Reinforcement learning (RL) is a popular paradigm for addressing sequential decision tasks in which the agent has only limited environmental feedback. Despite many advances over the past three decades, learning in many domains still requires a large amount of interaction with the environment, which can be prohibitively expensive in realistic scenarios. To address this problem, transfer learning has been applied to reinforcement learning such that experience gained in one task can be leveraged when starting to learn the next, harder task. More recently, several lines of research have explored how tasks, or data samples themselves, can be sequenced into a curriculum for the purpose of learning a problem that may otherwise be too difficult to learn from scratch. In this article, we present a framework for curriculum learning (CL) in reinforcement learning, and use it to survey and classify existing CL methods in terms of their assumptions, capabilities, and goals. Finally, we use our framework to find open problems and suggest directions for future RL curriculum learning research.

Machine learning using behavioral and text data can result in highly accurate prediction models, but these are often very difficult to interpret. Linear models require investigating thousands of coefficients, while the opaqueness of nonlinear models makes things even worse. Rule-extraction techniques have been proposed to combine the desired predictive behaviour of complex "black-box" models with explainability. However, rule-extraction in the context of ultra-high-dimensional and sparse data can be challenging, and has thus far received scant attention. Because of the sparsity and massive dimensionality, rule-extraction might fail in their primary explainability goal as the black-box model may need to be replaced by many rules, leaving the user again with an incomprehensible model. To address this problem, we develop and test a rule-extraction methodology based on higher-level, less-sparse "metafeatures". We empirically validate the quality of the rules in terms of fidelity, explanation stability and accuracy over a collection of data sets, and benchmark their performance against rules extracted using the original features. Our analysis points to key trade-offs between explainability, fidelity, accuracy, and stability that Machine Learning researchers and practitioners need to consider. Results indicate that the proposed metafeatures approach leads to better trade-offs between these, and is better able to mimic the black-box model. There is an average decrease of the loss in fidelity, accuracy, and stability from using metafeatures instead of the original fine-grained features by respectively 18.08%, 20.15% and 17.73%, all statistically significant at a 5% significance level. Metafeatures thus improve a key "cost of explainability", which we define as the loss in fidelity when replacing a black-box with an explainable model.

The government of Singapore is preparing to transition to a facial recognition program it hopes will eliminate the need for ID cards by 2022. Beginning in June, kiosks fitted with cameras will be installed at a limited number of government agencies, and instead of presenting an ID card citizens will be able to check in for services with just their faces. The facial recognition system is a major expansion of the Smart Nation Initiative, which began in 2014 under Prime Minister Lee Hsien Loong and through which the state has built up a biometric database on more than four million Singaporeans over the age of 15. The facial recognition kiosks will crosscheck each new scan against this database to verify a person's identity, according to a report in The Strait Times. The kiosks will also work in tandem with SingPass Mobile, an app launched in 2018 that allows people to register their own finger print and face data with the government's biometric database.

WASHINGTON, D.C. - Today, the U.S. Department of Energy (DOE) announced plans to provide up to $40 million over three years for new research in data, artificial intelligence and machine learning to address the challenges of producing and managing data at DOE scientific user facilities. "Major scientific facilities at our DOE national laboratories are generating vast and growing amounts of data for researchers every day," said Dr. Chris Fall, Director of DOE's Office of Science. Proposals are expected to focus on each of a range of different challenges, including extracting information from complex data sets, managing facility operations in real time, and optimizing experiments through the creation of virtual laboratory environments, among other topics. "Artificial intelligence's ability to analyze and divine insights from massive data sets has the power to transform the world around us," said Cheryl Ingstad, Director of DOE's Artificial Intelligence & Technology Office. "DOE is determined to lead by example in AI application by turning this power on ourselves to optimize the way we operate facilities and push the boundaries of scientific discovery."

