We investigate whether model extraction can be used to "steal" the weights of sequential recommender systems, and the potential threats posed to victims of such attacks. This type of risk has attracted attention in image and text classification, but to our knowledge not in recommender systems. We argue that sequential recommender systems are subject to unique vulnerabilities due to the specific autoregressive regimes used to train them. Unlike many existing recommender attackers, which assume the dataset used to train the victim model is exposed to attackers, we consider a data-free setting, where training data are not accessible. Under this setting, we propose an API-based model extraction method via limited-budget synthetic data generation and knowledge distillation. We investigate state-of-the-art models for sequential recommendation and show their vulnerability under model extraction and downstream attacks. We perform attacks in two stages. (1) Model extraction: given different types of synthetic data and their labels retrieved from a black-box recommender, we extract the black-box model to a white-box model via distillation. (2) Downstream attacks: we attack the black-box model with adversarial samples generated by the white-box recommender. Experiments show the effectiveness of our data-free model extraction and downstream attacks on sequential recommenders in both profile pollution and data poisoning settings.

We consider the query recommendation problem in closed loop interactive learning settings like online information gathering and exploratory analytics. The problem can be naturally modelled using the Multi-Armed Bandits (MAB) framework with countably many arms. The standard MAB algorithms for countably many arms begin with selecting a random set of candidate arms and then applying standard MAB algorithms, e.g., UCB, on this candidate set downstream. We show that such a selection strategy often results in higher cumulative regret and to this end, we propose a selection strategy based on the maximum utility of the arms. We show that in tasks like online information gathering, where sequential query recommendations are employed, the sequences of queries are correlated and the number of potentially optimal queries can be reduced to a manageable size by selecting queries with maximum utility with respect to the currently executing query. Our experimental results using a recent real online literature discovery service log file demonstrate that the proposed arm selection strategy improves the cumulative regret substantially with respect to the state-of-the-art baseline algorithms.

Google's Nest Audio has been one of our favorite smart speakers since it came out almost a year ago. When compared to other $100 devices, it packs a lot of value and will be especially attractive for those who already use the Google Assistant a lot. But now you can grab the speaker for even less because Best Buy and B&H Photo have the Nest Audio for only $80. While we did see the speaker drop to $75 ahead of Amazon Prime Day back in June, this is the best price we've seen since then. The Nest Audio is Google's answer to Amazon's Echo and Apple's HomePod mini and it holds its own against both of those devices. We like its attractive, minimalist design and you have five colors to choose from, so you'll likely find one that fits well with the rest of your home decor.

Edge computing, as known to many, has surpassed the expectation levels in terms of performance delivered and objectives achieved. Over the last couple of years, it has been a common scenario to observe companies making huge tech investments as a part of their digital transformation journey. On the same lines, cloud companies see new opportunities by fusing Edge computing and AI, or Edge AI. This form of AI has left everyone spellbound with the impact it has left behind. Virtual assistants like Amazon's Alexa and Apple's Siri that have dominated the world of late are a result of Edge AI Here is everything you need to know about Edge AI and beyond. To know about Edge AI, it is important to know about AI and Edge computing individually for the sole reason that the amalgamation of the two would result in Edge AI.

User behavior has been validated to be effective in revealing personalized preferences for commercial recommendations. However, few user-item interactions can be collected for new users, which results in a null space for their interests, i.e., the cold-start dilemma. In this paper, a two-tower framework, namely, the model-agnostic interest learning (MAIL) framework, is proposed to address the cold-start recommendation (CSR) problem for recommender systems. In MAIL, one unique tower is constructed to tackle the CSR from a zero-shot view, and the other tower focuses on the general ranking task. Specifically, the zero-shot tower first performs cross-modal reconstruction with dual auto-encoders to obtain virtual behavior data from highly aligned hidden features for new users; and the ranking tower can then output recommendations for users based on the completed data by the zero-shot tower. Practically, the ranking tower in MAIL is model-agnostic and can be implemented with any embedding-based deep models. Based on the co-training of the two towers, the MAIL presents an end-to-end method for recommender systems that shows an incremental performance improvement. The proposed method has been successfully deployed on the live recommendation system of NetEase Cloud Music to achieve a click-through rate improvement of 13% to 15% for millions of users. Offline experiments on real-world datasets also show its superior performance in CSR. Our code is available.

Amazon might have made your choice of true wireless earbuds a little easier. The internet retailer is running a sale on the second-generation Echo Buds that lowers the price to $90 for the standard version, and $105 for the model with a wireless charging case. That makes them less expensive than many no-frills earbuds, let alone ones with comparable features like active noise cancellation. These aren't the absolute best-sounding earbuds you'll buy, but they pair solid quality with perks difficult to find even at their normal prices, including ANC, IPX4 water resistance and built-in Alexa support. These may be just the ticket if you're interested in hushing the outside world or adding a soundtrack to your workouts.

Artificial intelligence and machine learning are two popular and often hyped terms these days. And people often use them interchangeably to describe an intelligent software or system. But even though both AI and ML are based on statistics and mathematics, they are not the same thing. In this article, you will learn the differences between AI and ML with some practical examples to help clear up any confusion. Artificial intelligence, or AI, is the ability of a computer or machine to mimic or imitate human intelligent behavior and perform human-like tasks.

AI technology is radically changing the outlook for the ecommerce sector in a number of very promising ways.

Our personal devices would know us more than we know ourselves. They might even increase our life span. That Artificial Intelligence (AI) will change the ways of the world in the 2020s is a foregone conclusion. Perhaps its greatest--and most defining--impact would be felt on personal devices and the way humans interact with them. 'Emotion AI' systems are becoming so nuanced and powerful that our devices will soon know more about our emotional being than our friends or family ever did.

As we enter another new year, there are a fresh set of forecasts, hopes, and fears about the path of digital marketing in the coming year and beyond. The recurrence of two themes: Artificial Intelligence (AI) and Machine Learning (ML), is one of the fascinating topics found across numerous trends. Marketing automation, digital creativity, customization, and marketing technology are just a few of the areas where these two closely related topics appear to be gaining traction. When it comes to finding new information, 84 percent of individuals say they look to the internet first. Mobile devices account for 50% of all search inquiries.