantum chemistry, fluid dynamics research, open problems, and even machine learning, all with promising results.
Continuing this trend, UK-based startup Cambridge Quantum Computing (CQC) has now demonstrated that quantum computers "can learn to reason". Confusing at first, this claim is based upon new research coming out of CQC. Dr. Mattia Fiorentini, Head of Quantum Machine Learning at the firm, and his team of researchers investigated using quantum computers for Variational Inference.
Variational Inference is a process through which we approximate a given probability distribution using stochastic optimization and other learning techniques. Jargon aside, this means a quantum computer outputs potential solutions to inferential questions such as that given the grass is wet and it's cloudy, what's the more probable cause for it? Rain or sprinklers? Formally, the question is posed as follows:
The team's work, titled Variational inference with a quantum computer, has been published on the pre-print repository arXiv and highlights what the firm believes to be a promising indicator that quantum computers are great for Variational Inference, and by extension, at reasoning.
In the paper, the researchers demonstrate their results on Bayesian Networks. Three different problem sets were tested. First, was the classic cloud-sprinkler-rain problem that was described above. Second, was the prediction of market regime switches (bull or bear) in a Hidden Markov Model of simulated financial time series. Third, was the task of inferring likely diseases in patients given some information about symptoms and risk factors.
Using adversarial training and the kernelized Stein discrepancy, the details of both which can be found in the paper, the firm optimized a classical probabilistic classifier and a probabilistic quantum model, called Born machine, in tandem.
Adversarial method Kernelized Stein discrepancy method Once trained, inference was carried out on the three problems defined earlier, both on a quantum simulator and on IBM Q's real quantum computers. In the truncated histograms shown below, the magenta bars represent the true probability distribution, blue bars indicate outputs from a quantum computing simulator, and grey bars indicate the output from real quantum hardware from IBM Q. The results on real quantum computer hardware are marred by noise and this causes slower convergence compared to the simulation. That is to be expected in the NISQ era, however.
Truncated histogram of the posterior distribution for a hidden Markov model Histogram of the posterior distribution for a medical diagnosis task The probability distribution of the quantum simulator closely resembles the true probability distribution, indicating that the quantum algorithm has trained well, and that the firm's adversarial training and the kernelized Stein discrepancy methods are powerful algorithms for the intended purpose.
The firm believes that this is yet an indicator that "sampling from complex distributions is the most promising way towards a quantum advantage with today’s noisy quantum devices." And that its new inference methods "can incorporate domain expertise". Moving forward, the firm envisions "a combination with other machine learning tasks in generative modeling or natural language processing for a meaningful impact."
Further details can be found in this blog post and the paper on arXiv. If you are interested, you can check out Dr. Mattia's interview on YouTube here.
TWIRL 6: Blue Origin prepares for a human space flight
by Paul Hill
Welcome to This Week in Rocket Launches 6, it looks set to be a bit of an interesting week with Blue Origin planning to launch its New Shepard NS-15 mission which will prep the firm for a crewed flight next time around. We’ve also got launches from SpaceX, Rocket Lab, Roscosmos, China, and ExPace which will re-attempt a mission originally slated for last week.
Monday, March 29
The first rocket we could see take off this week is ExPace’s Kuaizhou KZ-1A carrying the Jilin Gaofen 2D satellite that will join the Jilin 1 Earth observation constellation. The payload, also known as Jilin 28, is a 230 kg satellite that will be used to take photos of Earth from 535 km. The launch of this rocket is not set in stone but it could go up from Monday.
SpaceX is also looking to carry out its Starship SN11 mission on Monday. The firm was aiming for a launch last Friday but it was ultimately scrubbed. Similarly to the last two test flights, SN11 will fly to 10 km before attempting to land. During the SN10 mission, Starship did land but it also managed to catch fire which caused the ship's destruction several minutes later. The SN11 mission will be streamed by SpaceX on its YouTube channel when Starship is ready to launch.
Thursday, April 1
The first rocket that we could see launch on Thursday is Rocket Lab’s Electron. As part of its STP-27RM mission, Rocket Lab will launch an experimental payload for the U.S. Air Force. The Payload is a space weather instrument called Monolith and is part of the Space Test Program. It will demonstrate the ability of small satellites to support large aperture payloads. While the mission is eligible for launch on Thursday, it could launch later.
Another mission that will take off from Thursday is the New Shepard NS-15. Blue Origin, who makes New Shepard, is the firm owned by Amazon’s Jeff Bezos. It will be using a booster rocket called Tail 4 which itself comes equipped with an improved BE-3PM engine. The CC 2.0-2 RSS First Step capsule has been upgraded for astronauts who will also come aboard and then leave again before the rocket launch – this will prepare them for NS-16 which will be a crewed mission. You’ll be able to find a stream of the event here. As with the other launches so far, this mission is also marked with a No Earlier Than (NET) tag so it may happen after Thursday. Below is a video from the NS-14 mission:
The final launch on Thursday will come from Roscosmos who is launching a Soyuz 2.1b with the 4th Resurs-P satellite. The Resurs-P series of satellites are Earth observation satellites that are used by several Russian governmental agencies including Russia’s meteorological agency. The launch was delayed from November 2019 and November 2020.
