Conventional wisdom holds that science fiction was written almost exclusively by men until the advent of feminism in the 1960s and '70s. But when Lisa Yaszek, who teaches science fiction studies at Georgia Tech, went digging through old magazines, she discovered a very different story. "I was so surprised to see how many women there were in science fiction before women really came into the genre in the 1970s with feminist science fiction," Yaszek says in Episode 346 of the Geek's Guide to the Galaxy podcast. "I kept uncovering these anthologies with all of these women who were clearly well-known and celebrated in their day, and who I had never heard of." In fact, women writers were relatively common throughout the pulp era, and the proportion of women readers was even higher.

Robotic process automation is becoming an integral part of how businesses operate, and these solutions are making employees both more efficient and effective. PwC estimates that companies can automate approximately 45 percent of workplace activities. By 2024, robotics process automation is forecasted to be a $3.5-billion-dollar market in North America alone. However, robotics won't just change operations -- process automation will change everything from the boardroom to education. The rise of RPA will pave the way for a new role in the C-suite: the Chief Robotics Officer.

In this paper we consider the problem of how a reinforcement learning agent that is tasked with solving a sequence of reinforcement learning problems (a sequence of Markov decision processes) can use knowledge acquired early in its lifetime to improve its ability to solve new problems. We argue that previous experience with similar problems can provide an agent with information about how it should explore when facing a new but related problem. We show that the search for an optimal exploration strategy can be formulated as a reinforcement learning problem itself and demonstrate that such strategy can leverage patterns found in the structure of related problems. We conclude with experiments that show the benefits of optimizing an exploration strategy using our proposed approach.

Since technology is not going anywhere and does more good than harm, adapting is the best course of action. That is where The Tech Edvocate comes in. We plan to cover the PreK-12 and Higher Education EdTech sectors and provide our readers with the latest news and opinion on the subject. From time to time, I will invite other voices to weigh in on important issues in EdTech. We hope to provide a well-rounded, multi-faceted look at the past, present, the future of EdTech in the US and internationally.

Issues around talent management are becoming more and more complex, from the role that artificial intelligence and talent analytics plays, to the shift in how people are looking to get compensated. To succeed in attracting, developing and retaining top talent, it's critical to be agile and forward thinking. Recruiters are no longer looking at employment gaps with raised eyebrows as the stigma of taking time off between jobs to raise children, travel, or learn new skills wears off. Tactics to reach professionals who have been out of the workforce include targeted proactive sourcing, talent communities, workshops, alumni networks for those who have left the company and may consider returning, and'buddy' systems for effective onboarding. Making artificial intelligence more'intelligent' Artificial intelligence (AI) has been touted as a new holy grail in recruiting – particularly in helping to source qualified candidates.

Authorities say a 13-year-old northwestern Indiana boy is charged with intimidation after telling Apple's digital assistant Siri he planned a school shooting and posting an iPhone screenshot of the response on social media as an apparent joke.

A team of researchers from the University of North Carolina at Chapel Hill have unveiled a 3D printer that uses light to create an entire object at once. It's called the Replicator, named after the machines in the Star Trek universe that can synthesize food, water, air and various objects seemingly out of nothing. Before you get too excited, the researchers didn't quite create an exact replica of that fictional machine, but it still offers a new and promising 3D printing technique. According to the team's paper published in Science, the Replicator works like a reverse CT scan. When a patient undergoes the procedure, an X-ray tube rotates around their body to take multiple photos that a computer can use to create 3D images.

This article was co-written by Chris Davis and Brandon Metzger. From detailed homework review to back office automation, progress in artificial intelligence will continue to explode in the year ahead. In 2018, Metis Strategy interviewed nearly 40 CIOs, CDOs and CTOs of companies with over $1 billion in revenue as part of our Technovation podcast and column. When asked to identify the emerging technologies that are of growing interest or are making their way onto their 2019 roadmap, 75 percent of the technology leaders highlighted artificial intelligence, while 40 percent said blockchain and 13 percent cited the Internet of Things. AI, an umbrella term for technologies that enable machines to accomplish tasks that previously required human intelligence, could rapidly upend the competitive landscape across industries.

We present a particle flow realization of Bayes' rule, where an ODE-based neural operator is used to transport particles from a prior to its posterior after a new observation. We prove that such an ODE operator exists and its neural parameterization can be trained in a meta-learning framework, allowing this operator to reason about the effect of an individual observation on the posterior, and thus generalize across different priors, observations and to online Bayesian inference. We demonstrated the generalization ability of our particle flow Bayes operator in several canonical and high dimensional examples.

We present DANTE, a novel method for training neural networks using the alternating minimization principle. DANTE provides an alternate perspective to traditional gradient-based backpropagation techniques commonly used to train deep networks. It utilizes an adaptation of quasi-convexity to cast training a neural network as a bi-quasi-convex optimization problem. We show that for neural network configurations with both differentiable (e.g. sigmoid) and non-differentiable (e.g. ReLU) activation functions, we can perform the alternations very effectively. DANTE can also be extended to networks with multiple hidden layers. In experiments on standard datasets, neural networks trained using the proposed method were found to be very promising and competitive to traditional backpropagation techniques, both in terms of quality of the solution, as well as training speed.