We will use a test to determine what AI is. The test will produce a certain score, and we say that this is the program's IQ. Then we decide that all computer programs the IQ of which is above a certain level satisfy the AI definition. In order to explain this concept, let us make an analogy with the admission tests given to candidates who wish to become university students. The problems given at the test are selected randomly, but all candidate students receive the same problems. Withal, solving the problems should require logical thinking, because we aim to enroll students who think logically rather than the lucky ones that may hit the right answers haphazardly. The score is based on the number of problems solved by each candidate student. We cannot say how many problems should be solved, because we do not know how many candidates will show up at the test, nor do we know how well or unwell they will perform. We may say set a certain score (e.g.

In the field of machine learning, data understanding is the practice of getting initial insights in unknown datasets. Such knowledge-intensive tasks require a lot of documentation, which is necessary for data scientists to grasp the meaning of the data. Usually, documentation is separate from the data in various external documents, diagrams, spreadsheets and tools which causes considerable look up overhead. Moreover, other supporting applications are not able to consume and utilize such unstructured data. That is why we propose a methodology that uses a single semantic model that interlinks data with its documentation. Hence, data scientists are able to directly look up the connected information about the data by simply following links. Equally, they can browse the documentation which always refers to the data. Furthermore, the model can be used by other approaches providing additional support, like searching, comparing, integrating or visualizing data. To showcase our approach we also demonstrate an early prototype.

We introduce a new approach to hierarchy formation and task decomposition in hierarchical reinforcement learning. Our method is based on the Hierarchy Of Abstract Machines (HAM) framework because HAM approach is able to design efficient controllers that will realize specific behaviors in real robots. The key to our algorithm is the introduction of the internal or "mental" environment in which the state represents the structure of the HAM hierarchy. The internal action in this environment leads to changes the hierarchy of HAMs. We propose the classical Q-learning procedure in the internal environment which allows the agent to obtain an optimal hierarchy. We extends the HAM framework by adding on-model approach to select the appropriate sub-machine to execute action sequences for certain class of external environment states. Preliminary experiments demonstrated the prospects of the method.

Fairness in algorithmic decision-making processes is attracting increasing concern. When an algorithm is applied to human-related decision-making an estimator solely optimizing its predictive power can learn biases on the existing data, which motivates us the notion of fairness in machine learning. while several different notions are studied in the literature, little studies are done on how these notions affect the individuals. We demonstrate such a comparison between several policies induced by well-known fairness criteria, including the color-blind (CB), the demographic parity (DP), and the equalized odds (EO). We show that the EO is the only criterion among them that removes group-level disparity. Empirical studies on the social welfare and disparity of these policies are conducted.

Deep networks often perform well on the data manifold on which they are trained, yet give incorrect (and often very confident) answers when evaluated on points from off of the training distribution. This is exemplified by the adversarial examples phenomenon but can also be seen in terms of model generalization and domain shift. We propose Manifold Mixup which encourages the network to produce more reasonable and less confident predictions at points with combinations of attributes not seen in the training set. This is accomplished by training on convex combinations of the hidden state representations of data samples. Using this method, we demonstrate improved semi-supervised learning, learning with limited labeled data, and robustness to adversarial examples. Manifold Mixup requires no (significant) additional computation. Analytical experiments on both real data and synthetic data directly support our hypothesis for why the Manifold Mixup method improves results.

Uncertainty computation in deep learning is essential to design robust and reliable systems. Variational inference (VI) is a promising approach for such computation, but requires more effort to implement and execute compared to maximum-likelihood methods. In this paper, we propose new natural-gradient algorithms to reduce such efforts for Gaussian mean-field VI. Our algorithms can be implemented within the Adam optimizer by perturbing the network weights during gradient evaluations, and uncertainty estimates can be cheaply obtained by using the vector that adapts the learning rate. This requires lower memory, computation, and implementation effort than existing VI methods, while obtaining uncertainty estimates of comparable quality. Our empirical results confirm this and further suggest that the weight-perturbation in our algorithm could be useful for exploration in reinforcement learning and stochastic optimization.

Machine learning practitioners are often ambivalent about the ethical aspects of their products. We believe anything that gets us from that current state to one in which our systems are achieving some degree of fairness is an improvement that should be welcomed. This is true even when that progress does not get us 100% of the way to the goal of "complete" fairness or perfectly align with our personal belief on which measure of fairness is used. Some measure of fairness being built would still put us in a better position than the status quo. Impediments to getting fairness and ethical concerns applied in real applications, whether they are abstruse philosophical debates or technical overhead such as the introduction of ever more hyper-parameters, should be avoided. In this paper we further elaborate on our argument for this viewpoint and its importance.

Many data mining tasks cannot be completely addressed by automated processes, such as sentiment analysis and image classification. Crowdsourcing is an effective way to harness the human cognitive ability to process these machine-hard tasks. Thanks to public crowdsourcing platforms, e.g., Amazon Mechanical Turk and CrowdFlower, we can easily involve hundreds of thousands of ordinary workers (i.e., the crowd) to address these machine-hard tasks. In this tutorial, we will survey and synthesize a wide spectrum of existing studies on crowd-powered data mining. We first give an overview of crowdsourcing, and then summarize the fundamental techniques, including quality control, cost control, and latency control, which must be considered in crowdsourced data mining. Next we review crowd-powered data mining operations, including classification, clustering, pattern mining, outlier detection, knowledge base construction and enrichment. Finally, we provide the emerging challenges in crowdsourced data mining.

We consider networked control systems consisting of multiple independent controlled subsystems, operating over a shared communication network. Such systems are ubiquitous in cyber-physical systems, Internet of Things, and large-scale industrial systems. In many large-scale settings, the size of the communication network is smaller than the size of the system. In consequence, scheduling issues arise. The main contribution of this paper is to develop a deep reinforcement learning-based \emph{control-aware} scheduling (\textsc{DeepCAS}) algorithm to tackle these issues. We use the following (optimal) design strategy: First, we synthesize an optimal controller for each subsystem; next, we design a learning algorithm that adapts to the chosen subsystems (plants) and controllers. As a consequence of this adaptation, our algorithm finds a schedule that minimizes the \emph{control loss}. We present empirical results to show that \textsc{DeepCAS} finds schedules with better performance than periodic ones.

This paper describes an architecture for controlling non-player characters (NPC) in the First Person Shooter (FPS) game Unreal Tournament 2004. Specifically, the DRE-Bot architecture is made up of three reinforcement learners, Danger, Replenish and Explore, which use the tabular Sarsa({\lambda}) algorithm. This algorithm enables the NPC to learn through trial and error building up experience over time in an approach inspired by human learning. Experimentation is carried to measure the performance of DRE-Bot when competing against fixed strategy bots that ship with the game. The discount parameter, {\gamma}, and the trace parameter, {\lambda}, are also varied to see if their values have an effect on the performance.