One factor needed to ensure data quality is subject matter expertise. Domain experts can provide invaluable feedback when deploying, for instance, machine learning for cybersecurity. "Typically, you do something with your data so that an AI algorithm can interpret it," Ramzan said. That step might involve deleting incomplete or inaccurate data or adding contextual information to an additional data set. "When you do this step, it should be geared toward allowing the algorithm to look at the data elements that matter the most and coming up with some determination about the problem you are trying to solve," Ramzan said.

AI has the potential to improve human lives and a company's bottom line, but it can also accelerate inequality and eliminate jobs during the worst U.S. recession since the Great Depression. This dual promise and peril led members of the House Budget Committee to hold a hearing today to discuss the impact of AI on economic recovery, the future of work, and the federal budget. Expert witnesses recommended approaches that ranged from giving people lifelong upskilling accounts to creating regional investment districts and portable benefits. Daron Acemoglu warned the committee about the dangers of excessive automation. The MIT professor and economist recently found that every robot replaces 3.3 human jobs in the U.S. In a working paper published by the National Bureau of Economic Research, Acemoglu detailed how excessive automation looks for ways to replace workers with machines or algorithms but produces few new jobs.

Most state-of-the-art machine learning algorithms induce black-box models, preventing their application in many sensitive domains. Hence, many methodologies for explaining machine learning models have been proposed to address this problem. In this work, we introduce strategies to improve local explanations taking into account the distribution of the data used to train the black-box models. We show that our approach, MeLIME, produces more meaningful explanations compared to other techniques over different ML models, operating on various types of data. MeLIME generalizes the LIME method, allowing more flexible perturbation sampling and the use of different local interpretable models. Additionally, we introduce modifications to standard training algorithms of local interpretable models fostering more robust explanations, even allowing the production of counterfactual examples. To show the strengths of the proposed approach, we include experiments on tabular data, images, and text; all showing improved explanations. In particular, MeLIME generated more meaningful explanations on the MNIST dataset than methods such as GuidedBackprop, SmoothGrad, and Layer-wise Relevance Propagation. MeLIME is available on https://github.com/tiagobotari/melime.

A 2016 report claims that annually upwards of 235 000 Canadians endure periods of homelessness, with approximately 35 000 individuals lacking a place to stay each night [1]. Between 2005 and 2014, there was a downward trend in the total number of Canadians using shelters; however, the occupancy rates of shelters has been increasing [1]. One factor accounting for this ongoing decrease in the number of homeless individuals paired with an increase in shelter occupancy is an increase in chronic homelessness. London's Homeless Prevention division identifies an individual as chronically homelessness if they have spent 6 or more months ( 180 days) of the last year in a shelter, which was based on the definition of chronic homelessness outlined by the Canadian government's homelessness strategy directives [2]. In addition to this trend, the demographics of homelessness are changing in Canada. In preceding decades, older, single males are over-represented in the homeless population; in contrast, the homeless population of today is increasingly diverse, with families, women, and youth comprising a greater fraction [1].

Deepfakes is a class of synthetic media generated by AI and represents another dark side of technology -- this form of Artificial Intelligence stole the headlines last year when a LinkedIn user by the name Katie Jones, who appeared on the platform & started connecting with the Who's Who of the political elite in Washington DC. It was alarming, how deep learning created a real-life image of a person & then penetrated the social media spreading misinformation. With the U.S presidential elections looming, lawmakers in the country are worried about how deepfakes can greatly jeopardize the transparency of the democratic process. Many of the leading tech companies have been asked for help and are working on developing tools that can detect this fake synthetic media. Global software giant, Microsoft, has now released two new tools that can spot if a certain media has been artificially manipulated.

AI will face off against human pilots in real-world fighter aircraft by 2024, Secretary of Defense Mark Esper revealed on Wednesday. The Pentagon announced the plan a month after an AI system demolished an Air Force pilot in a virtual dogfight. An algorithm developed by defense contractor Heron Systems swept a best-of-five aerial duel versus an F-16 pilot wearing a VR helmet. The new trials will test how the AI's capabilities transfer to the real world, Esper explained on Wednesday at the Pentagon's first AI Symposium: The AI agent's resounding victory demonstrated the ability of advanced algorithms to out-perform humans in virtual dogfights. To be clear, AI's role in our lethality is to support human decision-makers, not replace them.

Pressure on US lawmakers to create federal regulations on facial recognition has been mounting. IBM, Amazon, and Microsoft stopped selling the technology to US police, and called on Congress to regulate its use. Amidst international protests against racism and police misconduct, news broke that Detroit police had wrongfully arrested a Black man based on a faulty facial recognition match. In response, House Democrats proposed a bill last week that would ban police from using facial recognition. Against that backdrop, industry groups have quietly lobbied to soften regulations and avoid an outright ban.

Drone footage shows streets of houses that have been completely wiped out by wildfires in the US state of Oregon. Some residents have been allowed to return to where their houses once stood and are realising the extent of the damage to their communities. Read more: Half a million flee'unprecedented' Oregon fires

This article is authored by Duane Pozza, Megan Brown, and Rick Sofield. The U.S. government is increasingly focused on competition between the United States and China in the development of artificial intelligence (AI), as a national security issue. The Administration has oriented its approach to AI to position the U.S. as a leader in AI development and standards, explicitly stating that it is working with its allies in opposition to China. Given the Administration's aggressive stance on trade restrictions with China in a variety of areas, one question facing industry is how AI technology will be regulated by the U.S. government. A recent report by the congressionally formed National Security Commission on Artificial Intelligence (NSCAI) provides some insights on how the Administration – and in particular the U.S. Department of Commerce – might approach AI technology protection.

For quantum optical technologies to become more practical, there is a need for large-scale integration of quantum photonic circuits on chips. This integration calls for scaling up key building blocks of these circuits – sources of particles of light – produced by single quantum optical emitters. Purdue University engineers created a new machine learning-assisted method that could make quantum photonic circuit development more efficient by rapidly preselecting these solid-state quantum emitters. The work is published in the journal Advanced Quantum Technologies. Researchers around the world have been exploring different ways to fabricate identical quantum sources by "transplanting" nanostructures containing single quantum optical emitters into conventional photonic chips.