We know we are not anonymous online. Our every move in the digital sphere is tracked, collected, analyzed. It's all fascinating to our spies, who know our identity at every step. They can pinpoint us by the way we write our emails, use the mouse on our computer screens--even how we hold and swipe our cellphones. Soon we may not be anonymous in public either.

Machine-learning adoption in Australia hasn't been as fast as expected, experts say, with the coronavirus pandemic being one of the main reasons as it has both affected the data used for modelling but also put projects on hold across Australian enterprises. The intentions were high, as Gartner's 2021 CIO survey for Australia and New Zealand found that 9% of CIOs had already deployed artificial intelligence or machine learning, and 33% planned to within 12 months. Machine learning uses data and algorithms to attempt to imitate the way humans learn. Organisations build models that uses historical data to look for patterns, changes, or particular events. Gartner's applied analytics, data, and governance analyst, Ian Bertram, says that with COVID-19 all that data changes, and models created on existing data now perform poorly, so the push to operationalise and productionise machine learning needs to be reconsidered.

South Africa recently became what is believed to be the first country in the world to award a patent to an invention by an Artificial Intelligence (AI). An interlocking food and beverage container based on fractal geometry has been awarded a patent by South Africa's Companies and Intellectual Property Commission (CIPC). Confirmation of the patent was published in the commission's journal on 28 July. But unlike the hundreds of patents listed in the CIPC's latest journal, this container was not conceptualised by a human. The patent identifies Dabus – the Device for the Autonomous Bootstrapping of Unified Sentience – as the inventor.

Hate speech is a specific type of controversial content that is widely legislated as a crime that must be identified and blocked. However, due to the sheer volume and velocity of the Twitter data stream, hate speech detection cannot be performed manually. To address this issue, several studies have been conducted for hate speech detection in European languages, whereas little attention has been paid to low-resource South Asian languages, making the social media vulnerable for millions of users. In particular, to the best of our knowledge, no study has been conducted for hate speech detection in Roman Urdu text, which is widely used in the sub-continent. In this study, we have scrapped more than 90,000 tweets and manually parsed them to identify 5,000 Roman Urdu tweets. Subsequently, we have employed an iterative approach to develop guidelines and used them for generating the Hate Speech Roman Urdu 2020 corpus. The tweets in the this corpus are classified at three levels: Neutral-Hostile, Simple-Complex, and Offensive-Hate speech. As another contribution, we have used five supervised learning techniques, including a deep learning technique, to evaluate and compare their effectiveness for hate speech detection. The results show that Logistic Regression outperformed all other techniques, including deep learning techniques for the two levels of classification, by achieved an F1 score of 0.906 for distinguishing between Neutral-Hostile tweets, and 0.756 for distinguishing between Offensive-Hate speech tweets.

The study presents a new method for analyzing medical data based on the LogNNet neural network, which uses chaotic mappings to transform input information. The technique calculates risk factors for the presence of a disease in a patient according to a set of medical health indicators. The LogNNet architecture allows the implementation of artificial intelligence on medical pe-ripherals of the Internet of Things with low RAM resources, and the development of edge computing in healthcare. The efficiency of LogNNet in assessing perinatal risk is illustrated on cardiotocogram data of 2126 pregnant women, obtained from the UC Irvine machine learning repository. The classification accuracy reaches ~ 91%, with the ~ 3-10 kB of RAM used on the Arduino microcontroller. In addition, examples for diagnosing COVID-19 are provided, using LogNNet trained on a publicly available database from the Israeli Ministry of Health. The service concept has been developed, which uses the data of the express test for COVID-19 and reaches the classification accuracy of ~ 95% with the ~ 0.6 kB of RAM used on Arduino microcontrollers. In all examples, the model is tested using standard classification quality metrics: Precision, Recall, and F1-measure. The study results can be used in clinical decision support systems.

This paper introduces WeChat AI's participation in WMT 2021 shared news translation task on English->Chinese, English->Japanese, Japanese->English and English->German. Our systems are based on the Transformer (Vaswani et al., 2017) with several novel and effective variants. In our experiments, we employ data filtering, large-scale synthetic data generation (i.e., back-translation, knowledge distillation, forward-translation, iterative in-domain knowledge transfer), advanced finetuning approaches, and boosted Self-BLEU based model ensemble. Our constrained systems achieve 36.9, 46.9, 27.8 and 31.3 case-sensitive BLEU scores on English->Chinese, English->Japanese, Japanese->English and English->German, respectively. The BLEU scores of English->Chinese, English->Japanese and Japanese->English are the highest among all submissions, and that of English->German is the highest among all constrained submissions.

