Google's latest Chromecast streaming media dongle is a bit different. With a full interface and a remote, the new Chromecast with Google TV costs £59.99 and sits above the basic £30 Chromecast. You can still Google Cast to the new device, but the new flat plastic dongle is more than just a simple receiver, running the full Android TV software similar to the Nvidia Shield or smart TVs from Sony and others. Once plugged in, the new Chromecast is set up using the Google Home app on an Android, iPhone or iPad in about five minutes. Scan the QR code on your TV, log in with the required Google account, and choose some apps to install.

With the prevalence of Internet of Things (IoT)-based social media applications, the distance among people has been greatly shortened. As a result, recommender systems in IoT-based social media need to be developed oriented to groups of users rather than individual users. However, existing methods were highly dependent on explicit preference feedbacks, ignoring scenarios of implicit feedback. To remedy such gap, this paper proposes an implicit feedback-based group recommender system using probabilistic inference and non-cooperative game(GREPING) for IoT-based social media. Particularly, unknown process variables can be estimated from observable implicit feedbacks via Bayesian posterior probability inference. In addition, the globally optimal recommendation results can be calculated with the aid of non-cooperative game. Two groups of experiments are conducted to assess the GREPING from two aspects: efficiency and robustness. Experimental results show obvious promotion and considerable stability of the GREPING compared to baseline methods.

In the last few years, three major topics received increased interest: deep learning, NLP and conversational agents. Bringing these three topics together to create an amazing digital customer experience and indeed deploy in production and solve real-world problems is something innovative and disruptive. We introduce a new Portuguese financial domain language representation model called BERTa\'u. BERTa\'u is an uncased BERT-base trained from scratch with data from the Ita\'u virtual assistant chatbot solution. Our novel contribution is that BERTa\'u pretrained language model requires less data, reached state-of-the-art performance in three NLP tasks, and generates a smaller and lighter model that makes the deployment feasible. We developed three tasks to validate our model: information retrieval with Frequently Asked Questions (FAQ) from Ita\'u bank, sentiment analysis from our virtual assistant data, and a NER solution. All proposed tasks are real-world solutions in production on our environment and the usage of a specialist model proved to be effective when compared to Google BERT multilingual and the DPRQuestionEncoder from Facebook, available at Hugging Face. The BERTa\'u improves the performance in 22% of FAQ Retrieval MRR metric, 2.1% in Sentiment Analysis F1 score, 4.4% in NER F1 score and can also represent the same sequence in up to 66% fewer tokens when compared to "shelf models".

With the emergence of personality computing as a new research field related to artificial intelligence and personality psychology, we have witnessed an unprecedented proliferation of personality-aware recommendation systems. Unlike conventional recommendation systems, these new systems solve traditional problems such as the cold start and data sparsity problems. This survey aims to study and systematically classify personality-aware recommendation systems. To the best of our knowledge, this survey is the first that focuses on personality-aware recommendation systems. We explore the different design choices of personality-aware recommendation systems, by comparing their personality modeling methods, as well as their recommendation techniques. Furthermore, we present the commonly used datasets and point out some of the challenges of personality-aware recommendation systems.

Back at its fall hardware event last year, Amazon announced the Echo Show 10. And now it finally has a release date. You'll be able to pick the company's latest smart display on February 25th for $250. In case you need a refresher, the Echo Show 10 is one of Amazon's strangest devices to date. It features a rotating base and a computer vision algorithm that allows its 10-inch screen to reposition itself and face you head-on wherever you are in a room.

People have long used Tasker to take care of repetitive tasks on their Android device, or to customize its features based on things like whether they're at home or at the office. Now the app's features are a little easier to use since you can trigger them via Google Assistant. XDA points out a post by the developer on Reddit where he points out the currently available triggers, which you can use to run your favorite automations by name. Tasker is an incredibly powerful utility, but it can be a bit complex and intimidating, and voice control could be the difference in making it usable on a regular basis around your home. If you have the Google Play Pass subscription then access is free, and there's also a seven day trial available, otherwise it costs $3.49 in the Play Store.

Amazon's Alexa Guard Plus subscription service is now live in the US. The company unveiled Guard Plus back in September as a paid option for those who want the extra features that come with the premium version of its free Guard service. While the basic Guard feature can already turn Echo smart speakers and displays into home security devices, its premium version takes things a step further by giving subscribers hands-free access to emergency services and giving Alexa the power to deter intruders from breaking in. Customers who pay for a subscription will be able to ask Alexa to call Emergency Helpline for them to request medical, fire or police assistance. The service also gives Alexa the capability to detect sounds of activity in the house if its residents are away and to sound a siren from Echo devices if it does. Alexa could also play the sounds of dogs barking from the speakers if connected outdoor security cameras detect motion outside the door.

Five years ago, the world of artificial intelligence--and the algorithms it runs on--looked very different. Asking your Google Home to play Adele's chart-topping single wasn't possible yet. IBM Watson was still widely considered a beacon for AI advancement, and DeepMind's AI victory over a human at Go was still fresh. Machine learning engineers were facing earlier versions of today's image classification and speech recognition challenges. And though most tech giants hadn't earmarked corporate funding for ethical AI, the conversation was becoming more mainstream as the impact of algorithms on human lives became clearer.

Sony's top of the line noise-cancelling headphones have long had a winning formula and the latest edition has a much-requested addition – multiple device connectivity – to make them the best of class. The WH-1000XM4 have an RRP of £350 and on initial inspection little has changed for the fourth edition of the 1000X line, with its understated design. The high-quality plastic body is well made and lightweight at 254g but doesn't feel as premium as some metal or carbon fibre competitors that weigh more than 300g. They are some of the lightest-feeling headphones you can buy, matching the longstanding comfort kings, the Bose QC35 II. The ear cups are well padded with a gentle, even pressure on the side of your head while a soft leatherette headband sits on your dome.

Textual explanations have proved to help improve user satisfaction on machine-made recommendations. However, current mainstream solutions loosely connect the learning of explanation with the learning of recommendation: for example, they are often separately modeled as rating prediction and content generation tasks. In this work, we propose to strengthen their connection by enforcing the idea of sentiment alignment between a recommendation and its corresponding explanation. At training time, the two learning tasks are joined by a latent sentiment vector, which is encoded by the recommendation module and used to make word choices for explanation generation. At both training and inference time, the explanation module is required to generate explanation text that matches sentiment predicted by the recommendation module. Extensive experiments demonstrate our solution outperforms a rich set of baselines in both recommendation and explanation tasks, especially on the improved quality of its generated explanations. More importantly, our user studies confirm our generated explanations help users better recognize the differences between recommended items and understand why an item is recommended.