Standard automatic metrics (such as BLEU) are problematic for document-level MT evaluation. They can neither distinguish document-level improvements in translation quality from sentence-level ones nor can they identify the specific discourse phenomena that caused the translation errors. To address these problems, we propose an automatic metric BlonD for document-level machine translation evaluation. BlonD takes discourse coherence into consideration by calculating the recall and distance of check-pointing phrases and tags, and further provides comprehensive evaluation scores by combining with n-gram. Extensive comparisons between BlonD and existing evaluation metrics are conducted to illustrate their critical distinctions. Experimental results show that BlonD has a much higher document-level sensitivity with respect to previous metrics. The human evaluation also reveals high Pearson R correlation values between BlonD scores and manual quality judgments.

We take a step towards addressing the under-representation of the African continent in NLP research by creating the first large publicly available high-quality dataset for named entity recognition (NER) in ten African languages, bringing together a variety of stakeholders. We detail characteristics of the languages to help researchers understand the challenges that these languages pose for NER. We analyze our datasets and conduct an extensive empirical evaluation of state-of-the-art methods across both supervised and transfer learning settings. We release the data, code, and models in order to inspire future research on African NLP.

We introduce a grey-box adversarial attack and defence framework for sentiment classification. We address the issues of differentiability, label preservation and input reconstruction for adversarial attack and defence in one unified framework. Our results show that once trained, the attacking model is capable of generating high-quality adversarial examples substantially faster (one order of magnitude less in time) than state-of-the-art attacking methods. These examples also preserve the original sentiment according to human evaluation. Additionally, our framework produces an improved classifier that is robust in defending against multiple adversarial attacking methods. Code is available at: https://github.com/ibm-aur-nlp/adv-def-text-dist.

The increasing adoption of automation, artificial intelligence (AI), and other technologies suggests that the role of humans in the economy will shrink drastically, wiping out millions of jobs in the process. COVID-19 accelerated this effect in 2020 and will likely boost digitization, and perhaps establish it permanently, in some areas. However, the real picture is more nuanced: though these technologies will eliminate some jobs, they will create many others. Governments, companies, and individuals all need to understand these shifts when they plan for the future. BCG recently collaborated with Faethm, a firm specializing in AI and analytics, to study the potential impact of various technologies on jobs in three countries: the US, Germany, and Australia.

Today, however, another revolution is unfolding that has potentially further reaching ramifications. According to experts, artificial intelligence is going to significantly change and alter the way humans manufacture, produce and deliver. In other words, it will change the way we work, live and connect with one another. Moreover, the scale of this change will be unlike anything we have experienced before. AI entails all attempts to make machines and devices think just like humans do.

Looking at the America's Cup boats, you can't help but notice how futuristic they appear. But this year, Emirates Team New Zealand had a competitive edge, when they enlisted a new crew member to hyper-boost their boat design process: AI. The 2021 race officially began on March 10, when boats hit the water, but the real competition started three years earlier, when the new Class Rule was released. With each America's Cup, the Class Rule defines the parameters for yacht design -- and design can mean the difference between winning and losing. Once the rule is released, it becomes a race to design and build the fastest possible America's Cup boats.

In this paper, we propose a novel lightweight relation extraction approach of structural block driven - convolutional neural learning. Specifically, we detect the essential sequential tokens associated with entities through dependency analysis, named as a structural block, and only encode the block on a block-wise and an inter-block-wise representation, utilizing multi-scale CNNs. This is to 1) eliminate the noisy from irrelevant part of a sentence; meanwhile 2) enhance the relevant block representation with both block-wise and inter-block-wise semantically enriched representation. Our method has the advantage of being independent of long sentence context since we only encode the sequential tokens within a block boundary. Experiments on two datasets i.e., SemEval2010 and KBP37, demonstrate the significant advantages of our method. In particular, we achieve the new state-of-the-art performance on the KBP37 dataset; and comparable performance with the state-of-the-art on the SemEval2010 dataset.

We focus on the robustness of neural networks for classification. To permit a fair comparison between methods to achieve robustness, we first introduce a standard based on the mensuration of a classifier's degradation. Then, we propose natural perturbed training to robustify the network. Natural perturbations will be encountered in practice: the difference of two images of the same object may be approximated by an elastic deformation (when they have slightly different viewing angles), by occlusions (when they hide differently behind objects), or by saturation, Gaussian noise etc. Training some fraction of the epochs on random versions of such variations will help the classifier to learn better. We conduct extensive experiments on six datasets of varying sizes and granularity. Natural perturbed learning show better and much faster performance than adversarial training on clean, adversarial as well as natural perturbed images. It even improves general robustness on perturbations not seen during the training. For Cifar-10 and STL-10 natural perturbed training even improves the accuracy for clean data and reaches the state of the art performance. Ablation studies verify the effectiveness of natural perturbed training.

Most of existing outlier detection methods assume that the outlier factors (i.e., outlierness scoring measures) of data entities (e.g., feature values and data objects) are Independent and Identically Distributed (IID). This assumption does not hold in real-world applications where the outlierness of different entities is dependent on each other and/or taken from different probability distributions (non-IID). This may lead to the failure of detecting important outliers that are too subtle to be identified without considering the non-IID nature. The issue is even intensified in more challenging contexts, e.g., high-dimensional data with many noisy features. This work introduces a novel outlier detection framework and its two instances to identify outliers in categorical data by capturing non-IID outlier factors. Our approach first defines and incorporates distribution-sensitive outlier factors and their interdependence into a value-value graph-based representation. It then models an outlierness propagation process in the value graph to learn the outlierness of feature values. The learned value outlierness allows for either direct outlier detection or outlying feature selection. The graph representation and mining approach is employed here to well capture the rich non-IID characteristics. Our empirical results on 15 real-world data sets with different levels of data complexities show that (i) the proposed outlier detection methods significantly outperform five state-of-the-art methods at the 95%/99% confidence level, achieving 10%-28% AUC improvement on the 10 most complex data sets; and (ii) the proposed feature selection methods significantly outperform three competing methods in enabling subsequent outlier detection of two different existing detectors.

This paper addresses the interesting problem of processing and analyzing data in geographic information systems (GIS) to achieve a clear perspective on urban sprawl. The term urban sprawl refers to overgrowth and expansion of low-density areas with issues such as car dependency and segregation between residential versus commercial use. Sprawl has impacts on the environment and public health. In our work, spatiotemporal features related to real GIS data on urban sprawl such as population growth and demographics are mined to discover knowledge for decision support. We adapt data mining algorithms, Apriori for association rule mining and J4.8 for decision tree classification to geospatial analysis, deploying the ArcGIS tool for mapping. Knowledge discovered by mining this spatiotemporal data is used to implement a prototype spatial decision support system (SDSS). This SDSS predicts whether urban sprawl is likely to occur. Further, it estimates the values of pertinent variables to understand how the variables impact each other. The SDSS can help decision-makers identify problems and create solutions for avoiding future sprawl occurrence and conducting urban planning where sprawl already occurs, thus aiding sustainable development. This work falls in the broad realm of geospatial intelligence and sets the stage for designing a large scale SDSS to process big data in complex environments, which constitutes part of our future work.