Vector space models are to consider the relationship between data that are represented by vectors. It is popular in information retrieval systems but also useful for other purposes. Generally, this allows us to compare the similarity of two vectors from a geometric perspective. In this tutorial, we will see what is a vector space model and what it can do. A Gentle Introduction to Vector Space Models Photo by liamfletch, some rights reserved.

With the increasing adoption of artificial intelligence (AI) applications at the workplace, the debate about the future of work, workers, and the workplace has intensified. The polarised nature of debate ranges from job losses versus new-technology job creation through performance efficiency versus performance effectiveness to liberating humans from drudgery versus being controlled by machines. While several other polarities are evident in this debate, the truth always lies somewhere in between. In addition, there are other dark-side debates in the field about ethical, legal, and moral issues in the design and implementation of AI technologies for work and society. While popular discourse presents AI as a new phenomenon, the development of AI as an academic discipline dates back to 1956.

Rapid advances in artificial intelligence ("AI") are unlocking enhanced capabilities for machine learning, data interpretation and innovation, whilst also increasingly becoming useful in our everyday lives. AI now plays a key role in drug discovery, the advertisements we see recommended to us online, route suggestions for online mapping platforms, and auto-generated digital content. Recently, this has raised questions for traditional thinking around intellectual property law, with particular implications for patent ownership and invention. The question is – could AI be capable of being considered an inventor? An additional step, that an inventor must be human, was recently put to the test.

We focus on some shortcomings in current data protection regulation's ability to adequately address the ramifications of AI-driven data processing practices, in particular those of combining data sets. We propose that privacy regulation relies less on individuals' privacy expectations and recommend regulatory reform in two directions: (1) abolishing the distinction between personal and anonymized data for the purposes of triggering the application of data protection laws and (2) developing methods to prioritize regulatory intervention based on the level of privacy risk posed by individual data processing actions. This is an interdisciplinary paper that intends to build a bridge between the various communities involved in privacy research. We put special emphasis on linking technical notions with their regulatory implications and introducing the relevant technical and legal terminology in use to foster more efficient coordination between the policymaking and technical communities and enable a timely solution of the problems raised.

Disobeying the classical wisdom of statistical learning theory, modern deep neural networks generalize well even though they typically contain millions of parameters. Recently, it has been shown that the trajectories of iterative optimization algorithms can possess fractal structures, and their generalization error can be formally linked to the complexity of such fractals. This complexity is measured by the fractal's intrinsic dimension, a quantity usually much smaller than the number of parameters in the network. Even though this perspective provides an explanation for why overparametrized networks would not overfit, computing the intrinsic dimension (e.g., for monitoring generalization during training) is a notoriously difficult task, where existing methods typically fail even in moderate ambient dimensions. In this study, we consider this problem from the lens of topological data analysis (TDA) and develop a generic computational tool that is built on rigorous mathematical foundations. By making a novel connection between learning theory and TDA, we first illustrate that the generalization error can be equivalently bounded in terms of a notion called the 'persistent homology dimension' (PHD), where, compared with prior work, our approach does not require any additional geometrical or statistical assumptions on the training dynamics. Then, by utilizing recently established theoretical results and TDA tools, we develop an efficient algorithm to estimate PHD in the scale of modern deep neural networks and further provide visualization tools to help understand generalization in deep learning. Our experiments show that the proposed approach can efficiently compute a network's intrinsic dimension in a variety of settings, which is predictive of the generalization error.

--In the trial-and-error mechanism of reinforcement learning (RL), a notorious contradiction arises when we expect to learn a safe policy: how to learn a safe policy without enough data and prior model about the dangerous region? Existing methods mostly use the posterior penalty for dangerous actions, which means that the agent is not penalized until experiencing danger . This fact causes that the agent cannot learn a zero-violation policy even after convergence . Otherwise, it would not receive any penalty and lose the knowledge about danger . In this paper, we propose the safe set actor-critic (SSAC) algorithm, which confines the policy update using safety-oriented energy functions, or the safety indexes . The safety index is designed to increase rapidly for potentially dangerous actions, which allow us to locate the safe set on the action space, or the control safe set . Therefore, we can identify the dangerous actions prior to taking them, and further obtain a zero constraint-violation policy after convergence. We claim that we can learn the energy function in a model-free manner similar to learning a value function. By using the energy function transition as the constraint objective, we formulate a constrained RL problem. We prove that our Lagrangian-based solutions make sure that the learned policy will converge to the constrained optimum under some assumptions. The proposed algorithm is evaluated on both the complex simulation environments and a hardware-in-loop (HIL) experiment with a real controller from the autonomous vehicle. Experimental results suggest that the converged policy in all environments achieve zero constraint violation and comparable performance with model-based baseline. EINFORCEMENT learning has drawn rapidly growing attention for its superhuman learning capabilities in many sequential decision making problems like Go [1], Atari Games [2], and Starcraft [3].

