The term "artificial intelligence" has been around since 1956, and yet many journalists are unfamiliar with its history and impact on the world today, even as its influence grows everywhere, including on how we gather and report the news. The next massive open online course (MOOC) in Spanish, and the Knight Center's first in partnership with Microsoft, will familiarize students with the foundations of artificial intelligence (AI) and how it impacts the news industry. "Artificial Intelligence: How to cover AI and understand its impact on journalism," will run from Oct. 22 to Nov. 25, 2018 and will be taught by Sandra Crucianelli, a veteran instructor for Knight Center MOOCs and a member of the International Consortium of Investigative Journalists (ICIJ). "The course will be a wonderful opportunity for those who have not yet become familiar with artificial intelligence technologies," Crucianelli said. "We will be sharing definitions, but also analyzing applications, examples and there also will be online discussions. For example, will robots replace journalists? This is a question that many of us ask and I believe the exchange of opinions will be very interesting."

The head of Air Force training said Tuesday that the service wants artificial intelligence to become the go-to coach that helps airmen learn faster and better than ever before. Lt. Gen. Steven Kwast, commander of Air Education and Training Command, said he hopes that the results of testing, scheduled to be completed next year, will show that futuristic tools such as AI, virtual reality and super-computing can improve the speed and effectiveness of the human brain. "The data is very promising that we can accentuate the adult human brain to learn faster, better and, I'll say, more sticky, meaning when you learn something longer and better," Kwast told a group of defense reporters at the Air Force Association's Air, Space & Cyber Conference. He used pilot training as an example of how artificial intelligence can be used as a coach in a flight simulator. "Let's take a loop: A pilot has to do a loop, and the artificial intelligence is watching you do that loop. And, as you pull back on the stick, it can tell what you are doing and says, 'Hey, you are pulling back too much. Keep your eye on the horizon,' " Kwast said.

The biggest trend in video surveillance and indeed the security industry as a whole over the past several years has been the rise of artificial intelligence (AI) and the revolutionary promise it holds for the market. What once seemed like capabilities reserved only for the characters of science fiction novels and movies, are quickly becoming a reality because of advancements in computing combined with parallel breakthroughs in machine learning technology. This has subsequently resulted in a renaissance for video analytics, which were frequently written off by systems integrators and end users as being an over-hyped solutions. Indeed, many of the vendors that offered video analytics as a standalone solution as little as a decade ago have now been largely relegated to the ash heap of industry history via company or patent acquisition. Now, however, the technology is once again flourishing with fresh venture capital funds flowing into an evergrowing number of companies.

The irony is not lost on Kate Saenko. Now that humans have programmed computers to learn, they want to know exactly what the computers have learned, and how they make decisions after their learning process is complete. To do that, Saenko, a Boston University College of Arts & Sciences associate professor of computer science, used humans--asking them to look at dozens of pictures depicting steps that the computer may have taken on its road to a decision, and identify its most likely path. The humans gave Saenko answers that made sense, but there was a problem: they made sense to humans, and humans, Saenko knew, have biases. In fact, humans don't even understand how they themselves make decisions.

As recently as last year, nearly one in five students who committed to attending Georgia State University (GSU) never showed up for classes in the fall. This problem isn't unique to GSU, and it's commonly referred to as the "summer melt." But GSU has taken an innovative approach to solving this challenge, using an artificially intelligent (AI) chatbot that has led to a significant increase in student enrollment. Summer melt most commonly affects low-income students, many of whom are the first in their family to be accepted into college. Navigating the complex student enrollment process can be intimidating for anyone, but especially these students--and many just give up before they complete the process.

In an open-world setting, it is inevitable that an intelligent agent (e.g., a robot) will encounter visual objects, attributes or relationships it does not recognize. In this work, we develop an agent empowered with visual curiosity, i.e. the ability to ask questions to an Oracle (e.g., human) about the contents in images (e.g., What is the object on the left side of the red cube?) and build visual recognition model based on the answers received (e.g., Cylinder). In order to do this, the agent must (1) understand what it recognizes and what it does not, (2) formulate a valid, unambiguous and informative language query (a question) to ask the Oracle, (3) derive the parameters of visual classifiers from the Oracle response and (4) leverage the updated visual classifiers to ask more clarified questions. Specifically, we propose a novel framework and formulate the learning of visual curiosity as a reinforcement learning problem. In this framework, all components of our agent, visual recognition module (to see), question generation policy (to ask), answer digestion module (to understand) and graph memory module (to memorize), are learned entirely end-to-end to maximize the reward derived from the scene graph obtained by the agent as a consequence of the dialog with the Oracle. Importantly, the question generation policy is disentangled from the visual recognition system and specifics of the environment. Consequently, we demonstrate a sort of double generalization. Our question generation policy generalizes to new environments and a new pair of eyes, i.e., new visual system. Trained on a synthetic dataset, our results show that our agent learns new visual concepts significantly faster than several heuristic baselines, even when tested on synthetic environments with novel objects, as well as in a realistic environment.

