Lifeboat Foundation

Summary: Machine learning algorithm produced fewer decision-making errors than professionals when it came to clinical diagnosis of patients.

Source: University of Montreal.

It’s an old adage: there’s no harm in getting a second opinion. But what if that second opinion could be generated by a computer, using artificial intelligence? Would it come up with better treatment recommendations than your professional proposes?

A new type of artificial intelligence (AI) algorithm, developed by the Mayo Clinic and the Google Research Brain Team, can potentially pave the way toward more directed brain stimulation for the treatment of Parkinson’s disease and other movement-related disorders.

According to researchers, this algorithm can more accurately determine the interaction between different regions of the brain — data that will be key for improving the way brain stimulation devices are used in the real world for treating Parkinson’s.

“Our findings show that this new type of algorithm may help us understand which brain regions directly interact with one another, which in turn may help guide placement of electrodes for stimulating devices to treat network brain diseases,” Kai Miller, MD, PhD, a neurosurgeon at Mayo Clinic and the first author of the study, said in a press release.

Now DeepMind has set its sights on another grand challenge: bridging the worlds of deep learning and classical computer science to enable deep learning to do everything. If successful, this approach could revolutionize AI and software as we know them.

Petar Veličković is a senior research scientist at DeepMind. His entry into computer science came through algorithmic reasoning and algorithmic thinking using classical algorithms. Since he started doing deep learning research, he has wanted to reconcile deep learning with the classical algorithms that initially got him excited about computer science.

Continue reading “DeepMind aims to marry deep learning and classic algorithms” »

Circa 2012

Quantum ground-state problems are computationally hard problems for general many-body Hamiltonians; there is no classical or quantum algorithm known to be able to solve them efficiently. Nevertheless, if a trial wavefunction approximating the ground state is available, as often happens for many problems in physics and chemistry, a quantum computer could employ this trial wavefunction to project the ground state by means of the phase estimation algorithm (PEA). We performed an experimental realization of this idea by implementing a variational-wavefunction approach to solve the ground-state problem of the Heisenberg spin model with an NMR quantum simulator. Our iterative phase estimation procedure yields a high accuracy for the eigenenergies (to the 10–5 decimal digit).

What we need now is an expansion of public and private investment that does justice to the opportunity at hand. Such investments may have a longer time horizon, but their eventual impact is without parallel. I believe that net-energy gain is within reach in the next decade; commercialization, based on early prototypes, will follow in very short order.

But such timelines are heavily dependent on funding and the availability of resources. Considerable investment is being allocated to alternative energy sources — wind, solar, etc. — but fusion must have a place in the global energy equation. This is especially true as we approach the critical breakthrough moment.

If laser-driven nuclear fusion is perfected and commercialized, it has the potential to become the energy source of choice, displacing the many existing, less ideal energy sources. This is because fusion, if done correctly, offers energy that is in equal parts clean, safe and affordable. I am convinced that fusion power plants will eventually replace most conventional power plants and related large-scale energy infrastructure that are still so dominant today. There will be no need for coal or gas.

A study in which machine-learning models were trained to assess over 1 million companies has shown that artificial intelligence (AI) can accurately determine whether a startup firm will fail or become successful. The outcome is a tool, Venhound, that has the potential to help investors identify the next unicorn.

It is well known that around 90% of startups are unsuccessful: Between 10% and 22% fail within their first year, and this presents a significant risk to venture capitalists and other investors in early-stage companies. In a bid to identify which companies are more likely to succeed, researchers have developed machine-learning models trained on the historical performance of over 1 million companies. Their results, published in KeAi’s The Journal of Finance and Data Science, show that these models can predict the outcome of a company with up to 90% accuracy. This means that potentially 9 out of 10 companies are correctly assessed.

“This research shows how ensembles of non-linear machine-learning models applied to big data have huge potential to map large feature sets to business outcomes, something that is unachievable with traditional linear regression models,” explains co-author Sanjiv Das, Professor of Finance and Data Science at Santa Clara University’s Leavey School of Business in the US.

CERN Courier

Jennifer Ngadiuba and Maurizio Pierini describe how ‘unsupervised’ machine learning could keep watch for signs of new physics at the LHC that have not yet been dreamt up by physicists.

In the 1970s, the robust mathematical framework of the Standard Model ℠ replaced data observation as the dominant starting point for scientific inquiry in particle physics. Decades-long physics programmes were put together based on its predictions. Physicists built complex and highly successful experiments at particle colliders, culminating in the discovery of the Higgs boson at the LHC in 2012.

Continue reading “Hunting anomalies with an AI trigger” »

A three-qubit entangled state has been realized in a fully controllable array of spin qubits in silicon.

An all-RIKEN team has increased the number of silicon-based spin qubits that can be entangled from two to three, highlighting the potential of spin qubits for realizing multi-qubit quantum algorithms.

Quantum computers have the potential to leave conventional computers in the dust when performing certain types of calculations. They are based on quantum bits, or qubits, the quantum equivalent of the bits that conventional computers use.

Reinforcement learning (RL) is the most widely used machine learning algorithm, besides supervised and unsupervised learning and the less common self-supervised and semi-supervised learning. RL focuses on the controlled learning process, where a machine learning algorithm is provided with a set of actions, parameters, and end values. It teaches the machine trial and error.

From a data efficiency perspective, several methods have been proposed, including online setting, reply buffer, storing experience in a transition memory, etc. In recent years, off-policy actor-critic algorithms have been gaining prominence, where RL algorithms can learn from limited data sets entirely without interaction (offline RL).