The applications of artificial intelligence have grown exponentially over the past decade. Here are some examples of artificial intelligence at work today. The words artificial intelligence may seem like a far-off concept that has nothing to do with us. But the truth is that we encounter several examples of artificial intelligence in our daily lives. From Netflix's movie recommendation to Amazon's Alexa, we now rely on various AI models without knowing it.

Personalized recommendation based on multi-arm bandit (MAB) algorithms has shown to lead to high utility and efficiency as it can dynamically adapt the recommendation strategy based on feedback. However, unfairness could incur in personalized recommendation. In this paper, we study how to achieve user-side fairness in personalized recommendation. We formulate our fair personalized recommendation as a modified contextual bandit and focus on achieving fairness on the individual whom is being recommended an item as opposed to achieving fairness on the items that are being recommended. We introduce and define a metric that captures the fairness in terms of rewards received for both the privileged and protected groups. We develop a fair contextual bandit algorithm, Fair-LinUCB, that improves upon the traditional LinUCB algorithm to achieve group-level fairness of users. Our algorithm detects and monitors unfairness while it learns to recommend personalized videos to students to achieve high efficiency. We provide a theoretical regret analysis and show that our algorithm has a slightly higher regret bound than LinUCB. We conduct numerous experimental evaluations to compare the performances of our fair contextual bandit to that of LinUCB and show that our approach achieves group-level fairness while maintaining a high utility.

Large scale recommender models find most relevant items from huge catalogs, and they play a critical role in modern search and recommendation systems. To model the input space with large-vocab categorical features, a typical recommender model learns a joint embedding space through neural networks for both queries and items from user feedback data. However, with millions to billions of items, the power-law user feedback makes labels very sparse for a large amount of long-tail items. Inspired by the recent success in self-supervised representation learning research in both computer vision and natural language understanding, we propose a multi-task self-supervised learning (SSL) framework for large-scale item recommendations. The framework is designed to tackle the label sparsity problem by learning more robust item representations. Furthermore, we propose two self-supervised tasks applicable to models with categorical features within the proposed framework: (i) Feature Masking (FM) and (ii) Feature Dropout (FD). We evaluate our framework using two large-scale datasets with 500M and 1B training examples respectively. Our results demonstrate that the proposed framework outperforms traditional supervised learning only models and state-of-the-art regularization techniques in the context of item recommendations. The SSL framework shows larger improvement with less supervision compared to the counterparts. We also apply the proposed techniques to a web-scale commercial app-to-app recommendation system, and significantly improve top-tier business metrics via A/B experiments on live traffic. Our online results also verify our hypothesis that our framework indeed improves model performance on slices that lack supervision.

The migration of conversational AI agents across different embodiments in order to maintain the continuity of the task has been recently explored to further improve user experience. However, these migratable agents lack contextual understanding of the user information and the migrated device during the dialog conversations with the user. This opens the question of how an agent might behave when migrated into an embodiment for contextually predicting the next utterance. We collected a dataset from the dialog conversations between crowdsourced workers with the migration context involving personal and non-personal utterances in different settings (public or private) of embodiment into which the agent migrated. We trained the generative and information retrieval models on the dataset using with and without migration context and report the results of both qualitative metrics and human evaluation. We believe that the migration dataset would be useful for training future migratable AI systems.

Due to the shallow structure, classic graph neural networks (GNNs) failed in modelling high-order graph structures that deliver critical insights of task relevant relations. The negligence of those insights lead to insufficient distillation of collaborative signals in recommender systems. In this paper, we propose PEAGNN, a unified GNN framework tailored for recommendation tasks, which is capable of exploiting the rich semantics in metapaths. PEAGNN trains multilayer GNNs to perform metapath-aware information aggregation on collaborative subgraphs, $h$-hop subgraphs around the target user-item pairs. After the attentive fusion of aggregated information from different metapaths, a graph-level representation is then extracted for matching score prediction. To leverage the local structure of collaborative subgraphs, we present entity-awareness that regularizes node embedding with the presence of features in a contrastive manner. Moreover, PEAGNN is compatible with the mainstream GNN structures such as GCN, GAT and GraphSage. The empirical analysis on three public datasets demonstrate that our model outperforms or is at least on par with other competitive baselines. Further analysis indicates that trained PEAGNN automatically derives meaningful metapath combinations from the given metapaths.

