Pretrained language models (LMs) do not capture factual knowledge very well. This has led to the development of a number of knowledge integration (KI) methods which aim to incorporate external knowledge into pretrained LMs. Even though KI methods show some performance gains over vanilla LMs, the inner-workings of these methods are not well-understood. For instance, it is unclear how and what kind of knowledge is effectively integrated into these models and if such integration may lead to catastrophic forgetting of already learned knowledge. This paper revisits the KI process in these models with an information-theoretic view and shows that KI can be interpreted using a graph convolution operation. We propose a probe model called \textit{Graph Convolution Simulator} (GCS) for interpreting knowledge-enhanced LMs and exposing what kind of knowledge is integrated into these models. We conduct experiments to verify that our GCS can indeed be used to correctly interpret the KI process, and we use it to analyze two well-known knowledge-enhanced LMs: ERNIE and K-Adapter, and find that only a small amount of factual knowledge is integrated in them. We stratify knowledge in terms of various relation types and find that ERNIE and K-Adapter integrate different kinds of knowledge to different extent. Our analysis also shows that simply increasing the size of the KI corpus may not lead to better KI; fundamental advances may be needed.

In an international collaboration, University of Queensland palaeontologist Dr Anthony Romilio used AI pattern recognition to re-analyse footprints from the Dinosaur Stampede National Monument, south-west of Winton in Central Queensland. "Large dinosaur footprints were first discovered back in the 1970s at a track site called the Dinosaur Stampede National Monument, and for many years they were believed to be left by a predatory dinosaur, like Australovenator, with legs nearly two metres long," said Dr Romilio. "The mysterious tracks were thought to be left during the mid-Cretaceous Period, around 93 million years ago. "But working out what dino species made the footprints exactly -- especially from tens of millions of years ago -- can be a pretty difficult and confusing business. "Particularly since these big tracks are surrounded by thousands of tiny dinosaur footprints, leading many to think that this predatory beast could have sparked a stampede of smaller dinosaurs. "So, to crack the case, we decided to employ an AI program called Deep Convolutional Neural Networks." It was trained with 1,500 dinosaur footprints, all of which were theropod or ornithopod in origin -- the groups of dinosaurs relevant to the Dinosaur Stampede National Monument prints. The results were clear: the tracks had been made by a herbivorous ornithopod dinosaur. Dr Jens Lallensack, lead author from Liverpool John Moores University in the UK, said that the computer assistance was vital, as the team was originally at an impasse. "We were pretty stuck, so thank god for modern technology," Dr Lallensack said. "In our research team of three, one person was pro-meat-eater, one person was undecided, and one was pro-plant-eater.

It is our pleasure to invite you to submit a chapter for inclusion in the “Adversarial Multimedia Forensics” book to be Published by Springer – Advances in Information Security. The submitted chapter should have 15-20 pages of single-space single-column in latex and include sufficient details to be useful for Cybersecurity Applications experts and readers with […]

AI has enormous promise for improving and enriching our lives. However, serious concerns exist about its use, intrusion, and abuse. The Google AI arm revealed a variety of artificial intelligence projects it was working on, including one focused on preventing blindness. At its annual developer conference, Google unveiled 12 new AI project categories, some of which could lead to improved healthcare, others that could be used for creative purposes, and others that might be fun to play with. Google's new wildfire tracking feature is now available in the United States, Canada, Mexico, and some parts of Australia.

Google Play has been committed to giving users safe and convenient ways to pay for their favorite apps and games – while helping developers of all sizes transact with millions of users globally and build successful businesses on the platform. In line with this commitment, Google announced that it is introducing UPI Autopay as a payment option for subscription-based purchases on Google Play in India. Introduced under UPI 2.0 by NPCI, UPI Autopay helps customers make recurring payments using any UPI application that supports the feature. Saurabh Agarwal, Head of Google Play Retail & Payments Activation – India, Vietnam, Australia & New Zealand said, "We are always looking at adding popular and effective forms of payment around the world to ensure people can pay for apps and in-app content conveniently. With the introduction of UPI Autopay on the platform, we aim to extend the convenience of UPI to subscription-based purchases, helping many more people access helpful and delightful services – while enabling local developers to grow their subscription-based businesses on Google Play."

Artificial intelligence has revealed that prehistoric footprints thought to be made by a vicious dinosaur predator were in fact from a timid herbivore. In an international collaboration, University of Queensland paleontologist Dr. Anthony Romilio used AI pattern recognition to re-analyze footprints from the Dinosaur Stampede National Monument, south-west of Winton in Central Queensland. "Large dinosaur footprints were first discovered back in the 1970s at a track site called the Dinosaur Stampede National Monument, and for many years they were believed to be left by a predatory dinosaur, like Australovenator, with legs nearly two meters long," said Dr. Romilio. "The mysterious tracks were thought to be left during the mid-Cretaceous Period, around 93 million years ago. "But working out what dino species made the footprints exactly--especially from tens of millions of years ago--can be a pretty difficult and confusing business.

