Unrolled computation graphs arise in many scenarios, including training RNNs, tuning hyperparameters through unrolled optimization, and training learned optimizers. Current approaches to optimizing parameters in such computation graphs suffer from high variance gradients, bias, slow updates, or large memory usage. We introduce a method called Persistent Evolution Strategies (PES), which divides the computation graph into a series of truncated unrolls, and performs an evolution strategies-based update step after each unroll. PES eliminates bias from these truncations by accumulating correction terms over the entire sequence of unrolls. PES allows for rapid parameter updates, has low memory usage, is unbiased, and has reasonable variance characteristics. We experimentally demonstrate the advantages of PES compared to several other methods for gradient estimation on synthetic tasks, and show its applicability to training learned optimizers and tuning hyperparameters.

Artificial Intelligence in Video Games Market research is an intelligence report with meticulous efforts undertaken to study the right and valuable information. The data which has been looked upon is done considering both, the existing top players and the upcoming competitors. Business strategies of the key players and the new entering market industries are studied in detail. Well explained SWOT analysis, revenue share and contact information are shared in this report analysis. It also provides market information in terms of development and its capacities.

At the end of last month, a set of principles and advice on ethics in the use of Artificial Intelligence (AI) was known, adopted for the first time jointly and unanimously by the 193 member states of the General Council of the UNESCO. Beyond the uniqueness of its universal character, it is about Unesco launched a guide to improve the relationship between humans and robots and combines ethical issues to a warning voice that has been heard for a long time. There are already several international political organizations that have been warning about the need to provide an ethical component to what is undoubtedly the most notable advance in applied science of our time. In fact, in November but from '19 the European Union (EU) had published its Ethical Guidelines for a reliable artificial intelligence whose proposal revolves around the collateral effects, or unforeseen risks, that the implementation of disruptive technologies like this can generate. Likewise, in April of this year we learned about the European Commission regulation regarding the use of algorithms able to learn and make decisions.

Many online platforms today (such as Amazon, Netflix, Spotify, LinkedIn, and AirBnB) can be thought of as two-sided markets with producers and customers of goods and services. Traditionally, recommendation services in these platforms have focused on maximizing customer satisfaction by tailoring the results according to the personalized preferences of individual customers. However, our investigation reinforces the fact that such customer-centric design of these services may lead to unfair distribution of exposure to the producers, which may adversely impact their well-being. On the other hand, a pure producer-centric design might become unfair to the customers. As more and more people are depending on such platforms to earn a living, it is important to ensure fairness to both producers and customers. In this work, by mapping a fair personalized recommendation problem to a constrained version of the problem of fairly allocating indivisible goods, we propose to provide fairness guarantees for both sides. Formally, our proposed {\em FairRec} algorithm guarantees Maxi-Min Share ($\alpha$-MMS) of exposure for the producers, and Envy-Free up to One Item (EF1) fairness for the customers. Extensive evaluations over multiple real-world datasets show the effectiveness of {\em FairRec} in ensuring two-sided fairness while incurring a marginal loss in overall recommendation quality. Finally, we present a modification of FairRec (named as FairRecPlus) that at the cost of additional computation time, improves the recommendation performance for the customers, while maintaining the same fairness guarantees.

In recent years, there have been unprecedented technological advances in sensor technology, and sensors have become more affordable than ever. Thus, sensor-driven data collection is increasingly becoming an attractive and practical option for researchers around the globe. Such data is typically extracted in the form of time series data, which can be investigated with data mining techniques to summarize behaviors of a range of subjects including humans and animals. While enabling cheap and mass collection of data, continuous sensor data recording results in datasets which are big in size and volume, which are challenging to process and analyze with traditional techniques in a timely manner. Such collected sensor data is typically extracted in the form of time series data. There are two main approaches in the literature, namely, shape-based classification and feature-based classification. Shape-based classification determines the best class according to a distance measure. Feature-based classification, on the other hand, measures properties of the time series and finds the best class according to the set of features defined for the time series. In this dissertation, we demonstrate that neither of the two techniques will dominate for some problems, but that some combination of both might be the best. In other words, on a single problem, it might be possible that one of the techniques is better for one subset of the behaviors, and the other technique is better for another subset of behaviors. We introduce a hybrid algorithm to classify behaviors, using both shape and feature measures, in weakly labeled time series data collected from sensors to quantify specific behaviors performed by the subject. We demonstrate that our algorithm can robustly classify real, noisy, and complex datasets, based on a combination of shape and features, and tested our proposed algorithm on real-world datasets.

