If you've been hankering for a chance to play with Google Assistant but don't own an Android phone, Google Home or hate the Allo messaging app, we have good news: Google announced today that the standalone Assistant app has begun rolling out for iOS in the UK, Germany and France, following its US debut back in May. The app lets you converse with Google's AI using your voice, which means you'll likely ask it to answer your questions or use it interact with your smart home devices. Assistant can interact with smart lighting and thermostats, as well as handling calls, sending messages, setting reminders and calendar events, playing music (albeit via YouTube) and directing you home. Google says the app is designed to operate on devices running iOS 9.1 and above, which should cover the majority of iPhone and iPad users. As we noted in our recent Google Home breakdown, Assistant utilises Google's search smarts, so it can help you find what you want faster.

The American criminal-justice system's sentencing system is among the fairest and most equitable in the world ... assuming you're wealthy, white and male. Everybody else is generally SOL. During the past three decades, America's prison population has quadrupled to more than 2.3 million people. Of those incarcerated, 58 percent are either black or Latino (despite those groups constituting barely a quarter of the general US population). The racial disparity in America's justice system is both obvious and endemic, which is why some courts have started looking for technological solutions. But can an artificial intelligence really make better sentencing recommendations than the people who designed it?

DenseNet architecture is new, it is a logical extension of ResNet. ResNet architecture has a fundamental building block (Identity) where you merge (additive) a previous layer into a future layer. Reasoning here is by adding additive merges we are forcing the network to learn residuals (errors i.e. diff between some previous layer and current one). In contrast, DenseNet paper proposes concatenating outputs from the previous layers instead of using the summation. Recent work has shown that convolutional networks can be substantially deeper, more accurate, and efficient to train if they contain shorter connections between layers close to the input and those close to the output.

As Earth-observing satellites become more plentiful and climate models more powerful, researchers who study global warming are facing a deluge of data. Some are now turning to the latest trend in artificial intelligence (AI) to help trawl through all the information, in the hope of discovering new climate patterns and improving forecasts. "Climate is now a data problem," says Claire Monteleoni, a computer scientist at George Washington University in Washington DC who has helped to pioneer the marriage of machine-learning techniques with climate science. In machine learning, AI systems improve in performance as the amount of data that they analyse grows. This approach is a natural fit for climate science: a single run of a high-resolution climate model can produce a petabyte of data, and the archive of climate data maintained by the UK Met Office, the national weather service, now holds about 45 petabytes of information--and adds 0.085 petabytes a day.

Researchers at the Massachusetts Institute of Technology's Computer Science and Artificial Intelligence Laboratory have developed a machine learning algorithm that leverages large amounts of intensive care unit (ICU) data to predict actionable interventions for patients and improve health outcomes. By tapping into an MIT database of de-identified data for 40,000 critical care patients--including demographics, laboratory tests, medications and vital signs--the research team is able to use deep learning to determine what kinds of treatments are needed for different symptoms. The approach--called ICU Intervene--was presented in a paper this past weekend at the Machine Learning for Healthcare Conference in Boston. According to the authors, their model is the first to use deep neural networks to predict both onset and weaning of interventions using all available modalities of ICU data. "The decisions that are made in the ICU are made in a particularly high-stress and high-demand environment," says Harini Suresh, a PhD student and lead author on the paper, who adds that clinicians in these situations are bombarded with different types of data for many patients and as a result it can be difficult to make real-time treatment decisions.

Many of today's machine learning (ML) systems are not built from scratch, but are compositions of an array of {\em modular learning components} (MLCs). The increasing use of MLCs significantly simplifies the ML system development cycles. However, as most MLCs are contributed and maintained by third parties, their lack of standardization and regulation entails profound security implications. In this paper, for the first time, we demonstrate that potentially harmful MLCs pose immense threats to the security of ML systems. We present a broad class of {\em logic-bomb} attacks in which maliciously crafted MLCs trigger host systems to malfunction in a predictable manner. By empirically studying two state-of-the-art ML systems in the healthcare domain, we explore the feasibility of such attacks. For example, we show that, without prior knowledge about the host ML system, by modifying only 3.3{\textperthousand} of the MLC's parameters, each with distortion below $10^{-3}$, the adversary is able to force the misdiagnosis of target victims' skin cancers with 100\% success rate. We provide analytical justification for the success of such attacks, which points to the fundamental characteristics of today's ML models: high dimensionality, non-linearity, and non-convexity. The issue thus seems fundamental to many ML systems. We further discuss potential countermeasures to mitigate MLC-based attacks and their potential technical challenges.

