Eating a diet of junk food for just one week was enough to damage part of the brain that stops us eating more when we are already full, research suggests. Study participants who ate an abundance of fast food and high-fat milkshakes had increased cravings for more after seven days. They performed worse on cognitive tests, with results suggesting an area of the brain called the hippocampus was impaired. The hippocampus normally stops us from gorging on more food when we are full by suppressing memories of how tasty it is. When it's not working properly, the memories are more powerful and we are left unable to resist more cake, chocolates and crisps in front of us, the researchers believe.

A major California university has dropped plans to use facial recognition for the surveillance of the campus. The idea was to have the University of California Los Angeles use facial recognition as a way to gain access to buildings, to prove authenticity and deny entry to people with restricted access to the campus, matching their faces against a database. Advocacy group Fight for the Future says UCLA was the first major university exploring using facial recognition to monitor students. The group had tested facial recognition software and found that "dozens" of student-athletes and professors were incorrectly matched with photos from a mugshot database, "and the overwhelming majority of those misidentified were people of color." Why your face is the key: Do you really control how your face is being used?

AI is changing more than what computers can do and how we communicate and interact with technology. AI is changing the very nature of work, of hiring, reinforcing the imperative for life-long learning, and serving as a catalyst for organisation-wide change. How to Prepare a Generation of AI-first Workers The NHS is set to lose an estimated 350,000 staff by 2030 in the UK--a quarter of its workforce. It's a Catch 22 situation: staff are leaving because of their intolerable workloads, caused by an already acute level of staff shortages. While the National Health Service (NHS) has no magic wand to summon up a small army of suitably qualified staff, they are looking to Artificial Intelligence (AI) to help.

This 1-day workshop will not only equip artists to understand the implications and opportunities of artificial intelligence but also imagine the appropriate artistic and political responses to world that will be significantly altered by the introduction of these technologies. We need artists not only using these tools, but informing the conversation about how these tools will be deployed, and to whose benefit. This workshop is designed for ALL levels of technical expertise.

Gabriel is a professional transcriber, and for years he earned a middle-class living. He'd sit at his desk, "knock it out" for hours using custom keystrokes, and watch the money roll in. "I sent my son to private schools and university on transcribing," he tells me. "It was a nice life." But in the past decade, the bottom fell out. As audio recordings went digital and broadband spread, clients could ship work to India and the Philippines.

Optimizing a given metric is a central aspect of most current AI approaches, yet overemphasizing metrics leads to manipulation, gaming, a myopic focus on short-term goals, and other unexpected negative consequences. This poses a fundamental contradiction for AI development. Through a series of real-world case studies, we look at various aspects of where metrics go wrong in practice and aspects of how our online environment and current business practices are exacerbating these failures. Finally, we propose a framework towards mitigating the harms caused by overemphasis of metrics within AI by: (1) using a slate of metrics to get a fuller and more nuanced picture, (2) combining metrics with qualitative accounts, and (3) involving a range of stakeholders, including those who will be most impacted.

Ensembles, where multiple neural networks are trained individually and their predictions are averaged, have been shown to be widely successful for improving both the accuracy and predictive uncertainty of single neural networks. However, an ensemble's cost for both training and testing increases linearly with the number of networks, which quickly becomes untenable. In this paper, we propose BatchEnsemble, an ensemble method whose computational and memory costs are significantly lower than typical ensembles. BatchEnsemble achieves this by defining each weight matrix to be the Hadamard product of a shared weight among all ensemble members and a rank-one matrix per member. Unlike ensembles, BatchEnsemble is not only parallelizable across devices, where one device trains one member, but also parallelizable within a device, where multiple ensemble members are updated simultaneously for a given mini-batch. Across CIFAR-10, CIFAR-100, WMT14 EN-DE/EN-FR translation, and out-of-distribution tasks, BatchEnsemble yields competitive accuracy and uncertainties as typical ensembles; the speedup at test time is 3X and memory reduction is 3X at an ensemble of size 4. We also apply BatchEnsemble to lifelong learning, where on Split-CIFAR-100, BatchEnsemble yields comparable performance to progressive neural networks while having a much lower computational and memory costs. We further show that BatchEnsemble can easily scale up to lifelong learning on Split-ImageNet which involves 100 sequential learning tasks.

Networks with large receptive field (RF) have shown advanced fitting ability in recent years. In this work, we utilize the short-term residual learning method to improve the performance and robustness of networks for image denoising tasks. Here, we choose a multi-wavelet convolutional neural network (MWCNN), one of the state-of-art networks with large RF, as the backbone, and insert residual dense blocks (RDBs) in its each layer. We call this scheme multi-wavelet residual dense convolutional neural network (MWRDCNN). Compared with other RDB-based networks, it can extract more features of the object from adjacent layers, preserve the large RF, and boost the computing efficiency. Meanwhile, this approach also provides a possibility of absorbing advantages of multiple architectures in a single network without conflicts. The performance of the proposed method has been demonstrated in extensive experiments with a comparison with existing techniques.

In this paper, we provide a unified convergence analysis for a class of shuffling-type gradient methods for solving a well-known finite-sum minimization problem commonly used in machine learning. This algorithm covers various variants such as randomized reshuffling, single shuffling, and cyclic/incremental gradient schemes. We consider two different settings: strongly convex and non-convex problems. Our main contribution consists of new non-asymptotic and asymptotic convergence rates for a general class of shuffling-type gradient methods to solve both non-convex and strongly convex problems. While our rate in the non-convex problem is new (i.e. not known yet under standard assumptions), the rate on the strongly convex case matches (up to a constant) the best-known results. However, unlike existing works in this direction, we only use standard assumptions such as smoothness and strong convexity. Finally, we empirically illustrate the effect of learning rates via a non-convex logistic regression and neural network examples.

Partial-label learning is one of the important weakly supervised learning problems, where each training example is equipped with a set of candidate labels that contains the true label. Most existing methods elaborately designed learning objectives as constrained optimizations that must be solved in specific manners, making their computational complexity a bottleneck for scaling up to big data. The goal of this paper is to propose a novel framework of partial-label learning without implicit assumptions on the model or optimization algorithm. More specifically, we propose a general estimator of the classification risk, theoretically analyze the classifier-consistency, and establish an estimation error bound. We then explore a progressive identification method for approximately minimizing the proposed risk estimator, where the update of the model and identification of true labels are conducted in a seamless manner. The resulting algorithm is model-independent and loss-independent, and compatible with stochastic optimization. Thorough experiments demonstrate it sets the new state of the art.