Distinguishing between tridactyl (three-toed) dinosaur tracks of the herbivorous ornithischians and the predominantly carnivorous theropods is a complex and long-standing problem [1–9]. Broadly, ornithischian tracks are expected to be wider and more symmetric than theropod tracks, with digit impression III less projecting beyond digit impressions II and IV, and with digit impressions being broader, more splayed apart, and terminating in blunt hoof marks instead of sharp claw marks. However, any of these characteristics can be found in both groups, and which are the most important depends on the particular track type in question. Moratalla et al. [1] presented a quantitative approach to discriminate these groups, albeit limited to larger theropod and ornithopod tracks. Limitations of this approach include the small sample size, issues with the measurement scheme and omission of relevant shape characteristics [1,5,6]; despite this, the method has found wide application [3,9–12]. To overcome the limitations of previous statistical approaches, and to remove as much subjectivity as possible, we trained and then employed an artificial neural network to categorize outlines of tridactyl dinosaur footprints as theropod or ornithischian. Artificial neural networks are a type of nonlinear model that can learn from data, and a principal component of machine learning and artificial intelligence. Inspired by the structure of the human brain, such neural networks comprise interconnected nodes (or neurons), with each connection represented by a number (weight).

Deep learning has achieved great success in learning features from massive remote sensing images (RSIs). To better understand the connection between feature learning paradigms (e.g., unsupervised feature learning (USFL), supervised feature learning (SFL), and self-supervised feature learning (SSFL)), this paper analyzes and compares them from the perspective of feature learning signals, and gives a unified feature learning framework. Under this unified framework, we analyze the advantages of SSFL over the other two learning paradigms in RSIs understanding tasks and give a comprehensive review of the existing SSFL work in RS, including the pre-training dataset, self-supervised feature learning signals, and the evaluation methods. We further analyze the effect of SSFL signals and pre-training data on the learned features to provide insights for improving the RSI feature learning. Finally, we briefly discuss some open problems and possible research directions.

Clinical Cohort Studies (CCS), such as randomized clinical trials, are a great source of documented clinical research. Ideally, a clinical expert inspects these articles for exploratory analysis ranging from drug discovery for evaluating the efficacy of existing drugs in tackling emerging diseases to the first test of newly developed drugs. However, more than 100 articles are published daily on a single prevalent disease like COVID-19 in PubMed. As a result, it can take days for a physician to find articles and extract relevant information. Can we develop a system to sift through the long list of these articles faster and document the crucial takeaways from each of these articles? In this work, we propose CCS Explorer, an end-to-end system for relevance prediction of sentences, extractive summarization, and patient, outcome, and intervention entity detection from CCS. CCS Explorer is packaged in a web-based graphical user interface where the user can provide any disease name. CCS Explorer then extracts and aggregates all relevant information from articles on PubMed based on the results of an automatically generated query produced on the back-end. For each task, CCS Explorer fine-tunes pre-trained language representation models based on transformers with additional layers. The models are evaluated using two publicly available datasets. CCS Explorer obtains a recall of 80.2%, AUC-ROC of 0.843, and an accuracy of 88.3% on sentence relevance prediction using BioBERT and achieves an average Micro F1-Score of 77.8% on Patient, Intervention, Outcome detection (PIO) using PubMedBERT. Thus, CCS Explorer can reliably extract relevant information to summarize articles, saving time by $\sim \text{660}\times$.

Quantification, variously called "supervised prevalence estimation" or "learning to quantify", is the supervised learning task of generating predictors of the relative frequencies (a.k.a. "prevalence values") of the classes of interest in unlabelled data samples. While many quantification methods have been proposed in the past for binary problems and, to a lesser extent, single-label multiclass problems, the multi-label setting (i.e., the scenario in which the classes of interest are not mutually exclusive) remains by and large unexplored. A straightforward solution to the multi-label quantification problem could simply consist of recasting the problem as a set of independent binary quantification problems. Such a solution is simple but na\"ive, since the independence assumption upon which it rests is, in most cases, not satisfied. In these cases, knowing the relative frequency of one class could be of help in determining the prevalence of other related classes. We propose the first truly multi-label quantification methods, i.e., methods for inferring estimators of class prevalence values that strive to leverage the stochastic dependencies among the classes of interest in order to predict their relative frequencies more accurately. We show empirical evidence that natively multi-label solutions outperform the na\"ive approaches by a large margin. The code to reproduce all our experiments is available online.