In conventional binary classification, examples are labeled as either positive (P) or negative (N), and we train a classifier on these labeled examples. On the contrary, positive-unlabeled (PU) learning addresses the problem of learning a classifier from P and unlabeled (U) data, without need of explicitly identifying N data (Elkan & Noto, 2008; Ward et al., 2009). PU learning finds its usefulness in many real-world problems. For example, in one-class remote sensing classification (Li et al., 2011), we seek to extract a specific land-cover class from an image. While it is easy to label examples of this specific land-cover class of interest, examples not belonging to this class are too diverse to be exhaustively annotated. The same problem arises in text classification, as it is difficult or even impossible to compile a set of N samples that provides a comprehensive characterization of everything that is not in the P class (Liu et al., 2003; Fung et al., 2006). Besides, PU learning has also been applied to other domains such as outlier detection (Hido et al., 2008; Scott & Blanchard, 2009), medical diagnosis (Zuluaga et al., 2011), or time series classification (Nguyen et al., 2011). By carefully examining the above examples, we find out that the most difficult step is often to collect a fully representative N set, whereas only labeling a small portion of all possible N data is relatively easy. Therefore, in this paper, we propose to study the problem of learning from P, U and biased N (bN) data, which we name PUbN learning hereinafter.

We present Predicted Variables (PVars), an approach to making machine learning (ML) a first class citizen in programming languages. There is a growing divide in approaches to building systems: using human experts (e.g. programming) on the one hand, and using behavior learned from data (e.g. ML) on the other hand. PVars aim to make ML in programming as easy as `if' statements and with that hybridize ML with programming. We leverage the existing concept of variables and create a new type, a predicted variable. PVars are akin to native variables with one important distinction: PVars determine their value using ML when evaluated. We describe PVars and their interface, how they can be used in programming, and demonstrate the feasibility of our approach on three algorithmic problems: binary search, Quicksort, and caches. We show experimentally that PVars are able to improve over the commonly used heuristics and lead to a better performance than the original algorithms. As opposed to previous work applying ML to algorithmic problems, PVars have the advantage that they can be used within the existing frameworks and do not require the existing domain knowledge to be replaced. PVars allow for a seamless integration of ML into existing systems and algorithms. Our PVars implementation currently relies on standard Reinforcement Learning (RL) methods. To learn faster, PVars use the heuristic function, which they are replacing, as an initial function. We show that PVars quickly pick up the behavior of the initial function and then improve performance beyond that without ever performing substantially worse -- allowing for a safe deployment in critical applications.

Knowledge distillation (KD) is a popular method for reducing the computational overhead of deep network inference, in which the output of a teacher model is used to train a smaller, faster student model. Hint training (i.e., FitNets) extends KD by regressing a student model's intermediate representation to a teacher model's intermediate representation. In this work, we introduce bLock-wise Intermediate representation Training (LIT), a novel model compression technique that extends the use of intermediate representations in deep network compression, outperforming KD and hint training. LIT has two key ideas: 1) LIT trains a student of the same width (but shallower depth) as the teacher by directly comparing the intermediate representations, and 2) LIT uses the intermediate representation from the previous block in the teacher model as an input to the current student block during training, avoiding unstable intermediate representations in the student network. We show that LIT provides substantial reductions in network depth without loss in accuracy -- for example, LIT can compress a ResNeXt-110 to a ResNeXt-20 (5.5) on CIFAR10 and a VDCNN-29 to a VDCNN-9 (3.2) on Amazon Reviews without loss in accuracy, outperforming KD and hint training in network size for a given accuracy. We also show that applying LIT to identical student/teacher architectures increases the accuracy of the student model above the teacher model, outperforming the recently-proposed Born Again Networks procedure on ResNet, ResNeXt, and VDCNN. Finally, we show that LIT can effectively compress GAN generators, which are not supported in the KD framework because GANs output pixels as opposed to probabilities. Modern deep networks have achieved increased accuracy by continuing to introduce more layers (Ioffe & Szegedy, 2015; He et al., 2016) at the cost of higher computational overhead. In response, researchers have proposed many techniques to reduce this computational overhead at inference time, which broadly fall under two categories. First, in deep compression (Han et al., 2015; Zhu et al., 2016; Li et al., 2016; Hubara et al., 2017), parts of a model are removed or quantized to reduce the number of weights and/or the computational footprint.

Machine learning libraries are becoming faster and more accessible with each passing year, showing no signs of slowing down. While traditionally Python has been the go-to language for machine learning, nowadays neural networks can run in any language, including JavaScript! The web ecosystem has made a lot of progress in recent times and although JavaScript and Node.js are still less performant than Python and Java, they are now powerful enough to handle many machine learning problems. Web languages also have the advantage of being super accessible - all you need to run a JavaScript ML project is your web browser. Most JavaScript machine learning libraries are fairly new and still in development, but they do exist and are ready for you to try them.