Artificial intelligence technology has continued to grow in recent years, stunning the world with its latest innovations. But, some are admittedly growing weary about AI and its continuous growth. With talk of robots one day replacing humans for labor, concerns of an increasingly tech dependent world grow stronger. A report from Oxford researchers stated that 47% of American jobs will be at risk by 2030 because of automation. However, AI is truly changing the world - providing innovation that can change how we approach healthcare, the environment, and the day to day act of living.

Oftentimes, we are surprised by the accuracy of recommendations on what to buy on Amazon, watch on Netflix, or listen on Spotify. We feel that somehow these companies know how our brain works and monetizing this magical guessing game. They have a deep foundation on behavioral sciences, and our job is to make all these concepts real in a way that is both easy to understand and covers the most important concepts. Remember: People are extremely predictable. Personality builds our conduct and our conduct determines our decisions.

Text-based interactive recommendation provides richer user preferences and has demonstrated advantages over traditional interactive recommender systems. However, recommendations can easily violate preferences of users from their past natural-language feedback, since the recommender needs to explore new items for further improvement. To alleviate this issue, we propose a novel constraint-augmented reinforcement learning (RL) framework to efficiently incorporate user preferences over time. Specifically, we leverage a discriminator to detect recommendations violating user historical preference, which is incorporated into the standard RL objective of maximizing expected cumulative future rewards. Our proposed framework is general and is further extended to the task of constrained text generation.

Voice Assistants (VAs) such as Amazon Alexa or Google Assistant rely on wake-word detection to respond to people's commands, which could potentially be vulnerable to audio adversarial examples. In this work, we target our attack on the wake-word detection system. Our goal is to jam the model with some inconspicuous background music to deactivate the VAs while our audio adversary is present. We implemented an emulated wake-word detection system of Amazon Alexa based on recent publications. Then we computed our audio adversaries with consideration of expectation over transform and we implemented our audio adversary with a differentiable synthesizer.

In this paper, we model the dependencies among the items that are recommended to a user in a collaborative-filtering problem via a Gaussian Markov Random Field (MRF). We build upon Besag's auto-normal parameterization and pseudo-likelihood, which not only enables computationally efficient learning, but also connects the areas of MRFs and sparse inverse covariance estimation with autoencoders and neighborhood models, two successful approaches in collaborative filtering. We propose a novel approximation for learning sparse MRFs, where the trade-off between recommendation-accuracy and training-time can be controlled. At only a small fraction of the training-time compared to various baselines, including deep nonlinear models, the proposed approach achieved competitive ranking-accuracy on all three well-known data-sets used in our experiments, and notably a 20% gain in accuracy on the data-set with the largest number of items. Papers published at the Neural Information Processing Systems Conference.

Large-scale industrial recommender systems are usually confronted with computational problems due to the enormous corpus size. To retrieve and recommend the most relevant items to users under response time limits, resorting to an efficient index structure is an effective and practical solution. The previous work Tree-based Deep Model (TDM) \cite{zhu2018learning} greatly improves recommendation accuracy using tree index. By indexing items in a tree hierarchy and training a user-node preference prediction model satisfying a max-heap like property in the tree, TDM provides logarithmic computational complexity w.r.t. the corpus size, enabling the use of arbitrary advanced models in candidate retrieval and recommendation. In tree-based recommendation methods, the quality of both the tree index and the user-node preference prediction model determines the recommendation accuracy for the most part.

Being an early adopter of artificial intelligence and automation, Amazon always had an edge in using AI to improve its business efficiencies. Not only has it been using AI to enhance its customer experience but has been heavily focused internally. From using AI to predict the number of customers willing to buy a new product to running a cashier-less grocery store, Amazon's AI capabilities are designed to provide customised recommendations to its customers. According to a report, Amazon's recommendation engine is driving 35% of its total sales. One of the main areas where Amazon is applying continuous AI is to better understand their customer search queries and what is the reason they are looking for a particular product.

Artificial intelligence (AI) is one of the most exciting technological growth areas in recent years, with some investors like technologically-focused entrepreneur Tej Kohli predicting the sector will be worth $150 trillion (£125tn) by 2025, but why do we need the technology in our phones? Flagship devices today all come equipped with specialised AI processing chips, known and neural engines or neural processing units, from Apple's A12 Bionic CPU to Huawei's Kirin 980 or Qualcomm's Snapdragon 845, and more and more tasks are using their advanced processing capabilities. The most obvious artificial intelligence in our phones are the voice assistants that learn to understand our voice commands and then act appropriately from telling us the weather to playing our favourite song or adding an appointment to our calendar. Google, Apple, and Amazon have steered clear of labelling their services as AI so as not to scare away users fearful of a robot takeover, but these services rely on machine learning to function – understanding what you are telling them to do and then performing the right action. Possibly the most advanced implementation of any digital assistant is Google's Duplex service that will make calls and interact with other people and businesses on your behalf.

It means giving the ability to computer or machines to do and think their own task and ability to learn and improve them. It's is also possible AI helps us and takes care of us as we protect animals but Ai might be evil as we treat sometimes to some animals. AI is made with neural and complex coding which gives them the ability to learn and develop it too. Some commonly used languages to program ai are Python, lisp, java, R, etc. It means to give machine ability to learn, understand, think commonly giving them consciousness.

Contextual bandits are a common problem faced by machine learning practitioners in domains as diverse as hypothesis testing to product recommendations. There have been a lot of approaches in exploiting rich data representations for contextual bandit problems with varying degree of success. Self-supervised learning is a promising approach to find rich data representations without explicit labels. In a typical self-supervised learning scheme, the primary task is defined by the problem objective (e.g. clustering, classification, embedding generation etc.) and the secondary task is defined by the self-supervision objective (e.g. rotation prediction, words in neighborhood, colorization, etc.). In the usual self-supervision, we learn implicit labels from the training data for a secondary task. However, in the contextual bandit setting, we don't have the advantage of getting implicit labels due to lack of data in the initial phase of learning. We provide a novel approach to tackle this issue by combining a contextual bandit objective with a self supervision objective. By augmenting contextual bandit learning with self-supervision we get a better cumulative reward. Our results on eight popular computer vision datasets show substantial gains in cumulative reward. We provide cases where the proposed scheme doesn't perform optimally and give alternative methods for better learning in these cases.