Last June, a team at Harvard Medical School and MIT showed that it's pretty darn easy to fool an artificial intelligence system analyzing medical images. Researchers modified a few pixels in eye images, skin photos and chest X-rays to trick deep learning systems into confidently classifying perfectly benign images as malignant. These so-called "adversarial attacks" implement small, carefully designed changes to data--in this case pixel changes imperceptible to human vision--to nudge an algorithm to make a mistake. That's not great news at a time when medical AI systems are just reaching the clinic, with the first AI-based medical device approved in April and AI systems besting doctors at diagnosis across healthcare sectors. Now, in collaboration with a Harvard lawyer and ethicist, the same team is out with an article in the journal Science to offer suggestions about when and how the medical industry might intervene against adversarial attacks.

Food waste could become a thing of the past thanks to an AI powered smart bin that let's you know the type of items you throw away most regularly. The system uses a camera, a set of smart scales and the same type of machine learning technology found in self-driving cars. It comes pre-programmed with common items and learns to recognise different foods being thrown away regularly. It uses this information to calculate the financial and environmental cost of this wasted food, so that you can tailor your next food order accordingly. The smart bin is currently aimed at commercial kitchens but could one day be a common feature in people's homes, the firm hopes. Food waste could become a thing of the past thanks to an AI powered smart bin that let's you know the type of items you throw away most regularly.

Autonomous vehicle development is a time and resource-intensive business, requiring dozens of test vehicles, thousands of hours of data collection and millions of miles of driving to hone the artificial brains of the cars of tomorrow. What if you could do most of that in the cloud? That's the question Nvidia hopes to answer with the release of its Nvidia Drive Constellation testing platform for self-driving cars. The announcement came during the keynote address at Nvidia's 2019 GPU Technology Conference in San Jose Monday. Drive Constellation is, basically, a simulation and validation platform that allows automakers and developers to test their autonomous vehicles and technologies in a virtual environment that lives in a specially-designed cloud server.

With any change comes the fear of the unknown, but this is especially true when it comes to artificial intelligence. Universities today have so much to gain by leveraging AI across the student lifecycle, but many are hesitant. Taking a step back, this somewhat nebulous concept of AI is already taking root in our everyday lives in so many forms. Today, you can wake up with a reminder and a playlist of your favorite motivational morning music via a voice-activated assistant, then get traffic advice on your way to work from a maps app. A quick tap on a suggestion based on previous purchases, and your favorite variety of coffee is waiting at your favorite store, already paid for in-app.

The TriRhenaTech alliance universities and their partners presented their competences in the field of artificial intelligence and their cross-border cooperations with the industry at the tri-national conference 'Artificial Intelligence : from Research to Application' on March 13th, 2019 in Offenburg. The TriRhenaTech alliance is a network of universities in the Upper Rhine Trinational Metropolitan Region comprising of the German universities of applied sciences in Furtwangen, Kaiserslautern, Karlsruhe, and Offenburg, the Baden-Wuerttemberg Cooperative State University Loerrach, the French university network Alsace Tech (comprised of 14 'grandes \'ecoles' in the fields of engineering, architecture and management) and the University of Applied Sciences and Arts Northwestern Switzerland. The alliance's common goal is to reinforce the transfer of knowledge, research, and technology, as well as the cross-border mobility of students.

Recall assistance methods are among the key aspects that improve the accuracy of online dietary assessment surveys. These methods still mainly rely on experience of trained interviewers with nutritional background, but data driven approaches could improve cost-efficiency and scalability of automated dietary assessment. We evaluated the effectiveness of a recommender algorithm developed for an online dietary assessment system called Intake24, that automates the multiple-pass 24-hour recall method. The recommender builds a model of eating behavior from recalls collected in past surveys. Based on foods they have already selected, the model is used to remind respondents of associated foods that they may have omitted to report. The performance of prompts generated by the model was compared to that of prompts hand-coded by nutritionists in two dietary studies. The results of our studies demonstrate that the recommender system is able to capture a higher number of foods omitted by respondents of online dietary surveys than prompts hand-coded by nutritionists. However, the considerably lower precision of generated prompts indicates an opportunity for further improvement of the system.

