Explicit latent variable models provide a flexible yet powerful framework for data synthesis, enabling controlled manipulation of generative factors. With latent variables drawn from a tractable probability density function that can be further constrained, these models enable continuous and semantically rich exploration of the output space by navigating their latent spaces. Structured latent representations are typically obtained through the joint minimization of regularization loss functions. In variational information bottleneck models, reconstruction loss and Kullback-Leibler Divergence (KLD) are often linearly combined with an auxiliary Attribute-Regularization (AR) loss. However, balancing KLD and AR turns out to be a very delicate matter. When KLD dominates over AR, generative models tend to lack controllability; when AR dominates over KLD, the stochastic encoder is encouraged to violate the standard normal prior. We explore this trade-off in the context of symbolic music generation with explicit control over continuous musical attributes. We show that existing approaches struggle to jointly minimize both regularization objectives, whereas suitable attribute transformations can help achieve both controllability and regularization of the target latent dimensions.

Tractors are commonly sold to farmers at agricultural fairs and announced in the trade press. But machinery makers are falling over themselves to get a slice of a much more unlikely advertising vehicle: the Farming Simulator video game. The developer, Giants Software, now receives hundreds of queries a year from manufacturers of equipment – from tractors and combine harvesters to trailers, balers and seed drills – about how they can feature in the game, where players create their own virtual farm. Farming Simulator is important enough that some firms even launch products at the same time as the game is updated. Search for news about Göweil, and you're just as likely to find details of the nine products in the Farming Simulator update pack released last week as coverage of its hay balers in the real world.

But no amount of amateur water drumming arranged in GarageBand could be even half as weird--and fun--as some of the music that AI can make these days. Welcome back to Science, Quickly.

In early 19th-century England, the Luddites rebelled against the introduction of machinery in the textile industry. The Luddites' name originates from the mythical tale of a weaver's apprentice called Ned Ludd who, in an act of anger against increasingly dangerous and poor working conditions, supposedly destroyed two knitting machines. Contrary to popular belief, the Luddites were not against technology because they were ignorant or inept at using it (1). In fact, the Luddites were perceptive artisans who cared about their craft, and some even operated machinery. Moreover, they understood the consequences of introducing machinery to their craft and working conditions.

Score-based generative models and diffusion probabilistic models have been successful at generating high-quality samples in continuous domains such as images and audio. However, due to their Langevin-inspired sampling mechanisms, their application to discrete and sequential data has been limited. In this work, we present a technique for training diffusion models on sequential data by parameterizing the discrete domain in the continuous latent space of a pre-trained variational autoencoder. Our method is non-autoregressive and learns to generate sequences of latent embeddings through the reverse process and offers parallel generation with a constant number of iterative refinement steps. We apply this technique to modeling symbolic music and show strong unconditional generation and post-hoc conditional infilling results compared to autoregressive language models operating over the same continuous embeddings.

America has been glued to their TV screens since the MLB playoffs began on October 1. As the field has whittled down to just four teams, odds makers are eager to figure out which team has the edge. Researchers at DataRobot thought it would be a fun exercise to pull all of the MLB data from the last few decades and have their AI figure out who will win the 2019 World Series. SEE: Artificial intelligence: A business leader's guide (free PDF) (TechRepublic Premium) At the start of the playoffs, the AI predicted the Los Angeles Dodgers were most likely to win the pennant, followed closely by the Houston Astros. In the American League, DataRobot's AI said the Houston Astros had a 40% probability of winning the American League, followed by the New York Yankees at 25% and Minnesota Twins at 18%.

Why has machine learning become so critical to cybersecurity? With machine learning, cybersecurity systems can analyze patterns and learn from them to help prevent similar attacks and respond to changing behavior. Machine learning helps cybersecurity teams be more pro-active in preventing threats and responding to active attacks in real time. Machine learning can reduce the amount of time spent on routine tasks and enable organizations to use their resources more strategically. In short, machine learning can make cybersecurity simpler, more proactive, less expensive and far more effective.

Some of the world's finest oil paintings have been self-destructing, developing mysterious lumps and bumps known as'art acne'. Works by Georgia O'Keeffe and Rembrandt are among the hundreds of works blighted by the condition. For decades, art conservators have struggled to control the outbreaks, which look like grains of sand to the naked eye. But now, a team at Northwestern University in Chicago has developed an iPad software that can zoom in on the pigments closer than ever before, revealing the chemical issue at hand. In 20 seconds, the technology can scan a painting to produce a three-dimensional image of it.

A group of scientists from the Massachusetts Institute of Technology (United States) has created a machine learning system that processes sounds like people. This model can understand the meaning of a word and classify a song according to its genre or style: classical, jazz, pop, rock, blues, soul, hip hop, techno, house, etc. It is the first invention of this type that mimics the way the brain works. As the experiments carried out at MIT show, it can compete in precision with humans. The research, published in the journal Neuron, is based on deep neural networks, that is, a structure inspired by brain cells that analyses information by layers.

Not for the student, but for the teacher, who plays a short piano melody. Without missing a measure, the student follows with an improvised, yet derivative, cello run. The student plays the same run again, and then again. "I have it looping, actually, so you can hear the response over and over again," says the teacher, Jesse Engel, a computer scientist with Google Brain. "And you can hear some similarities with what I played, but it's not doing the job of trying to replicate what I played. It's trying to continue it in a meaningful way."