Researchers from Google's AI division and Harvard University have created an AI model capable of predicting the location of aftershocks up to one year after a major earthquake. The model was trained with 199 major earthquake events in recent decades followed by 130,000 aftershocks, and was found to be more accurate than a method used to predict aftershocks today. Aftershocks included in the dataset used to train the neural network took place in a perimeter that stretches 50 kilometers vertically and 100 kilometers horizontally from each earthquake epicenter. "We found that after feeding these model stress changes into the neural network, the neural network could sort of predict aftershock locations in the testing dataset more accurately that the sort of baseline Coulomb failure stress change criterion that's used a lot in studies of aftershock locations," Phoebe DeVries of the Department of Earth and Planetary Sciences at Harvard University told VentureBeat in a phone interview. Data used to train the model came from noteworthy earthquakes such as the 2004 Sumatra earthquake, the 2011 earthquake in Japan, the 1989 Loma Prieta earthquake in the San Francisco Bay Area, and the 1994 Northride earthquake near Los Angeles.

"In two years, China will be ahead of the United States in AI (artificial intelligence)," states Denis Barrier, CEO of global venture firm Cathay Innovation. If so, China will largely determine how this technology transforms the world. Today's contest is more than a race for dominance in a new technology -- it's one between authoritarianism and democracy. "AI is the world's next big inflection point," says Ajeet Singh, CEO of ThoughtSpot in Palo Alto. Artificial intelligence is machine learning, which self-learns programmed tasks, using data, and the more it gets, the more learned it becomes.

It turns out it doesn't take much for hackers to see what's being displayed on your computer screen. A team of researchers have discovered that ultrasonic sounds picked up by a webcam microphone can be analyzed using machine learning to determine what's being shown on a remote computer screen. The attack could allow anyone savvy enough to spy on someone's browsing activity with ease. Researchers found that ultrasonic sounds picked up by a webcam mic can be analyzed using machine learning to determine what's being shown on a remote computer screen Researchers observed frequencies inaudible to the human ear being emitted from LCD displays. They found that microphones from webcams, voice-activated speakers and smartphones were able to record those noises.

When Argho Chatterjee decided to pursue UpGrad and IIIT Bangalore's PG Program in Machine Learning and Artificial Intelligence, he knew he was diving straight into coding his own artificial neural networks, and had a fair idea that this technology could help him solve real-world problems. What came as a pleasant surprise was that he had the access to a personalised learning environment provided by the prestigious institute through its partnership with distinguished online education venture – UpGrad. The two institutes have been working seamlessly to provide learners with an advanced curriculum, projects created in collaboration with the industry experts, and tailor-made support for AI career choices. In fact, the acclaimed degree went on to help Argho make a transition to the role of a Data Scientist ( Deep Learning (AI)) at Samsung R&D with 80% CTC hike! Learning in a personalised environment under great faculty, Argho brushed up on the basics, imbibed conceptual knowledge, and acquired full-fledged knowledge of the field.

Meta-learning is increasingly used to support the recommendation of machine learning algorithms and their configurations. Such recommendations are made based on meta-data, consisting of performance evaluations of algorithms on prior datasets, as well as characterizations of these datasets. These characterizations, also called meta-features, describe properties of the data which are predictive for the performance of machine learning algorithms trained on them. Unfortunately, despite being used in a large number of studies, meta-features are not uniformly described and computed, making many empirical studies irreproducible and hard to compare. This paper aims to remedy this by systematizing and standardizing data characterization measures used in meta-learning, and performing an in-depth analysis of their utility. Moreover, it presents MFE, a new tool for extracting meta-features from datasets and identify more subtle reproducibility issues in the literature, proposing guidelines for data characterization that strengthen reproducible empirical research in meta-learning.

Ensemble learning is a method of combining multiple trained models to improve the model accuracy. We introduce the usage of such methods, specifically ensemble average inside Convolutional Neural Networks (CNNs) architectures. By Inner Average Ensemble (IEA) of multiple convolutional neural layers (CNLs) replacing the single CNLs inside the CNN architecture, the accuracy of the CNN increased. A visual and a similarity score analysis of the features generated from IEA explains why it boosts the model performance. Empirical results using different benchmarking datasets and well-known deep model architectures shows that IEA outperforms the ordinary CNL used in CNNs.

