Our technological prowess is increasing every day, but the machines we use are still limited by the knowledge that the designers put into them. Researchers have started experimenting with a branch of computer science called “machine learning,” where computers can learn and adapt. The name evokes sinister Terminator imagery, but researchers at UC San Diego have been using machine learning to develop the most advanced music search engine ever created.

The researchers are using the same methodology in their system as hospitals use to train new doctors.

“When doctors learn how to diagnose diseases, they are trained to find patterns in patient data that indicate which disease the patient has. With many patients, patterns begin to emerge,” said Gert Lanckriet, a professor of electrical engineering at the UC San Diego Jacobs School of Engineering.

People identify music by its characteristics, just like doctors identify diseases based on symptoms; is it jazz, rock or dubstep?  Is it happy or sad, slow or fast, energizing or relaxing, piano or saxophone? The researchers started with a library of about 13,000 songs in their program, called Herd It, and started labeling each song with tags like those above. The program would then analyze patterns in the music and associate those patterns with the tags.

As more songs were tagged, Herd It began to learn for itself what patterns constituted each genre or style, and it began to make its own tag suggestions. The researchers would simply confirm or reject the program’s suggestions. The more data Herd It received from confirmations and rejections, the better it became at suggesting tags.

“The algorithms continuously adapt,” said Doug Turnbull, a graduate student working in Lanckriet’s lab on the project. “People labeled music for us in the fun context of a game. People label the song with any tag [they think is appropriate] and we use the most common ones.”

The music streaming service Pandora already does something similar. However, Pandora pays music experts to label songs. Each expert at Pandora takes about 30 minutes to tag each song by hand, making Pandora’s process very expensive. As such, Pandora has only about 900,000 songs in their library, compared to the near 20 million in the iTunes library.

Herd It is an AI (artificial intelligence) that can tag any song it encounters in a matter of seconds, using plain english terminology to describe genre, feelings and settings. It does the same thing that paid musicologists do, but cheaper, faster and in simpler terms. The program works by analyzing the waveform of a song and identifying patterns in the frequencies that are indicative of each tag.

The machine learning process is a continuous feedback loop. Users confirm or reject the program’s predictions, the machine learns something, and then it refines its predictions for the next guess. Every little bit of input from a user refines the algorithms’ accuracy on its future guesses.

Once the program has correctly identified songs, it adds the song name to a database with its related tags. Currently, Herd It uses between 200 and 300 tags to identify music, but it will soon use close to 1000.

According to the researchers, the machine learning was not the most impressive part of this project. Instead, the most unique feature was the use of crowdsourcing to create training data for a computer program, where humans are teaching a machine.

The final goal is for Herd It to not just identify a music type, but to identify the preferences of a listener, even if the listener is not consciously aware of those preferences. By analyzing listening patterns, Herd It will be able to give you the best possible match for your next song.

For Luke Barrington, founder of the UC San Diego Computer Audition Lab where Herd It was created, the advantage of Herd It is that people can search for music in natural language, rather than technical jargon.

“[Herd It] describes any piece of music in human words,” Barrington said. “You give [Herd It] words, it gives you music.”

HUDSON LOFCHIE can be reached at science@theaggie.org.

Last July a NASA-sponsored spacecraft called “Dawn” slipped into orbit about the first of its planned destinations — the asteroid Vesta, which is the second-largest asteroid in the Asteroid Belt between Mars and Jupiter. About 5 percent of meteorites recovered after falling to earth are believed to have come from Vesta’s surface.

The Asteroid Belt is the name given to hundreds of thousands of asteroids, both large and small, that orbit the Sun in a band of space that stretches partway between the orbits of Mars and Jupiter.

Were the asteroids once a planet in the past that broke apart?

“Quite the opposite,” said Christopher T. Russell, the principal investigator of the Dawn mission. “The material tried to accumulate into a planet but never was able do so.”

Russell coordinates and directs the science aspects of the Dawn mission as a part of his work as a professor in the Institute of Geophysics and Planetary Physics at UC Los Angeles. The Jet Propulsion Laboratory (JPL) in Pasadena is in charge of the actual operation and flight of the space vehicle.

