Lifeboat Foundation

In the ongoing effort to scale AI systems without incurring prohibitively high training and compute costs, sparse mixture-of-expert models (MoE) have shown their potential for achieving impressive neural network pretraining speedups by dynamically selecting only the related parameters for each input. This enables such networks to vastly expand their parameters while keeping their FLOPs per token (compute) roughly constant. Advancing MoE models to state-of-the-art performance has however been hindered by training instabilities and uncertain quality during fine-tuning.

To address these issues, a research team from Google AI and Google Brain has published a set of guidelines for designing more practical and reliable sparse expert models. The team tested their recommendations by pretraining a 269B sparse model, which it says is the first to achieve state-of-the-art results on natural language processing (NLP) benchmarks.

The team summarizes their main contributions as:

ATLAS has already spotted more than 66 comets and 700 near-Earth asteroids (two of which actually hit Earth’s atmosphere!).

Now, a new algorithm developed by Brown University bioengineers could be an important step toward such adaptive DBS. The algorithm removes a key hurdle that makes it difficult for DBS systems to sense brain signals while simultaneously delivering stimulation.

“We know that there are electrical signals in the brain associated with disease states, and we’d like to be able to record those signals and use them to adjust neuromodulation therapy automatically,” said David Borton, an assistant professor of biomedical engineering at Brown and corresponding author of a study describing the algorithm. “The problem is that stimulation creates electrical artifacts that corrupt the signals we’re trying to record. So we’ve developed a means of identifying and removing those artifacts, so all that’s left is the signal of interest from the brain.”

Despite having remarkable utility in treating movement disorders such as Parkinson’s disease, deep brain stimulation (DBS) has confounded researchers, with a general lack of understanding of why it works at some frequencies and does not at others. Now a University of Houston biomedical engineer is presenting evidence in Nature Communications Biology that electrical stimulation of the brain at higher frequencies (100Hz) induces resonating waveforms which can successfully recalibrate dysfunctional circuits causing movement symptoms.

“We investigated the modulations in local field potentials induced by electrical stimulation of the subthalamic nucleus (STN) at therapeutic and non-therapeutic frequencies in Parkinson’s disease patients undergoing DBS surgery. We find that therapeutic high-frequency stimulation (130−180 Hz) induces high-frequency oscillations (~300 Hz, HFO) similar to those observed with pharmacological treatment,” reports Nuri Ince, associate professor of biomedical engineering.

For the past couple of decades, deep brain stimulation (DBS) has been the most important therapeutic advancement in the treatment of Parkinson’s disease, a progressive nervous system disorder that affects movement in 10 million people worldwide. In DBS, electrodes are surgically implanted in the deep brain and electrical pulses are delivered at certain rates to control tremors and other disabling motor signs associated with the disease.

Recordings of neural activity during therapeutic stimulation can be used to predict subsequent changes in brain connectivity, according to a study of epilepsy patients published in JNeurosci. This approach could inform efforts to improve brain stimulation treatments for depression and other psychiatric disorders.

Corey Keller and colleagues delivered electrical stimulation from implanted electrodes in 14 patients while recording participants’ brain activity.

Repeated sets of stimulation resulted in progressive changes to the brain’s response to simulation, with stronger responses in brain regions connected to the stimulation site.

Our complicated emotional lives can often feel like a prison. Insecurities, depression and anxiety can all hold us back in life. But what if we could just eliminate the mental states that we don’t want? Or enhance the moods we do? There’s every reason to believe that this may be commonplace in the future. In fact, a lot of the technology that could achieve this already exists.

More than half of us will have experienced an extended period of sadness or low mood during our lives, and about a fifth will have been diagnosed with major depression, although these figures depend a lot on the culture in which you live. The fact that mood disorders are so common – and also so difficult to treat – means that research into the future of mood modulation is constantly evolving.

If you go to a doctor in the UK with suspected depression today, you will start on a pathway of care including “talking cures” such as cognitive behavioural therapy, or drug treatments including serotonin re-uptake inhibitors like Prozac. People who do not respond to these treatments may progress to heavier regimes or combinations of drug treatments. Since most psychoactive drug treatments are associated with side effects, there is pressure to develop new treatment options that are better tolerated by most people.

Stress-susceptible animals that behaved as if they were depressed or anxious were restored to relatively normal behavior by tweaking the system, according to a study appearing in the July 20 issue of Neuron.

“If you ‘turn the volume up’ on animals that hadn’t experienced stress, they start normal and then they have a problem,” said lead researcher Kafui Dzirasa, an assistant professor of psychiatry and behavioral sciences, and neurobiology. “But in the animals that had experienced stress and didn’t do well with it, you had to turn their volume up to get them back to normal. It looked like stress had turned the volume down.”

Artificial Intelligence has made tremendous progress these past few years, but what are the biggest AI Researchers expecting Artificial Intelligence to look like in the year 2030. They’ve made some amazing futurism predictions on all the amazing human abilities and efficiencies these AI’s will likely have. Human level AI will likely be a thing and other technology predictions like the metaverse will likely turn out to be true aswell.
–
TIMESTAMPS:
00:00 The Future of Artificial Intelligence.
00:53 Artificial Intelligence Predictions for 2030
02:15 The Metaverse in 2030
04:00 Other Technology Predictions for 2030
06:41 Last Words.
–
#future #ai #2030

Nobel Prize in physics in 2018 and professor emeritus at the École polytechnique, Gérard Mourou is a scientist that nothing can stop. After revolutionizing ophthalmic surgery with the invention of a new laser technique, the physicist launched a challenging scientific project, which only a researchers of this fame could imagine: the transmutation of radioactive waste by high-power laser. Andra met him to find out more.

It is on the plateau of Saclay, south of Paris, that we meet Gérard Mourou. Here at École Polytechnique, the Nobel Prize in Physics has been working in his laboratory for many years. His enthusiasm remains intact when it comes to addressing the issue of lasers. His research on the subject represents the project of a lifetime. “For a long time, the power of lasers was limited, due to the risk of destroying them. Alongside Donna Strickland, with whom I share the Nobel Prize, we invented the technique of CPA (Chirped Pulse Amplification): the laser emits an ultrashort pulse that we will stretch a colossal factor before amplifying it. Thanks to the CPA one can produce considerable power, to the order of the petawatt (10e15W), without destroying the laser. This represents the equivalent of a hundred times the world electricity grid, ” explains Gérard Mourou.

For the physicist, this new invention opens perspectives in several areas, starting with ophthalmic surgery. An application that came to light as a result of an unlikely combination of circumstances: One of my students was aligning the laser for an experiment when it got the pulse in the eye. We went to the hospital where an intern found that the damage to the retina was absolutely perfect. This laser was the cleanest knife possible…