We propose an efficient interpretable neuro-symbolic model to solve Inductive Logic Programming (ILP) problems. In this model, which is built from a set of meta-rules organised in a hierarchical structure, first-order rules are invented by learning embeddings to match facts and body predicates of a meta-rule. To instantiate it, we specifically design an expressive set of generic meta-rules, and demonstrate they generate a consequent fragment of Horn clauses. During training, we inject a controlled \pw{Gumbel} noise to avoid local optima and employ interpretability-regularization term to further guide the convergence to interpretable rules. We empirically validate our model on various tasks (ILP, visual genome, reinforcement learning) against several state-of-the-art methods.

Is knowledge definable as justified true belief ("JTB")? We argue that one can legitimately answer positively or negatively, depending on whether or not one's true belief is justified by what we call adequate reasons. To facilitate our argument we introduce a simple propositional logic of reason-based belief, and give an axiomatic characterization of the notion of adequacy for reasons. We show that this logic is sufficiently flexible to accommodate various useful features, including quantification over reasons. We use our framework to contrast two notions of JTB: one internalist, the other externalist. We argue that Gettier cases essentially challenge the internalist notion but not the externalist one. Our approach commits us to a form of infallibilism about knowledge, but it also leaves us with a puzzle, namely whether knowledge involves the possession of only adequate reasons, or leaves room for some inadequate reasons. We favor the latter position, which reflects a milder and more realistic version of infallibilism.

Seeking a firm foothold in the United States, Turkish startups are revving up their investments in the world's most prominent market. Following in the footsteps of gaming ventures that have placed themselves on top among most downloaded games, Turkey's ultrafast grocery delivery company Getir just recently launched operations in the U.S. only a few months after expanding into Europe. On the other hand, Vispera, which offers tech solutions in the field of machine vision and machine learning for the fast consumption and retail sectors across 25 countries, has decided to develop and strengthen its subsidiary in the U.S., Vispera Corp. The company also plans to further develop the strong business partnerships and product developments it has established in 2021. Offering a range of image recognition tools that can solve common issues in retail spaces like stocking inventory levels and more effective displays of current promotions, Vispera has announced it would complete a new investment round in the first months of 2022. The Istanbul-based company seeks to boost its activities in the U.S. through its subsidiary Vispera Corp.

Bengaluru features in the top five cities on Harvard Business review where diversity is high in the AI sector. Diversity and inclusive pool of talent developing AI matters to the reviewers as AI developers are influenced by their own world views, which, in turn, guide them in their selection of applications, datasets, and training of algorithms. The data from the Fletcher school, Tufts university is derived and pitted against indicators such as talent pool, investments, diversity of talent, evolution of the country's digital foundations or TIDE. The reviewers believe that the factors collectively give companies a way to prioritise their AI talent sourcing choices by scoring the different locations on the concentration, quality and diversity of the AI talent pool. Top four cities are San Francisco, New York, Boston and Seattle respectively.

For many years, research in Artificial Intelligence (AI) has directed efforts towards automating processes to perform specific academic, industrial or economic tasks for society. However, the investigation and development of procedures for the automation of human artistic and creative processes has not had as much attention due to the complexities involved in these activities. Procedures developed for these purposes involve mathematical-computational methods designed to process and learn from a large quantity of digital data, so as to detect patterns in order to simulate the creative process (CP), as explained by Boden in [3]. In this paper, we introduce a model for the generation of rhymes with literary components. Our proposal is based on findings detailed in [11], where Automatic Text Generation (ATG) techniques are combined with neural network (NN) based models, such as the Word2vec algorithm [9], for the generation of literary texts.