BEIJING, BM, ($1 6.40 yuan) – The Chinese state media Global Times published a material in which it reveals that the pilots of the People's Liberation Army are pretending to use simulators with artificial intelligence. According to Liu Xuanzun, the process of training with artificial intelligence is two-way – the pilot learns to improve his skills, while artificial intelligence learns from the pilot's actions. There is already evidence that artificial intelligence developed in China is taking a serious advantage over pilots. In a recent "battle" between Chinese pilot Fang Guoyu, artificial intelligence won for the second time, according to the publication. China says the future of their fifth-generation J-20 fighter and next-generation fighters is based on the use of artificial intelligence.

Sanmay Das, Professor, Computer Science, is conducting an exploratory study in the use of techniques from artificial intelligence (AI) to improve early screening and the delivery of targeted assistance to households that are at risk of future homelessness and child maltreatment. Das and the other members of the research team seek to develop novel methods for allocation of scarce housing support to at-risk households, taking into account considerations of both overall efficiency and fairness. This work will necessitate novel problem formulation and algorithm development in AI as well as creating new ethical methods for deciding on how to effectively deliver social services while considering the vast complexity of human behavior. Das is collaborating with Patrick J. Fowler, Associate Professor at Washington University in St. Louis, on this project. The researchers will explore the feasibility of using novel algorithmic techniques to inform societal decision-making on the allocation of scarce resources, with the specific goal of improving service system outcomes for both homelessness and child welfare.

Artificial intelligence is also on the advance in IT security. According to a survey of 300 managers, 96 percent reported preparations in their companies for AI-supported IT attacks. In doing so, they partly rely on the help of "defensive AI". The survey was carried out with the assistance of the AI cybersecurity provider Darktrace. A survey of around 200 IT managers in medium-sized companies came to a more differentiated result.

Indeed, there has been a slew of research projects making a wide variety of predictions about automation-caused job loss, but those predictions differ by tens of millions of jobs, even when comparing similar time frames. This is irresponsible as new legislation might use any one of these predictions as a base for new laws, and ought to be using accurate calculations. In fact, most workers should not be in full panic territory just yet: automation will come in three distinct waves, and we're only riding the first one. Data analysis and theoretically simple digital tasks are already becoming obsolete thanks to the creation of "basic" A.Is trained through machine learning, but this is unlikely to go much further in the next couple of years.

India is slowly yet steadily taking steps ahead to establish artificial intelligence in several domains. Where, NASSCOM President Debjani Ghosh had mentioned that, in order to make India realize its full AI potential, a coalition of government, industries, and startups are required to come up with the necessary infrastructure and policy support. This will aid in enabling AI innovation across various sectors thereby mentoring and providing financial backing to the startups. In addition, according to a report by Accenture, artificial intelligence is anticipated to add on $957 billion or 15% of India's present GDP value by the year 2035. The amalgamation of technology, information along with the possibility and ability to create intelligent systems has reached critical mass thereby navigating towards exceptional growth in AI investment.

Brussels – NATO leaders warned Monday that China's military ambitions pose "systemic challenges" to their alliance, and agreed to enhance ties with Japan and other Asia-Pacific nations to back the rules-based international order. The tough line against Beijing, taken in a communique released after the NATO summit, came as U.S. President Joe Biden rallies allies to counter what he calls autocracies like China and Russia that are challenging an open international order. "China's stated ambitions and assertive behavior present systemic challenges to the rules-based international order and to areas relevant to alliance security," said the communique from the 30-member organization that brings together North American and European countries. The leaders also expressed concerns over what they called China's coercive policies, while pointing out the country's rapid expansion of its nuclear arsenal and criticizing the opaqueness of its military modernization. The communique, meanwhile, named Australia, Japan, New Zealand and South Korea as countries with which NATO plans to strengthen its "political dialogue and practical cooperation" in a bid to promote cooperative security and support the rules-based international order.

Large-scale pre-trained models (PTMs) such as BERT and GPT have recently achieved great success and become a milestone in the field of artificial intelligence (AI). Owing to sophisticated pre-training objectives and huge model parameters, large-scale PTMs can effectively capture knowledge from massive labeled and unlabeled data. By storing knowledge into huge parameters and fine-tuning on specific tasks, the rich knowledge implicitly encoded in huge parameters can benefit a variety of downstream tasks, which has been extensively demonstrated via experimental verification and empirical analysis. It is now the consensus of the AI community to adopt PTMs as backbone for downstream tasks rather than learning models from scratch. In this paper, we take a deep look into the history of pre-training, especially its special relation with transfer learning and self-supervised learning, to reveal the crucial position of PTMs in the AI development spectrum. Further, we comprehensively review the latest breakthroughs of PTMs. These breakthroughs are driven by the surge of computational power and the increasing availability of data, towards four important directions: designing effective architectures, utilizing rich contexts, improving computational efficiency, and conducting interpretation and theoretical analysis. Finally, we discuss a series of open problems and research directions of PTMs, and hope our view can inspire and advance the future study of PTMs.

Artificial Intelligence (AI) started out with an ambition to reproduce the human mind, but, as the sheer scale of that ambition became manifest, it quickly retreated into either studying specialized intelligent behaviours, or proposing over-arching architectural concepts for interfacing specialized intelligent behaviour components, conceived of as agents in a kind of organization. This agent-based modeling paradigm, in turn, proves to have interesting applications in understanding, simulating, and predicting the behaviour of social and legal structures on an aggregate level. For these reasons, this chapter examines a number of relevant cross-cutting concerns, conceptualizations, modeling problems and design challenges in large-scale distributed Artificial Intelligence, as well as in institutional systems, and identifies potential grounds for novel advances.

Analysis of short text, such as social media posts, is extremely difficult because it relies on observing many document-level word co-occurrence pairs. Beyond topic distributions, a common downstream task of the modeling is grouping the authors of these documents for subsequent analyses. Traditional models estimate the document groupings and identify user clusters with an independent procedure. We propose a novel model that expands on the Latent Dirichlet Allocation by modeling strong dependence among the words in the same document, with user-level topic distributions. We also simultaneously cluster users, removing the need for post-hoc cluster estimation and improving topic estimation by shrinking noisy user-level topic distributions towards typical values. Our method performs as well as -- or better -- than traditional approaches to problems arising in short text, and we demonstrate its usefulness on a dataset of tweets from United States Senators, recovering both meaningful topics and clusters that reflect partisan ideology.

Many applications of data-driven models demand transparency of decisions, especially in health care, criminal justice, and other high-stakes environments. Modern trends in machine learning research have led to algorithms that are increasingly intricate to the degree that they are considered to be black boxes. In an effort to reduce the opacity of decisions, methods have been proposed to construe the inner workings of such models in a human-comprehensible manner. These post hoc techniques are described as being universal explainers - capable of faithfully augmenting decisions with algorithmic insight. Unfortunately, there is little agreement about what constitutes a "good" explanation. Moreover, current methods of explanation evaluation are derived from either subjective or proxy means. In this work, we propose a framework for the evaluation of post hoc explainers on ground truth that is directly derived from the additive structure of a model. We demonstrate the efficacy of the framework in understanding explainers by evaluating popular explainers on thousands of synthetic and several real-world tasks. The framework unveils that explanations may be accurate but misattribute the importance of individual features.