Graph Neural Networks (GNNs) have become increasingly popular and achieved impressive results in many graph-based applications. However, extensive manual work and domain knowledge are required to design effective architectures, and the results of GNN models have high variance with different training setups, which limits the application of existing GNN models. In this paper, we present AutoHEnsGNN, a framework to build effective and robust models for graph tasks without any human intervention. AutoHEnsGNN won first place in the AutoGraph Challenge for KDD Cup 2020, and achieved the best rank score of five real-life datasets in the final phase. Given a task, AutoHEnsGNN first applies a fast proxy evaluation to automatically select a pool of promising GNN models. Then it builds a hierarchical ensemble framework: 1) We propose graph self-ensemble (GSE), which can reduce the variance of weight initialization and efficiently exploit the information of local and global neighborhoods; 2) Based on GSE, a weighted ensemble of different types of GNN models is used to effectively learn more discriminative node representations. To efficiently search the architectures and ensemble weights, we propose AutoHEnsGNN$_{\text{Gradient}}$, which treats the architectures and ensemble weights as architecture parameters and uses gradient-based architecture search to obtain optimal configurations, and AutoHEnsGNN$_{\text{Adaptive}}$, which can adaptively adjust the ensemble weight based on the model accuracy. Extensive experiments on node classification, graph classification, edge prediction and KDD Cup challenge demonstrate the effectiveness and generality of AutoHEnsGNN

On November 19, Elisabeth Borne, Minister of Labour, Employment and Integration, visited the Matrice innovation institute to sign an agreement with Bruno Sportisse of Inria to create a laboratory dedicated to artificial intelligence. Called LaborIA and operated by Matrice, this resource and experimentation centre will have the mission of "better understanding artificial intelligence and its effects on work, employment, skills and social dialogue in order to develop business practices and public action". According to the OECD's 2019 Employment Outlook report, medium-skilled jobs are increasingly exposed to profound transformations. Over the next 15 to 20 years, the development of automation could lead to the disappearance of 14% of current jobs, and another 32% are likely to be profoundly transformed. The report states that the future of work is in our hands and will depend, to a large extent, on the public policy choices countries make.

Artificial intelligence (AI) is a force multiplier that has the potential to deliver faster, smarter and safer military effects with less resources. It is an essential technology that is at the heart of advances in decision support, situational awareness, logistics, robotic process automation, natural language processing and digital twin modelling. Massive investments have already been made by both allies and competitors seeking to lead or gain advantage through application of AI. To highlight this, a recent report from the US National Security Commission on AI recommends that, "by 2025, the (US) Department of Defense and Intelligence Community must be AI-ready". What does an "AI-ready" organisation look like, and what does Air Force need to do to realise the benefits?

Apartment building defects are not uncommon these days, but the NSW government has been developing new solutions using AI and blockchain to crackdown on this. Speaking at the 2021 digital.NSW event, Office of the NSW Building Commissioner digital director Yin Man explained how the state government has worked with KPMG, Microsoft, Australian Securities Exchange (ASX), Western Sydney University, and Mirvac to build what is being referred to as a trustworthy index, within the state government's building assurance solution. The solution, based on ASX's blockchain technology, has been designed to track a building's provenance -- from the materials that are used, the drawings of the building, and people involved in the construction -- to enable the building industry, current and prospective owners, regulators, and insurers to compare and assess the trustworthiness of different buildings. "You, as a consumer, can now see one building differentiated from another and that helps the insurance companies and the financiers as well, because at the moment, they do not want to be in the market because all the buildings look the same to them, everybody has an occupation certificate, but why are some buildings still defective as we find in our audits, and some are not," Man said. According to Man, the trustworthy index will be piloted for the next six months with a brand new Mirvac building, along with over 200 buildings where combustible cladding is being replaced.