Artificial intelligence is already everywhere. The technology is widely used in ways that are quite obvious, such as self-driving cars, and others that are inconspicuous. It keeps your mobile phone ticking over, translates for Alexa, helps doctors analyse medical images, controls robotics in factories, and so much more, quietly working behind the scenes to automate both simple and complicated tasks. Though these are small examples of its capabilities, AI is predicted to have a huge impact on our lives, with plenty predicting disruption to our jobs and work life and others seeing the benefits of churning through vast data sets. Keeping up with such changes requires understanding the various technologies behind AI, be it neural networks, deep learning and machine learning, and seeing how they're already being used.

Our unique multi-level hierarchy gives the James platform the ability to handle thousands of intents. Our ensemble of prediction models can interpret user intents with as little as 10 training messages - a feature unlikely to be found in other solutions.

Dating in 2018 can be a challenge. I'm sorry, let me rephrase: It suuuuuuuuccckkkkksssss. Apps like Tinder, Bumble, Hinge, Grindr, and others are the dater's tools of choice, and yet hating them is the one thing we can all agree on these days. They're often more hazard than help, and the forced psychoanalysis of every picture and witty answer can shake even the most durable of confidences loose. Why am I not getting more matches? But is it your fault, or the app's?

Item-item collaborative filtering (CF) models are a well known and studied family of recommender systems, however current literature does not provide any theoretical explanation of the conditions under which item-based recommendations will succeed or fail. We investigate the existence of an ideal item-based CF method able to make perfect recommendations. This CF model is formalized as an eigenvalue problem, where estimated ratings are equivalent to the true (unknown) ratings multiplied by a user-specific eigenvalue of the similarity matrix. Preliminary experiments show that the magnitude of the eigenvalue is proportional to the accuracy of recommendations for that user and therefore it can provide reliable measure of confidence.

Following recent successes in exploiting both latent factor and word embedding models in recommendation, we propose a novel Regularized Multi-Embedding (RME) based recommendation model that simultaneously encapsulates the following ideas via decomposition: (1) which items a user likes, (2) which two users co-like the same items, (3) which two items users often co-liked, and (4) which two items users often co-disliked. In experimental validation, the RME outperforms competing state-of-the-art models in both explicit and implicit feedback datasets, significantly improving Recall@5 by 5.9~7.0%, NDCG@20 by 4.3~5.6%, and MAP@10 by 7.9~8.9%. In addition, under the cold-start scenario for users with the lowest number of interactions, against the competing models, the RME outperforms NDCG@5 by 20.2% and 29.4% in MovieLens-10M and MovieLens-20M datasets, respectively. Our datasets and source code are available at: https://github.com/thanhdtran/RME.git.

Cold-start is a very common and still open problem in the Recommender Systems literature. Since cold start items do not have any interaction, collaborative algorithms are not applicable. One of the main strategies is to use pure or hybrid content-based approaches, which usually yield to lower recommendation quality than collaborative ones. Some techniques to optimize performance of this type of approaches have been studied in recent past. One of them is called feature weighting, which assigns to every feature a real value, called weight, that estimates its importance. Statistical techniques for feature weighting commonly used in Information Retrieval, like TF-IDF, have been adapted for Recommender Systems, but they often do not provide sufficient quality improvements. More recent approaches, FBSM and LFW, estimate weights by leveraging collaborative information via machine learning, in order to learn the importance of a feature based on other users opinions. This type of models have shown promising results compared to classic statistical analyzes cited previously. We propose a novel graph, feature-based machine learning model to face the cold-start item scenario, learning the relevance of features from probabilities of item-based collaborative filtering algorithms.

Today, Google rolled out a new ability for the Google Assistant. The helpful, disembodied entity that lives inside smartphones and Google Home devices can now interpret two languages at the same time, including French, German, Japanese, Spanish, Italian and English. But, how will users know about it? It's a question I recently encountered in my own personal experience. A few weeks ago, the familiar command that turns my Philips Hue lights on and off stopped working."OK, Google, turn off the light in the living room."

File this one under "Why has no one thought of its before?" And it packs a pretty good audio punch, based on the demo we got last week, thanks in part to a collaboration between Netgear and home audio powerhouse Harman Kardon. Mesh networks like Netgear's Orbi series are popular because they can blanket a home with strong Wi-Fi and eliminate dead zones. They consist of a router that you connect to your broadband gateway, and one or more satellite access points that you position around the house. Data hops from one satellite to the other and back to the router, with client devices automatically connecting to whichever access points can deliver the strongest signal.

The Google Home speaker is adding a bilingual Google Assistant feature. Hey, Google will be speaking your language more like you do if you're a bilingual home. The voice-activated Google Assistant inside Google Home smart speakers, Android smartphones and iPhones, too, will be able to recognize queries made in any pairing of English, German, French, Spanish, Italian and Japanese, with more languages to follow, and, assuming it works as promised, it will respond to requests in the language used for the question. The new feature, which Google claims as a first of its kind, began rolling out Thursday. . As Google explains, if you're listening for an answer in English, you can ask something along the lines of, "Hey Google, what's the weather like today?"

Google's Assistant is picking up the ability to speak with you in two languages without having to switch accounts. Now Google Home and Android smartphone owners will be able to speak in any two of the following languages: English, Spanish, French, German, Italian and Japanese. The Google Assistant will reply in the language of the query it's answering. The company first mentioned it was working on this feature in February, but there hadn't been an update on it for months. The new feature helps Google Assistant serve bilingual households, which make up an increasing percentage of American families.