Artificial intelligence (AI) is creeping into our everyday lives, often without us realizing it. Today, AI can be found in the digital assistants we use such as Apple's (NASDAQ:AAPL) Siri and Amazon's (NASDAQ:AMZN) Alexa to check our schedules and search for things on the internet; in the cars we own that now park themselves as they are able to recognize space around the vehicle; and in the small robots we use to clean our houses, such as the Roomba vacuum. Artificial intelligence is becoming more a part of our lives all the time, and will only grow in importance in coming years. In the not too distant future, AI will influence everything from how we shop for groceries to how diseases are diagnosed and treated by doctors. It all adds up to a fast growing market.

Following the launch of investigations last year, the U.S. Department of Justice (DOJ) together with attorney generals from 11 U.S. states filed a lawsuit against Google on Tuesday alleging that the company maintains monopolies in online search and advertising, and violates laws prohibiting anticompetitive business practices. It's the first antitrust lawsuit federal prosecutors filed against a tech company since the Department of Justice brought charges against Microsoft in the 1990s. "Back then, Google claimed Microsoft's practices were anticompetitive, and yet, now, Google deploys the same playbook to sustain its own monopolies," the complaint reads. "For the sake of American consumers, advertisers, and all companies now reliant on the internet economy, the time has come to stop Google's anticompetitive conduct and restore competition." Attorneys general from no Democratic states joined the suit.

Machine Learning is an algorithmic approach of creating computer models with the ability to learn and adapt from a given data-set, these models can then be used to make useful predictions of results against similar but never-seen-before data. It is often referred to as a subset of Artificial Intelligence and forms the very base on which AI models are created. The concept of Machine Learning is based on the idea that whether machines can be designed to imitate human behaviour of learning, adapting skills, and applying where necessary. Just like all living beings who learn from every experience in life and take future decisions, similarly, the Machine Learning approach creates models that are first trained to'learn' on a data-set distribution. The trained models predict results by applying the knowledge learned during training with reasonable high accuracy.

From instantaneous translation to conversational interfaces, artificial-intelligence (AI) technologies are making ever more evident impacts on our lives. This is particularly true in the financial-services sector, where challengers are already launching disruptive AI-powered innovations. To remain competitive, incumbent banks must become "AI first" in vision and execution, and as discussed in our previous article, 1 1. Suparna Biswas, Brant Carson, Violet Chung, Shwaitang Singh, and Renny Thomas, "AI-bank of the future: Can banks meet the AI challenge?," If fully integrated, these capabilities can strengthen engagement significantly, supporting customers' financial activities across diverse online and physical contexts with intelligent, highly personalized solutions delivered through an interface that is intuitive, seamless, and fast.

One of our favorite gadgets from 2019 was the Google-powered Lenovo Smart Clock. It doesn't have all the bells and whistles of a typical Google smart display, but its alarm clock features, affordable price point and small form factor more than make up for it. Recently, however, the company debuted an even simpler version of the device, appropriately called the Lenovo Smart Clock Essential. With the Essential, the pretense of a smart display is gone altogether; the LCD screen has been replaced with a basic LED display. As a result, I don't quite like it as much as the original Lenovo Smart Clock, but it's also $30 cheaper (the Essential retails for $50 while the original Smart Clock is $80) and if all you really want is an alarm clock with some Google Assistant smarts, then the Essential certainly fits the bill. At its core, the Lenovo Smart Clock Essential is simply a Google-powered smart speaker with a built-in alarm clock.