Elizabeth Palmer is a first-year graduate student who is working as a research assistant under Russell’s guidance at UC Los Angeles. She studied astronomy as an undergraduate at Case Western Reserve University in Cleveland, Ohio, but ended up deciding that planetary science was a better fit for what she wanted to do. She was able to find a planetary science internship at JPL.

While at JPL, Palmer began learning about the technique of interpreting radar signals which have been bounced or grazed off the surface of a planetary body and received on Earth. She hopes to be able to do the same type of work interpreting microwave transmissions from Dawn that have grazed the surface of Vesta in order to determine aspects of Vesta’s make-up, such as whether ice is present on the surface and in what amount.

“It’s an opportunistic experiment that’s not perfect, since it’s hard to alter such concrete mission plans to the way that would be most ideal for our microwave experiment to work best,” Palmer said. “We are just waiting and hoping the geometry will work at some point.”

If Dawn is able to successfully graze signals off Vesta, which are then successfully received here, Palmer will be able to compare that data with experimental data obtained from meteorites found on Earth. This mineral material was likely knocked off Vesta’s surface as a result of ancient collisions in space.

Dawn gets its name from its primary purpose, which is the study of our solar system’s early history. Russell often stresses that he views Dawn as being a kind of time machine traveling back in time, since both Vesta and Dawn’s second target, Ceres, have likely preserved evidence of the Solar System’s formation in the contours of their crust and in their composition. Both Vesta and Ceres were formed 4.6 billion years ago.

Dawn is currently scheduled to leave Vesta in August later this year and enter orbit about Ceres, the largest asteroid, in February 2015. According to Russell, it is the long-term thrust enabled by Dawn’s efficient ion-propulsion system which made the mission cost-effective and a visitation to two asteroids possible as a project goal.

BRIAN RILEY can be reached at science@theaggie.org.

A new study by UC San Francisco researchers indicates that the drug varenicline, a drug used to help stop smoking, can be used to reduce drinking.

Varenicline, commonly known by the brand name Chantix, has been FDA-approved since 2006 for helping people to stop smoking, but only recently did the researchers theorize that the drug could reduce alcohol consumption and began testing the drug on rats.

After yielding results that indicated this drug could be effective in reducing drinking, human testing was done and a significant reduction in heavy drinking was found.

Howard Fields, professor of neurology at UCSF and director of the Ernest Gallo Clinic and Research Center, and his colleagues split the participants into two groups and gave one group a placebo and the other group varenicline. He said that they started seeing results around the third week of testing.

“We didn’t know what we were going to find,” Fields said. “We thought we would see something similar to the rat reduction.”
The results confirmed the researchers’ hypothesis.

Fields believes the drug to be a great treatment for drinking. He hopes that people will read the paper he and his colleagues published and use the drug to treat heavy drinking.

One of the options patients have currently is to ask their doctors to write them a prescription of the drug to see if it helps.

Fields said that the medical industry, in general, is “pretty conservative,” meaning doctors may be reluctant to prescribe the drug at first.

Although it will probably take a larger clinical study for the drug to gain approval for drinking by the FDA, Fields believes the drug is safe and effective.

Varenicline can have some side effects, including nausea and heartburn. Fields also said that there is a black box warning of severe mood changes and suicidal thoughts, though the role of varenicline in these side effects is unclear.

However, these side effects don’t deter Fields from recommending the drug.

“It would save hundreds of lives every year if it [Varenicline] were being used widely,” Fields said.

Fields said that a good, large study might allow professionals to publish a guideline that will help physicians know to prescribe the drug to their patients.

“We’d like to see more people quit heavy drinking, but there’s not much we can do about it,” Fields said. “There are a lot of hurdles from the time a drug is discovered to be effective.”

Fields said that one of the reasons the drug might go unnoticed in the beginning is due to the fact most doctors won’t see the results just yet because they don’t have the time to read every journal.

Jennifer Mitchell, clinical project director at the Ernest Gallo Clinic and Research Center in San Francisco, also believes that the drug is effective and that patients can get the drug from their physicians.

“The FDA will reconsider [approving for treating drinking] once there is enough evidence,” Mitchell said. “Physicians can prescribe it ‘off-label’; it doesn’t mean you can’t get it for drinking.”