In the opening session of his 2020 introductory course on deep learning, Alexander Amini, a PhD student at the Massachusetts Institute of Technology (MIT), invited a famous guest: former US President Barack Obama. "Deep learning is revolutionizing so many fields, from robotics to medicine and everything in between," said Obama, who joined the class by video conference. After speaking a bit more on the virtues of artificial intelligence, Obama made an important revelation: "In fact, this entire speech and video are not real and were created using deep learning and artificial intelligence." Amini's Obama video was, in fact, a deepfake--an AI-doctored video in which the facial movements of an actor are transferred to that of a target. Since first appearing in 2018, deepfake technology has evolved from hobbyist experimentation to an effective and dangerous tool. Deepfakes have been used against celebrities and politicians and have become a threat to the very fabric of truth.

The US Navy is developing a robot submarine that is controlled by artificial intelligence that could kill without human control or input. The project is being run by the Office of Naval Research and has been described as an'autonomous undersea weapon system' according to a report by New Scientist. Details of the killer submersible were made available as part of the 2020 budget documents, which also revealed it has been named CLAWS by the US Navy. Very few details about the'top secret' project have been revealed beyond the fact it will use sensors and algorithms to carry out complex missions on its own. It's expected CLAWS will be installed on the new Orca class robot submarines that have 12 torpedo tubes and are being developed for the Navy by Boeing.

Amazon is cracking down price gougers on its platform who are looking to make a profit from the coronavirus that is wreaking havoc across the globe. The tech giant has pulled more than 530,000 listings from the site and suspended over 2,500 US sellers. The firm announced on Friday it is working with state attorneys general to identify and prosecute third-party sellers who are taking advantage of fears of the spreading coronavirus to engage in price-gouging on the Amazon website. Amazon also said it has begun manual audits of products in its online stores to spot sellers that evade its automated systems, which check for items that are'unfairly priced.' Amazon is cracking down price gougers on its platform who are looking to make a profit from the coronavirus that is wreaking havoc across the globe.

The S.A.R.I., an acronym which stands for Automatic Image Recognition System, is a new system available to the State Police to counter criminal activity; exploiting the A.F.I.S. system, Automated Fingerprint Identification System which collects the fingerprints, personal data, photographs and biometric notations of the subjects under investigation, law enforcement agencies can count on an identification system with a database of more than 10 million data; in this way those who stain a crime can be identified more quickly and efficiently. In the United States, another system is used, the C.O.M.P.A.S. (Correctional Offender Management Profiling for Alternative Sanctions), which is an algorithm used by judges to calculate the probability of recidivism within two years of a crime. These are two concrete and recent examples of Artificial Intelligence. Artificial Intelligence, for Stuart Russell and Peter Norvig, authors of "Artificial Intelligence: A Modern Approach, Global Edition", means a "field of studies in which intelligent agents are designed and built". What would be the etymological meaning of the term?

The S.A.R.I., an acronym which stands for Automatic Image Recognition System, is a new system available to the State Police to counter criminal activity; exploiting the A.F.I.S. system, Automated Fingerprint Identification System which collects the fingerprints, personal data, photographs and biometric notations of the subjects under investigation, law enforcement agencies can count on an identification system with a database of more than 10 million data; in this way those who stain a crime can be identified more quickly and efficiently. In the United States, another system is used, the C.O.M.P.A.S. (Correctional Offender Management Profiling for Alternative Sanctions), which is an algorithm used by judges to calculate the probability of recidivism within two years of a crime. These are two concrete and recent examples of Artificial Intelligence. Artificial Intelligence, for Stuart Russell and Peter Norvig, authors of "Artificial Intelligence: A Modern Approach, Global Edition", means a "field of studies in which intelligent agents are designed and built". What would be the etymological meaning of the term?

Curious how deep learning solutions are affecting your industry? On Wednesday, March 18th, join Upward, Connecticut Center for Advanced Technology (CCAT), and Connecticut Technology Council (CTC) for Hartford's first "teCTalk" focused on artificial intelligence/machine learning. Learn how machines are being taught to interact with the organic world around them and how this smart technology is working to elevate modern business. An esteemed panel of AI/ML experts across industries, including Upward Citizens GalaxE Solutions, VAANGO, and Saya Life, will navigate participants through the complex topic of deep learning. This event is designed to be interactive: pose your questions to the experts and engage with others in the crowd!