Commonsense reasoning is in principle a central problem in artificial intelligence, but it is a very difficult one. One approach that has been pursued since the earliest days of the field has been to encode commonsense knowledge as statements in a logic-based representation language and to implement commonsense reasoning as some form of logical inference. This paper surveys the use of logic-based representations of commonsense knowledge in artificial intelligence research.

Convolutional dictionary learning (CDL or sparsifying CDL) has many applications in image processing and computer vision. There has been growing interest in developing efficient algorithms for CDL, mostly relying on the augmented Lagrangian (AL) method or the variant alternating direction method of multipliers (ADMM). When their parameters are properly tuned, AL methods have shown fast convergence in CDL. However, the parameter tuning process is not trivial due to its data dependence and, in practice, the convergence of AL methods depends on the AL parameters for nonconvex CDL problems. To moderate these problems, this paper proposes a new practically feasible and convergent Block Proximal Gradient method using a Majorizer (BPG-M) for CDL. The BPG-M-based CDL is investigated with different block updating schemes and majorization matrix designs, and further accelerated by incorporating some momentum coefficient formulas and restarting techniques. All of the methods investigated incorporate a boundary artifacts removal (or, more generally, sampling) operator in the learning model. Numerical experiments show that, without needing any parameter tuning process, the proposed BPG-M approach converges more stably to desirable solutions of lower objective values than the existing state-of-the-art ADMM algorithm and its memory-efficient variant do. Compared to the ADMM approaches, the BPG-M method using a multi-block updating scheme is particularly useful in single-threaded CDL algorithm handling large datasets, due to its lower memory requirement and no polynomial computational complexity. Image denoising experiments show that, for relatively strong additive white Gaussian noise, the filters learned by BPG-M-based CDL outperform those trained by the ADMM approach.

As a new way of training generative models, Generative Adversarial Nets (GAN) that uses a discriminative model to guide the training of the generative model has enjoyed considerable success in generating real-valued data. However, it has limitations when the goal is for generating sequences of discrete tokens. A major reason lies in that the discrete outputs from the generative model make it difficult to pass the gradient update from the discriminative model to the generative model. Also, the discriminative model can only assess a complete sequence, while for a partially generated sequence, it is non-trivial to balance its current score and the future one once the entire sequence has been generated. In this paper, we propose a sequence generation framework, called SeqGAN, to solve the problems. Modeling the data generator as a stochastic policy in reinforcement learning (RL), SeqGAN bypasses the generator differentiation problem by directly performing gradient policy update. The RL reward signal comes from the GAN discriminator judged on a complete sequence, and is passed back to the intermediate state-action steps using Monte Carlo search. Extensive experiments on synthetic data and real-world tasks demonstrate significant improvements over strong baselines.

We investigate task clustering for deep-learning based multi-task and few-shot learning in a many-task setting. We propose a new method to measure task similarities with cross-task transfer performance matrix for the deep learning scenario. Although this matrix provides us critical information regarding similarity between tasks, its asymmetric property and unreliable performance scores can affect conventional clustering methods adversely. Additionally, the uncertain task-pairs, i.e., the ones with extremely asymmetric transfer scores, may collectively mislead clustering algorithms to output an inaccurate task-partition. To overcome these limitations, we propose a novel task-clustering algorithm by using the matrix completion technique. The proposed algorithm constructs a partially-observed similarity matrix based on the certainty of cluster membership of the task-pairs. We then use a matrix completion algorithm to complete the similarity matrix. Our theoretical analysis shows that under mild constraints, the proposed algorithm will perfectly recover the underlying "true" similarity matrix with a high probability. Our results show that the new task clustering method can discover task clusters for training flexible and superior neural network models in a multi-task learning setup for sentiment classification and dialog intent classification tasks. Our task clustering approach also extends metric-based few-shot learning methods to adapt multiple metrics, which demonstrates empirical advantages when the tasks are diverse.