Mitchell said that the drug can also bring new hope to people that have tried treatment for drinking before.

Mitchell believes that people need to focus more on the “enormous societal effects” of drinking.

“People don’t realize that there are drugs out there that can help,” Mitchell said. “They need to take control of their own destinies.”
According to the UC Davis Medical Center (UCDMC), drinking can have profound effects on the body.

“Long term use [of alcohol] can cause cancer, brain damage, cirrhosis of the liver, weight gain and birth defects if drinking while pregnant,” the UCDMC website states.

ERIC C. LIPSKY can be reached at science@theaggie.org.

Imagine that a fire has just devastated a small area of forest. Trees have burned to the ground, leaving stumps and cracked trunks. All of the small bushes or grasses are completely blackened.

What’s left? Tiny seeds. The healing of the forest is partially accomplished by these seeds that only germinate after a fire. When the old, dry trees burn down, the fire breaks these seeds’ dormancy and allows new, fresh trees to grow.

This example shows how plants are evolutionarily advanced organisms. Though they can’t walk around like animals can, they are capable of a great deal of strategy, as well as offensive and defensive action.

Simple defenses are common among the plant kingdom, all to make them less palatable or even deadly to their many enemies. Thorns, pricks, irritants and poisons are all to make sure that they won’t be devoured as soon as they sprout.

Except, of course, the parts they they want to be eaten.

The next time you bite into an apple, remember that what you’re eating is an ovary. The idea from the plant’s point of view is that an animal will eat the entire sweet, juicy fruit and walk a distance away before defecating out the seeds.

Seeds reach the ground with ready-made fertilizer, and the plant has successfully spread its genes.

Human-cultivated fruit has bypassed some of these evolutionary mechanisms, with worrying results. The most common type of banana is called the Cavendish banana, a seedless banana first cultivated in the 1860s. In order to propagate more trees, farmers have to carefully remove and transplant the underground stem where they want the new plant to go.

When Panama disease, a fungal infection, nearly wiped out the entire banana population of the popular Gros Michel strain, breeders scrambled to find a new type of banana that was resistant to the disease. The result was the Dwarf Cavendish, a strain chosen for its hardiness and ease of transport rather than superior taste.

Thankfully, we learned our lesson by … uh, making all of these bananas clones of each other. Well, live and learn. Researchers have learned their lesson somewhat, as they are working to make resistant hybrids that are still tasty enough for the mass market.

Some species of plant are much more resistant to being eaten, and in fact are sometimes the ones doing the eating.

Take a species of pitcher plant, Darlingtonia californica. This plant is also called the cobra lily for its almost predatory trickery. The basic strategy of the pitcher plant is to lure hapless insects inside of its container, which is full of digestive juices.

Why can’t the bug just fly out again, if they manage to get out of the deadly liquid fast enough? Well, the inside of the container is slippery, first of all. It also has clear areas, like little windows. The fly jumps up to what it thinks is an escape route, bumps against the wall and falls back into the liquid.

Rinse and repeat until the insect is dead. Plants are more diabolical than we give them credit for.

Most plants aren’t carnivorous and are just content to grow and spread their genes. This must involve some strategy as well, though. The leaves are the organs that absorb the most sunlight; if the plant grows in the deep forest and ends up covered by shade, how can they get the sunlight that they so desperately need?

They either must grow huge leaves (which is often what you’ll see in the deep underbrush of a tropical rainforest) or they must change where the leaves grow.

It’s difficult to see in real time, but if you’re able to find fast-motion growth of a plant on YouTube, find one and watch it. You’ll notice that often, the stem will grow in one direction, stop, then take a stem in a different direction. This is often because the original stem wasn’t getting enough sunlight and it just wasn’t worth the energy it would take to completely grow it.

Roots do this, too. When a plant is just starting to grow its roots, they usually spread in all directions. Soon enough, they find that one location happens to have more nutrients than another. The section of roots in the low-nutrient place will die, and the plant now has more energy to spread in the place with the highest nutrition.

So when you look at a plant, remember that you’re looking at one of nature’s most strategical organisms.

AMY STEWART can be reached at science@theaggie.org.