Reinforcement Learning (RL) algorithms have found limited success beyond simulated applications, and one main reason is the absence of safety guarantees during the learning process. Real world systems would realistically fail or break before an optimal controller can be learned. To address this issue, we propose a controller architecture that combines (1) a model-free RL-based controller with (2) model-based controllers utilizing control barrier functions (CBFs) and (3) on-line learning of the unknown system dynamics, in order to ensure safety during learning. Our general framework leverages the success of RL algorithms to learn high-performance controllers, while the CBF-based controllers both guarantee safety and guide the learning process by constraining the set of explorable polices. We utilize Gaussian Processes (GPs) to model the system dynamics and its uncertainties. Our novel controller synthesis algorithm, RL-CBF, guarantees safety with high probability during the learning process, regardless of the RL algorithm used, and demonstrates greater policy exploration efficiency. We test our algorithm on (1) control of an inverted pendulum and (2) autonomous car-following with wireless vehicle-to-vehicle communication, and show that our algorithm attains much greater sample efficiency in learning than other state-of-the-art algorithms and maintains safety during the entire learning process.

Standard stochastic optimization methods are brittle, sensitive to stepsize choices and other algorithmic parameters, and they exhibit instability outside of well-behaved families of objectives. To address these challenges, we investigate models for stochastic minimization and learning problems that exhibit better robustness to problem families and algorithmic parameters. With appropriately accurate models---which we call the aProx family---stochastic methods can be made stable, provably convergent and asymptotically optimal; even modeling that the objective is nonnegative is sufficient for this stability. We extend these results beyond convexity to weakly convex objectives, which include compositions of convex losses with smooth functions common in modern machine learning applications. We highlight the importance of robustness and accurate modeling with a careful experimental evaluation of convergence time and algorithm sensitivity.

Mini-batch sub-sampling is likely here to stay, due to growing data demands, memory-limited computational resources such as graphical processing units (GPUs), and the dynamics of on-line learning. Sampling a new mini-batch at every loss evaluation brings a number of benefits, but also one significant drawback: The loss function becomes discontinuous. These discontinuities are generally not problematic when using fixed learning rates or learning rate schedules typical of subgradient methods. However, they hinder attempts to directly minimize the loss function by solving for critical points, since function minimizers find spurious minima induced by discontinuities, while critical points may not even exist. Therefore, finding function minimizers and critical points in stochastic optimization is ineffective. As a result, attention has been given to reducing the effect of these discontinuities by means such as gradient averaging or adaptive and dynamic sampling. This paper offers an alternative paradigm: Recasting the optimization problem to rather find Non-Negative Associated Gradient Projection Points (NN-GPPs). In this paper, we demonstrate the NN-GPP interpretation of gradient information is more robust than critical points or minimizers, being less susceptible to sub-sampling induced variance and eliminating spurious function minimizers. We conduct a visual investigation, where we compare function value and gradient information for a variety of popular activation functions as applied to a simple neural network training problem. Based on the improved description offered by NN-GPPs over minimizers to identify true optima, in particular when using smooth activation functions with high curvature characteristics, we postulate that locating NN-GPPs can contribute significantly to automating neural network training.

Learning with less supervision is a major challenge in artificial intelligence. One sensible approach to decrease the amount of supervision is to leverage prior experience and transfer knowledge from tasks seen in the past. However, a necessary condition for a successful transfer is the ability to remember how to perform previous tasks. The Continual Learning (CL) setting, whereby an agent learns from a stream of tasks without seeing any example twice, is an ideal framework to investigate how to accrue such knowledge. In this work, we consider supervised learning tasks and methods that leverage a very small episodic memory for continual learning. Through an extensive empirical analysis across four benchmark datasets adapted to CL, we observe that a very simple baseline, which jointly trains on both examples from the current task as well as examples stored in the memory, outperforms state-of-the-art CL approaches with and without episodic memory. Surprisingly, repeated learning over tiny episodic memories does not harm generalization on past tasks, as joint training on data from subsequent tasks acts like a data dependent regularizer. We discuss and evaluate different approaches to write into the memory. Most notably, reservoir sampling works remarkably well across the board, except when the memory size is extremely small. In this case, writing strategies that guarantee an equal representation of all classes work better. Overall, these methods should be considered as a strong baseline candidate when benchmarking new CL approaches