Generative Adversarial Networks (GANs) have been shown to produce realistically looking synthetic images with remarkable success, yet their performance seems less impressive when the training set is highly diverse. In order to provide a better fit to the target data distribution when the dataset includes many different classes, we propose a variant of the basic GAN model, called Gaussian Mixture GAN (GM-GAN), where the probability distribution over the latent space is a mixture of Gaussians. We also propose a supervised variant which is capable of conditional sample synthesis. In order to evaluate the model's performance, we propose a new scoring method which separately takes into account two (typically conflicting) measures - diversity vs. quality of the generated data. Through a series of empirical experiments, using both synthetic and real-world datasets, we quantitatively show that GM-GANs outperform baselines, both when evaluated using the commonly used Inception Score, and when evaluated using our own alternative scoring method. In addition, we qualitatively demonstrate how the \textit{unsupervised} variant of GM-GAN tends to map latent vectors sampled from different Gaussians in the latent space to samples of different classes in the data space. We show how this phenomenon can be exploited for the task of unsupervised clustering, and provide quantitative evaluation showing the superiority of our method for the unsupervised clustering of image datasets. Finally, we demonstrate a feature which further sets our model apart from other GAN models: the option to control the quality-diversity trade-off by altering, post-training, the probability distribution of the latent space. This allows one to sample higher quality and lower diversity samples, or vice versa, according to one's needs.

Convolutional Neural Networks (CNNs) models have been recently introduced in the domain of top-$N$ session-based recommendations. An ordered collection of past items the user has interacted with in a session (or sequence) are embedded into a 2-dimensional latent matrix, and treated as an image. The convolution and pooling operations are then applied to the mapped item embeddings. In this paper, we first examine the typical session-based CNN recommender and show that both the generative model and network architecture are suboptimal when modeling long-range dependencies in the item sequence. To address the issues, we propose a simple, but very effective generative model that is capable of learning high-level representation from both short- and long-range dependencies. The network architecture of the proposed model is formed of a stack of holed convolutional layers, which can efficiently increase the receptive fields without relying on the pooling operation. Another contribution is the effective use of residual block structure in recommender systems, which can ease the optimization for much deeper networks. The proposed generative model attains state-of-the-art accuracy with less training time in the session-based recommendation task. It accordingly can be used as a powerful session-based recommendation baseline to beat in future, especially when there are long sequences of user feedback.

Multimodal learning has shown promising performance in content-based recommendation due to the auxiliary user and item information of multiple modalities such as text and images. However, the problem of incomplete and missing modality is rarely explored and most existing methods fail in learning a recommendation model with missing or corrupted modalities. In this paper, we propose LRMM, a novel framework that mitigates not only the problem of missing modalities but also more generally the cold-start problem of recommender systems. We propose modality dropout (m-drop) and a multimodal sequential autoencoder (m-auto) to learn multimodal representations for complementing and imputing missing modalities. Extensive experiments on real-world Amazon data show that LRMM achieves state-of-the-art performance on rating prediction tasks. More importantly, LRMM is more robust to previous methods in alleviating data-sparsity and the cold-start problem.

Empirical risk minimization (ERM), with proper loss function and regularization, is the common practice of supervised classification. In this paper, we study training arbitrary (from linear to deep) binary classifier from only unlabeled (U) data by ERM but not by clustering in the geometric space. A two-step ERM is considered: first an unbiased risk estimator is designed, and then the empirical training risk is minimized. This approach is advantageous in that we can also evaluate the empirical validation risk, which is indispensable for hyperparameter tuning when some validation data is split from U training data instead of labeled test data. We prove that designing such an estimator is impossible given a single set of U data, but it becomes possible given two sets of U data with different class priors. This answers a fundamental question in weakly-supervised learning, namely what the minimal supervision is for training any binary classifier from only U data. Since the proposed learning method is based on unbiased risk estimates, the asymptotic consistency of the learned classifier is certainly guaranteed. Experiments demonstrate that the proposed method could successfully train deep models like ResNet and outperform state-of-the-art methods for learning from two sets of U data.