I most appreciate AI when it augments human work and makes us stronger or more effective; when it performs tasks that we do not want to do or struggle to do, thereby freeing us to enjoy the activities we are good at and like. Remember how the introduction of the electronic spreadsheet did not make accountants or statisticians obsolete but instead gave them numerical "super-powers" that made their work easier and better? But what is critical is that companies and consumers understand how an AI algorithm uses data to make better decisions that protect all stakeholders from disappointment and harm. This means AI's decision-making process must be transparent to reinforce trust, fair to avoid bias, protective of data to ensure privacy, and vigilant against cybersecurity threats to prevent external abuse. As well as having negative implications for the consumer, the possible harm of misuse of data by AI can tarnish a company's brand, stripping away the trust that it has worked so hard to build over the years.

SEOUL – North Korea can get all the uranium it needs for nuclear weapons through its existing Pyongsan mill, and satellite imagery of tailings piles suggests the country can produce far more nuclear fuel than it is, a new academic study concludes. Despite a self-imposed moratorium on nuclear weapons tests since 2017, North Korea has said it is continuing to build its arsenal -- and this year appeared to have restarted a reactor widely believed to have produced weapons-grade plutonium. According to research published last month in the journal Science & Global Security by researchers at Stanford University and an Arizona-based mining consulting company, North Korea may be able to increase production and has no need for other uranium mills. "It is clear that the DPRK appears to have substantially more milling capacity than it has been using to date," said the report, using the initials of North Korea's official name, the Democratic People's Republic of Korea. "This means that the DPRK could produce much greater quantities of milled natural uranium if desired."

Human and artificial neural networks are very similar in design. The brain of a mammal comprises of cells called neurons which are responsible for sending messages throughout the body. The cell body contains the nucleus and has branches sprouting out from it called dendrites. Dendrites receive chemical signals from other neurons across a synapse space. Some signals are more important than others and form stronger connections.

Our lives are a succession of optimization problems. They occur when we search for the fastest route home from work or attempt to balance cost and quality on a trip to the store, or even when we decide how to spend limited free time before bed. These scenarios and many others can be represented as a mathematical optimization problem. Making the best decisions is a matter of finding their optimal solutions. And for a world steeped in optimization, two recent results provide both good and bad news.

The man responsible for ensuring the EU's institutions stick to its data protection laws believes Europe isn't ready for facial recognition tech that watches people in public. European "society is not ready," European Data Protection Supervisor (EDPS) Wojciech Wiewiórowski told POLITICO in an interview. The tech and its applications have divided Europe. The EU's proposed AI legislation bans most applications of remote biometric identification, such as facial recognition, in public places by law enforcement, but makes exceptions for fighting "serious" crime, which could include terrorism. Proponents of the technology, which include law enforcement and some security-minded governments, argue that the police need the technology to catch criminals.

Fox News contributor Joey Jones reacts to testimony from Pentagon officials on the Afghanistan withdrawal and slams the New York Times' proposed redesigns of the American flag. An independent Pentagon review has concluded that the U.S. drone strike that killed innocent Kabul civilians and children in the final days of the Afghanistan war was not caused by misconduct or negligence, and it doesn't recommend any disciplinary action. The review, done by Air Force Lt. Gen. Sami Said, found there were breakdowns in communication and in the process of identifying and confirming the target of the bombing. Said concluded that the mistaken strike happened despite prudent measures to prevent civilian deaths. "I found that given the information they had and the analysis that they did -- I understand they reached the wrong conclusion, but ... was it reasonable to conclude what they concluded based on what they had? It just turned out to be incorrect," Said said.

Airport performance prediction with a reasonable look-ahead time is a challenging task and has been attempted by various prior research. Traffic, demand, weather, and traffic management actions are all critical inputs to any prediction model. In this paper, a novel approach based on Temporal Fusion Transformer (TFT) was proposed to predict departure and arrival delays simultaneously for multiple airports at once. This approach can capture complex temporal dynamics of the inputs known at the time of prediction and then forecast selected delay metrics up to four hours into the future. When dealing with weather inputs, a self-supervised learning (SSL) model was developed to encode high-dimensional weather data into a much lower-dimensional representation to make the training of TFT more efficiently and effectively. The initial results show that the TFT-based delay prediction model achieves satisfactory performance measured by smaller prediction errors on a testing dataset. In addition, the interpretability analysis of the model outputs identifies the important input factors for delay prediction. The proposed approach is expected to help air traffic managers or decision makers gain insights about traffic management actions on delay mitigation and once operationalized, provide enough lead time to plan for predicted performance degradation.

Transformers have become the de-facto standard in the natural language processing (NLP) field. They have also gained momentum in computer vision and other domains. Transformers can enable artificial intelligence (AI) models to dynamically focus on certain parts of their input and thus reason more effectively. Inspired by the success of transformers, we adopted this technique to predict strategic flight departure demand in multiple horizons. This work was conducted in support of a MITRE-developed mobile application, Pacer, which displays predicted departure demand to general aviation (GA) flight operators so they can have better situation awareness of the potential for departure delays during busy periods. Field demonstrations involving Pacer's previously designed rule-based prediction method showed that the prediction accuracy of departure demand still has room for improvement. This research strives to improve prediction accuracy from two key aspects: better data sources and robust forecasting algorithms. We leveraged two data sources, Aviation System Performance Metrics (ASPM) and System Wide Information Management (SWIM), as our input. We then trained forecasting models with temporal fusion transformer (TFT) for five different airports. Case studies show that TFTs can perform better than traditional forecasting methods by large margins, and they can result in better prediction across diverse airports and with better interpretability.

Wildfire forecasting is of paramount importance for disaster risk reduction and environmental sustainability. We approach daily fire danger prediction as a machine learning task, using historical Earth observation data from the last decade to predict next-day's fire danger. To that end, we collect, pre-process and harmonize an open-access datacube, featuring a set of covariates that jointly affect the fire occurrence and spread, such as weather conditions, satellite-derived products, topography features and variables related to human activity. We implement a variety of Deep Learning (DL) models to capture the spatial, temporal or spatio-temporal context and compare them against a Random Forest (RF) baseline. We find that either spatial or temporal context is enough to surpass the RF, while a ConvLSTM that exploits the spatio-temporal context performs best with a test Area Under the Receiver Operating Characteristic of 0.926. Our DL-based proof-of-concept provides national-scale daily fire danger maps at a much higher spatial resolution than existing operational solutions.

The three-year contract will be split across multiple bidders. It replaces the 10-year, $10 billion JEDI cloud-computing contract terminated in July, which was planned to consolidate the Pentagon's patchwork of data systems to give defense personnel better access to real-time information and artificial-intelligence capabilities. The Pentagon said the contract was canceled because of its evolving needs. The project was mired in years of squabbling between Microsoft Corp. MSFT 0.26%, which won the bidding, and Amazon.com Inc., AMZN 2.15% which contended the process was politically motivated under the Trump administration.