Deep Learning methods have been proven to be flexible to model complex phenomena. This has also been the case of Intelligent Transportation Systems (ITS), in which several areas such as vehicular perception and traffic analysis have widely embraced Deep Learning as a core modeling technology. Particularly in short-term traffic forecasting, the capability of Deep Learning to deliver good results has generated a prevalent inertia towards using Deep Learning models, without examining in depth their benefits and downsides. This paper focuses on critically analyzing the state of the art in what refers to the use of Deep Learning for this particular ITS research area. To this end, we elaborate on the findings distilled from a review of publications from recent years, based on two taxonomic criteria. A posterior critical analysis is held to formulate questions and trigger a necessary debate about the issues of Deep Learning for traffic forecasting. The study is completed with a benchmark of diverse short-term traffic forecasting methods over traffic datasets of different nature, aimed to cover a wide spectrum of possible scenarios. Our experimentation reveals that Deep Learning could not be the best modeling technique for every case, which unveils some caveats unconsidered to date that should be addressed by the community in prospective studies. These insights reveal new challenges and research opportunities in road traffic forecasting, which are enumerated and discussed thoroughly, with the intention of inspiring and guiding future research efforts in this field.

Model compression is a ubiquitous tool that brings the power of modern deep learning to edge devices with power and latency constraints. The goal of model compression is to take a large reference neural network and output a smaller and less expensive compressed network that is functionally equivalent to the reference. Compression typically involves pruning and/or quantization, followed by re-training to maintain the reference accuracy. However, it has been observed that compression can lead to a considerable mismatch in the labels produced by the reference and the compressed models, resulting in bias and unreliability. To combat this, we present a framework that uses a teacher-student learning paradigm to better preserve labels. We investigate the role of additional terms to the loss function and show how to automatically tune the associated parameters. We demonstrate the effectiveness of our approach both quantitatively and qualitatively on multiple compression schemes and accuracy recovery algorithms using a set of 8 different real-world network architectures. We obtain a significant reduction of up to 4.1X in the number of mismatches between the compressed and reference models, and up to 5.7X in cases where the reference model makes the correct prediction.

Key components of current cybersecurity methods are the Intrusion Detection Systems (IDSs) were different techniques and architectures are applied to detect intrusions. IDSs can be based either on cross-checking monitored events with a database of known intrusion experiences, known as signature-based, or on learning the normal behavior of the system and reporting whether some anomalous events occur, named anomaly-based. This work is dedicated to the application to the Internet of Things (IoT) network where edge computing is used to support the IDS implementation. New challenges that arise when deploying an IDS in an edge scenario are identified and remedies are proposed. We focus on anomaly-based IDSs, showing the main techniques that can be leveraged to detect anomalies and we present machine learning techniques and their application in the context of an IDS, describing the expected advantages and disadvantages that a specific technique could cause.

Motivated by the EU's "Right To Be Forgotten" regulation, we initiate a study of statistical data deletion problems where users' data are accessible only for a limited period of time. This setting is formulated as an online supervised learning task with \textit{constant memory limit}. We propose a deletion-aware algorithm \texttt{FIFD-OLS} for the low dimensional case, and witness a catastrophic rank swinging phenomenon due to the data deletion operation, which leads to statistical inefficiency. As a remedy, we propose the \texttt{FIFD-Adaptive Ridge} algorithm with a novel online regularization scheme, that effectively offsets the uncertainty from deletion. In theory, we provide the cumulative regret upper bound for both online forgetting algorithms. In the experiment, we showed \texttt{FIFD-Adaptive Ridge} outperforms the ridge regression algorithm with fixed regularization level, and hopefully sheds some light on more complex statistical models.

We consider data-driven approaches that integrate a machine learning prediction model within distributionally robust optimization (DRO) given limited joint observations of uncertain parameters and covariates. Our framework is flexible in the sense that it can accommodate a variety of learning setups and DRO ambiguity sets. We investigate the asymptotic and finite sample properties of solutions obtained using Wasserstein, sample robust optimization, and phi-divergence-based ambiguity sets within our DRO formulations, and explore cross-validation approaches for sizing these ambiguity sets. Through numerical experiments, we validate our theoretical results, study the effectiveness of our approaches for sizing ambiguity sets, and illustrate the benefits of our DRO formulations in the limited data regime even when the prediction model is misspecified.

China's Chang'e-5 probe touched down on the surface of the Moon Tuesday just before 10:30am ET, and it's scheduled to spend the next few days collecting rocks and dust before heading back to Earth. Chinese state news outlets didn't air the touchdown live, but they reported the landing afterward with a brief video showing the probe's shadow encroaching on the pockmarked surface of the Moon. If successful, the Chang'e-5 mission will mark China's first sample-return Moon mission. It will also be the first time humans have harvested lunar rocks in 44 years. Both the United States and the Soviet Union collected Moon material throughout the Space Race in the 1960s and '70s, and US scientists are still studying these samples today.

This article is part of the Future Agenda, a series from Future Tense in which experts suggest specific, forward-looking actions the new Biden administration should implement. In October, Tesla offered some of its customers an upgrade to its "Autopilot" driver-assistance system called "Full Self-Driving." Anyone familiar with how Tesla cars work knows that "Autopilot" isn't really "autopilot," and "Full Self-Driving" isn't "full" either. For now, the feature allows a car to stay within lanes on a road, automatically brake in an emergency, turn, and respond to traffic signals on its own. But the company warns drivers to "not become complacent" because the vehicle "may do the wrong thing at the worst time."

Artificial Intelligence ("AI") is clearly on the horizon of the regulatory landscape. Alongside the use of technology to assist with navigating the regulatory process, regulators are now digitizing their enforcement efforts. The Canadian Securities Administrators ("CSA")1 have approached this challenge head-on. In 2018, the CSA put the capital markets on notice that they were strengthening their technological capabilities to assist in fighting securities misconduct.2 The CSA confirmed they would rely on AI technology to analyze large data sets, allowing them to detect misconduct faster and earlier, through the Market Analysis Platform ("MAP"), an automated centralized solution that the CSA believed could handle the size of the current market practices.

At the height of the exchange of accusations between the United States and China regarding the "Covid-19" disease, new signs of a war between the two countries appeared, the Artificial intelligence War, which lead us to ask: Is this technology ready to work in safety? And can military AI be deceived easily? Although military AI technologies dominate military strategy in the US and China; But what sparked the crisis was that last March, Chinese researchers launched a brilliant, and potentially devastating, attack against one of America's most valuable technological assets, the Tesla electric car. A research team from the security laboratory of the Chinese technology giant "Tencent" has succeeded in finding several ways to deceive the artificial intelligence algorithms in the Tesla electric car by carefully changing the data, which are fed to the car's sensors, and the team managed to trick and confuse the vehicle's AI. The team tricked Tesla's brilliant algorithms capable of detecting raindrops on the windshield or following the lines on the road, operating the windshield wipers to act as if there was rain, and the lane markings on the road were modified to confuse the autonomous driving system so that it passed in the opposite traffic lane in violation of traffic rules.

It's hard to believe, but a year in which the unprecedented seemed to happen every day is just weeks from being over. In AI circles, the end of the calendar year means the rollout of annual reports aimed at defining progress, impact, and areas for improvement. The AI Index is due out in the coming weeks, as is CB Insights' assessment of global AI startup activity, but two reports -- both called The State of AI -- have already been released. Last week, McKinsey released its global survey on the state of AI, a report now in its third year. Interviews with executives and a survey of business respondents found a potential widening of the gap between businesses that apply AI and those that do not.