Science is a catalyst for human progress. But a publishing monopoly and funding monopsony have inhibited research. We intend to improve incentives in science by developing smart research contracts. These will collectively reward scientific activities, including proposals, papers, replications, datasets, analyses, annotations, editorials, and more. Peer-to-peer review networks will be designed to help evaluate proposals and publications. Long term, these smart contracts help accelerate research by minimizing science friction, ensuring science quality, and maximizing science variance. Email bits@atoms.org or follow @atoms_org to help us build a flourishing research economy. Papers are the fundamental asset of the research economy: they serve as proof of work that valuable research has been completed. Funding agencies and research institutions evaluate scientists based on their publications. Principal investigators (PIs) attract prospective students and collaborators via papers. Investors and companies use scientific literature to conduct due diligence on research ranging from basic discoveries to clinical studies. Thus, the evaluation and dissemination of papers are vital to this research economy. Publishers are the sole arbiters of papers today. They assign a value -- denominated in "prestige" -- by accepting a paper into the appropriate journal based on selectivity and domain. To evaluate papers, journals typically outsource it to two or three PIs, who often outsource it further to their students. Reviewers are unpaid for this peer review work, as it is an expected part of their scientific duties. Peer review is believed to be necessary because of the industrialization of science. Research papers and proposals have become too specialized and too numerous, making it difficult to assess merit prima facie. As a result, scientific incentives have become distorted in two major ways: prestige capture and reviewer misalignment. Over half of all research papers in 2013 were published by five companies, who have used their centuries of brand equity to build an economic moat. This results in prestige capture, which akin to regulatory capture, causes public and scientific interest to be directed towards the regulators of prestige.

Account-Based Marketing (ABM) is a strategy that's been around for over 100 years, yet only lately has gained the interest of business marketers. In this article, you'll find everything you need to know about Account-Based Marketing and how it can work for your business. So without any further ado, let's get straight into it. For those not familiar with the terms, Account-Based Marketing (ABM) is an approach to marketing that focuses on identifying and prioritizing accounts or contacts within a target market. In most cases, ABM will also involve creating personalized value propositions for each account or contact, building individualized campaigns based around those unique value propositions, and then delivering customized content to each account in order to generate the best possible ROI.

Aspect-level sentiment classification (ASC) aims to predict the fine-grained sentiment polarity towards a given aspect mentioned in a review. Despite recent advances in ASC, enabling machines to preciously infer aspect sentiments is still challenging. This paper tackles two challenges in ASC: (1) due to lack of aspect knowledge, aspect representation derived in prior works is inadequate to represent aspect's exact meaning and property information; (2) prior works only capture either local syntactic information or global relational information, thus missing either one of them leads to insufficient syntactic information. To tackle these challenges, we propose a novel ASC model which not only end-to-end embeds and leverages aspect knowledge but also marries the two kinds of syntactic information and lets them compensate for each other. Our model includes three key components: (1) a knowledge-aware gated recurrent memory network recurrently integrates dynamically summarized aspect knowledge; (2) a dual syntax graph network combines both kinds of syntactic information to comprehensively capture sufficient syntactic information; (3) a knowledge integrating gate re-enhances the final representation with further needed aspect knowledge; (4) an aspect-to-context attention mechanism aggregates the aspect-related semantics from all hidden states into the final representation. Experimental results on several benchmark datasets demonstrate the effectiveness of our model, which overpass previous state-of-the-art models by large margins in terms of both Accuracy and Macro-F1.

To enable the reusability of massive scientific datasets by humans and machines, researchers aim to create scientific datasets that adhere to the principles of findability, accessibility, interoperability, and reusability (FAIR) for data and artificial intelligence (AI) models. This article provides a domain-agnostic, step-by-step assessment guide to evaluate whether or not a given dataset meets each FAIR principle. We then demonstrate how to use this guide to evaluate the FAIRness of an open simulated dataset produced by the CMS Collaboration at the CERN Large Hadron Collider. This dataset consists of Higgs boson decays and quark and gluon background, and is available through the CERN Open Data Portal. We also use other available tools to assess the FAIRness of this dataset, and incorporate feedback from members of the FAIR community to validate our results. This article is accompanied by a Jupyter notebook to facilitate an understanding and exploration of the dataset, including visualization of its elements. This study marks the first in a planned series of articles that will guide scientists in the creation and quantification of FAIRness in high energy particle physics datasets and AI models.