Artificial intelligence has existed as a field for more than 50 years, but in pace with technological developments in recent years, the area has found increasing numbers of applications and has been the subject of increasing attention. New methods and technologies mean that the mobile phone not only understands what we say, but also translates between languages as quickly as we speak, recognizes faces. There are methods and technologies of artificial intelligence that lie at the core of self-driving cars and robots that perform precise surgical procedures. Facial recognition in stores, robotic sellers who submit offers based on past behaviors, facial recognition in stores, language assistants (like Alexa or Google Home) who are always listening and making recommendations based on recorded conversations. Al is a subject area that is changing how we live and work and how the future will be.

Today's small children, aka Generation Alpha, are the first to grow up with robots as peers. Those winsome talking devices spawned by a booming education-tech industry can speed children's learning, but they also can be confusing to them, research shows. Many children think robots are smarter than humans or imbue them with magical powers. The long-term consequences of growing up surrounded by AI-driven devices won't be clear for a while. But an expanding body of research is lending new impetus to efforts to expand technology education beyond learning to code, to understanding how AI works.

A lot of the conversations happening today in education are focused on how we can best prepare our students for the future. What are the types of experiences that our students should have? How can we create authentic, innovative learning opportunities for students that will help them to develop the skills they might need in the future? These are probably the two most common questions that I hear and that I ask myself quite often. Besides the content that I teach, what else can I embed into the curriculum that will best prepare my students?

Recently, a special type of data poisoning (DP) attack targeting Deep Neural Network (DNN) classifiers, known as a backdoor, was proposed. These attacks do not seek to degrade classification accuracy, but rather to have the classifier learn to classify to a target class whenever the backdoor pattern is present in a test example. Launching backdoor attacks does not require knowledge of the classifier or its training process - it only needs the ability to poison the training set with (a sufficient number of) exemplars containing a sufficiently strong backdoor pattern (labeled with the target class). Here we address post-training detection of backdoor attacks in DNN image classifiers, seldom considered in existing works, wherein the defender does not have access to the poisoned training set, but only to the trained classifier itself, as well as to clean examples from the classification domain. This is an important scenario because a trained classifier may be the basis of e.g. a phone app that will be shared with many users. Detecting backdoors post-training may thus reveal a widespread attack. We propose a purely unsupervised anomaly detection (AD) defense against imperceptible backdoor attacks that: i) detects whether the trained DNN has been backdoor-attacked; ii) infers the source and target classes involved in a detected attack; iii) we even demonstrate it is possible to accurately estimate the backdoor pattern. We test our AD approach, in comparison with alternative defenses, for several backdoor patterns, data sets, and attack settings and demonstrate its favorability. Our defense essentially requires setting a single hyperparameter (the detection threshold), which can e.g. be chosen to fix the system's false positive rate.

Such an agent is typically characterised by: (i) a finite set of states 6 S {s i} that describe all possible situations in the environment; (ii) a finite set of actions A {a j} to change in the environment from one situation to another; (iii) a state transition function T (s,a,s null) that specifies the next state s null for having taken action a in the current state s; (iv) a reward function R (s,a,s null) that specifies a numerical value given to the agent for taking action a in state s and transitioning to state s null; and (v) a policy π: S A that defines a mapping from states to actions [2, 30]. The goal of a reinforcement learning agent is to find an optimal policy by maximising its cumulative discounted reward defined as Q (s,a) max π E[r t γr t 1 γ 2 r t 1 ... s t s,a t a,π ], where function Q represents the maximum sum of rewards r t discounted by factor γ at each time step. While a reinforcement learning agent takes actions with probability Pr ( a s) during training, it selects the best action at test time according to π (s) arg max a A Q (s,a). A deep reinforcement learning agent approximates Q using a multi-layer neural network [31]. The Q function is parameterised as Q(s,a; θ), where θ are the parameters or weights of the neural network (recurrent neural network in our case). Estimating these weights requires a dataset of learning experiences D {e 1,...e N} (also referred to as'experience replay memory'), where every experience is described as a tuple e t ( s t,a t,r t,s t 1). Inducing a Q function consists in applying Q-learning updates over minibatches of experience MB {( s,a,r,s null) U (D)} drawn uniformly at random from the full dataset D . This process is implemented in learning algorithms using Deep Q-Networks (DQN) such as those described in [31, 32, 33], and the following section describes a DQN-based algorithm for human-chatbot interaction.

The deep supervised and reinforcement learning paradigms (among others) have the potential to endow interactive multimodal social robots with the ability of acquiring skills autonomously. But it is still not very clear yet how they can be best deployed in real world applications. As a step in this direction, we propose a deep learning-based approach for efficiently training a humanoid robot to play multimodal games---and use the game of `Noughts & Crosses' with two variants as a case study. Its minimum requirements for learning to perceive and interact are based on a few hundred example images, a few example multimodal dialogues and physical demonstrations of robot manipulation, and automatic simulations. In addition, we propose novel algorithms for robust visual game tracking and for competitive policy learning with high winning rates, which substantially outperform DQN-based baselines. While an automatic evaluation shows evidence that the proposed approach can be easily extended to new games with competitive robot behaviours, a human evaluation with 130 humans playing with the Pepper robot confirms that highly accurate visual perception is required for successful game play.

This story is part of a series on how we learn--from augmented reality to music-training devices. By now you must have heard the good news about our savior, artificial intelligence. It makes you look better in selfies, prevents blindness, and can even turn water into tastier beer. Tech giants and governments say we're living in a golden age of AI. Truth is, most times you hear the term artificial intelligence, the specific technology at work is called machine learning.

Not the Terminator kind that will take over the world. But definitely, the kind that are taking over our jobs! We live in an automated world, where our Uber rides are booked by Alexa and Siri gives us driving directions. PwC estimates that 30% of jobs will be impacted by automation by 2030. You and I may become irrelevant.

In this tutorial, you will learn how to build a scalable image hashing search engine using OpenCV, Python, and VP-Trees. Back in 2017, I wrote a tutorial on image hashing with OpenCV and Python (which is required reading for this tutorial). That guide showed you how to find identical/duplicate images in a given dataset. However, there was a scalability problem with that original tutorial -- namely that it did not scale! To find near-duplicate images, our original image hashing method would require us to perform a linear search, comparing the query hash to each individual image hash in our dataset. In a practical, real-world application that's far too slow -- we need to find a way to reduce that search to sub-linear time complexity. But how can we reduce search time so dramatically?

Sometimes a potential partner or client decides to wait for the next best concept (sometimes without giving you a chance). Sometimes they are simply phishing (have no intention of taking a next step). My team is trying to establish a standard representation of learning expectations and build a community of conscientious learners who apply skills properly, so our strategy depends on building awareness; usually do not hold back in communicating our vision. Started to hear a phrase'future proof' and thought how this applies to our skills applications. I have made the case for the relevancy of skills in education, higher education, and professional development in many previous posts; here are a few of them: Where Superintendents Stand on Key Issues, Workforce Strategy Focuses on Skills, Root Causes of Skills Gap, Forecast Hire and Train Based on Skills, and Why Skills Mapping.