Friday, April 2
The final launch of the week will come from China. The Long March CZ-4C will carry the Gaofen 12-02 remote sensing satellite which will perform high-res Earth observation. The satellite has a sub-meter level resolution which is suited for urban planning, crop yield estimation, and disaster prevention. The mission is being launched for the China High-definition Earth Observation System (CHEOS).
By Ather Fawaz
Lack of thrust and an ad hoc solution to SN8's explosion led to Starship SN10's fiery ending
by Ather Fawaz
Less than a week ago, SpaceX almost succeeded in completing a flawless test flight for the Starship SN10. And things were looking up until a few moments after a successful ascent and touchdown back to earth. Just like the SN8 and SN9 predecessors before it, SN10 too met a fiery, explosive ending.
It was unclear why the rocket exploded, especially after it had touched down and remained stationary for close to a minute. SpaceX founder and CEO, Elon Musk, has now cleared up some confusion on that front. Musk tweeted that "the SN10 engine was low on thrust due (probably) to partial helium ingestion from the fuel header tank." And that the prototype plummeted towards the earth at an impact velocity of 10m/s that crushed some legs and parts of the skirt.
As spotted by Engadget, Chris Bergin of NASA Spaceflight pointed out that the helium ingestion was caused by the pressurization system that was added to the CH4 tank to prevent what caused the SN8 to explode. Musk replied to Bergin stating that this was a reasonable point, and that "...if autogenous pressurization had been used, CH4 bubbles would most likely have reverted to liquid. Helium in header was used to prevent ullage collapse from slosh, which happened in prior flight. My fault for approving. Sounded good at the time."
Despite this, the SN10's touchdown represents a big step forward for the Starship program and SpaceX at large. The firm was quick on bringing the next prototype, the SN11, to the stand at Boca Chica, preparing for its test flight sometime later. As Austin Barnard photographed, the team of engineers on-site inspected every landing leg of the SN11. SpaceX hopes that the insight gained from its predecessors, including the SN10, will be used to do away with past mistakes and progress towards a successful test flight for the SN11.
Apple shares Hearing Study insights for World Hearing Day
by Paul Hill
In time for World Hearing Day, which takes place globally on March 3, Apple has shared some of its findings from a study it is carrying out on the long-term auditory health of its users in the United States. Apple said the participation levels in this study are on a scale never seen before thanks to the ease of participation; users can partake in the study using an Apple Watch and iPhone and sharing their health data.
One of the interesting findings revealed that 25% of participants experience environmental sounds higher than the World Health Organization’s (WHO) recommended limit. This noise can come from anywhere including traffic, machinery, and public transport. The WHO has estimated that 700 million people around the world will be affected by profound hearing loss by 2050 - with its Noise app on Apple Watch, the tech firm hopes to save some of its users from hearing loss.
Another source of hearing loss is the way we use headphones to listen to music. One in ten of Apple’s participants listen to sounds through headphones at a level above WHO recommendations. Apple says that to protect hearing, users ought to listen to media at the lowest enjoyable volume rather than at full blast.
The study seems to suggest that a lot of people are neglecting their ears; around 10% of participants have been diagnosed with hearing loss by a professional and over these, three-quarters do not use tools such as hearing aids or cochlear implants even though they can reduce the impact of hearing loss. Furthermore, Apple found that a huge 50% of participants hadn’t had their hearing tested by a professional in the last 10 years and a quarter of all participants reported a ringing in their ears which is a sign of hearing loss.
To take part in the study yourself, head over to the App Store and download the Apple Research app. If you’re not sure how good your hearing is or lack access to testing facilities, the World Health Organization provides an app for Android and iOS called HearWHO where you have to listen out for numbers being read out in static, this can give you an estimate about how good your hearing is.
By Ather Fawaz
Elon Musk claims that his wired-up Neuralink monkey is happy and enjoys playing video games
by Ather Fawaz
Image via The Telegraph Neuralink is Elon Musk's bold initiative to create an interface between a brain and a computer chip. Revealed back in 2017, the company was funded entirely out of Musk's own pocket. While many raised suspicions about the potential misuse of the technology, Musk insists that its raison d'etre is to cure medical injuries related to the spinal cord and brain and improve cognition and memory. Over the years, there have been a bunch of updates on the project. Perhaps most importantly, the chip was successfully implanted in a monkey, allowing it to control a computer with its brain. While the exact nature of this control wasn't entirely clear, Musk dubbed that result as "very positive."
Building on this, Musk gave a couple of added details about Neuralink's tests on monkeys. Bloomberg reports that while speaking to Clubhouse, which is a private social app where users engage in informal conversations, Musk told several thousand listeners that his company has a happy monkey with Neuralink implanted in its skull who enjoys playing video games:
When Neuralink's working was first highlighted, many people raised a concern that implanting a device in the skull would not only be a notoriously difficult procedure, but it might not look all that cosmetically pleasing either. Speaking to Joe Rogan on his podcast, the billionaire clarified that the device would sit flush with the skull. He further added on Clubhouse today that:
Going back to his plugged-in monkeys, he further indicated that their videos could be released in about a month. As with the experiments before, the exact nature and details of the controllability of the computer are unclear. The videos could clear up the confusion, and for those, we shall be on the lookout.