Toggle light / dark theme

Computational molecular physics (CMP) aims to leverage the laws of physics to understand not just static structures but also the motions and actions of biomolecules. Applying CMP to proteins has required either simplifying the physical models or running simulations that are shorter than the time scale of the biological activity. Brini et al. reviewed advances that are moving CMP to time scales that match biological events such as protein folding, ligand unbinding, and some conformational changes. They also highlight the role of blind competitions in driving the field forward. New methods such as deep learning approaches are likely to make CMP an increasingly powerful tool in describing proteins in action.

Science, this issue p.

### BACKGROUND

The biggest computer chip in the world is so fast and powerful it can predict future actions “faster than the laws of physics produce the same result.”

That’s according to a post by Cerebras Systems, a that made the claim at the online SC20 supercomputing conference this week.

Working with the U.S. Department of Energy’s National Energy Technology Laboratory, Cerebras designed what it calls “the world’s most powerful AI compute system.” It created a massive chip 8.5 inch-square chip, the Cerebras CS-1, housed in a refrigerator-sized computer in an effort to improve on deep-learning training models.

Code Unto Caesar

Durendal’s algorithm wrote scripture about three topics: “the plague,” “Caesar,” and “the end of days.” So it’s not surprising that things took a grim turn. The full text is full of glitches characteristic of AI-written texts, like excerpts where over half of the nouns are “Lord.” But some passages are more coherent and read like bizarre doomsday prophecies.

For example, from the plague section: “O LORD of hosts, the God of Israel; When they saw the angel of the Lord above all the brethren which were in the wilderness, and the soldiers of the prophets shall be ashamed of men.”

Whole-body positron emission tomography combined with computed tomography (PET/CT) is a cornerstone in the management of lymphoma (cancer in the lymphatic system). PET/CT scans are used to diagnose disease and then to monitor how well patients respond to therapy. However, accurately classifying every single lymph node in a scan as healthy or cancerous is a complex and time-consuming process. Because of this, detailed quantitative treatment monitoring is often not feasible in clinical day-to-day practice.

Researchers at the University of Wisconsin-Madison have recently developed a deep-learning model that can perform this task automatically. This could free up valuable physician time and make quantitative PET/CT treatment monitoring possible for a larger number of patients.

To acquire PET/CT scans, patients are injected with a sugar molecule marked with radioactive fluorine-18 (18 F-fluorodeoxyglucose). When the fluorine atom decays, it emits a positron that instantly annihilates with an electron in its immediate vicinity. This annihilation process emits two back-to-back photons, which the scanner detects and uses to infer the location of the radioactive decay.

Artificial intelligence is being developed that can analyze whether it’s own decision or prediction is reliable.

…An AI that is aware/determine or analyze it’s own weaknesses. Basically, it should help doctors or passengers of the AI know quickly the risk involved.


How might The Terminator have played out if Skynet had decided it probably wasn’t responsible enough to hold the keys to the entire US nuclear arsenal? As it turns out, scientists may just have saved us from such a future AI-led apocalypse, by creating neural networks that know when they’re untrustworthy.

These deep learning neural networks are designed to mimic the human brain by weighing up a multitude of factors in balance with each other, spotting patterns in masses of data that humans don’t have the capacity to analyse.

While Skynet might still be some way off, AI is already making decisions in fields that affect human lives like autonomous driving and medical diagnosis, and that means it’s vital that they’re as accurate as possible. To help towards this goal, this newly created neural network system can generate its confidence level as well as its predictions.

Scholars have a nifty way of alerting colleagues to lengthy treatises that they find simply not worth their time to read.

They tag such documents “tl;dr”—too long, didn’t read.

It’s kind of a 21st century spin on the 420-year-old notion Shakespeare’s Polonius relayed to the king and queen in “Hamlet”: “Brevity,” he suggested, “is the soul of wit.”

Another great advantage is the ability to incorporate AI at early stages of image acquisition. Among other things, this enables us to reduce the amount of radiation needed to acquire a high-resolution CT or shorten the duration needed for an MRI scan. And this leads to patient welfare improvements as well as healthcare cost reductions.

AI applications

In recent years there has been tremendous work in this field mainly focusing on cardiovascular, ophthalmology, neurology, and cancer detection.