Unsupervised domain adaptation seeks to mitigate the distribution discrepancy between source and target domains, given labeled samples of the source domain and unlabeled samples of the target domain. Generative adversarial networks (GANs) have demonstrated significant improvement in domain adaptation by producing images which are domain specific for training. However, most of the existing GAN based techniques for unsupervised domain adaptation do not consider semantic information during domain matching, hence these methods degrade the performance when the source and target domain data are semantically different. In this paper, we propose an end-to-end novel semantic consistent generative adversarial network (SCGAN). This network can achieve source to target domain matching by capturing semantic information at the feature level and producing images for unsupervised domain adaptation from both the source and the target domains. We demonstrate the robustness of our proposed method which exceeds the state-of-the-art performance in unsupervised domain adaptation settings by performing experiments on digit and object classification tasks.

Graph neural networks (GNN) have been ubiquitous in graph learning tasks such as node classification. Most of GNN methods update the node embedding iteratively by aggregating its neighbors' information. However, they often suffer from negative disturbance, due to edges connecting nodes with different labels. One approach to alleviate this negative disturbance is to use attention, but current attention always considers feature similarity and suffers from the lack of supervision. In this paper, we consider the label dependency of graph nodes and propose a decoupling attention mechanism to learn both hard and soft attention. The hard attention is learned on labels for a refined graph structure with fewer inter-class edges. Its purpose is to reduce the aggregation's negative disturbance. The soft attention is learned on features maximizing the information gain by message passing over better graph structures. Moreover, the learned attention guides the label propagation and the feature propagation. Extensive experiments are performed on five well-known benchmark graph datasets to verify the effectiveness of the proposed method.

In many urban areas, traffic load and noise pollution are constantly increasing. Automated systems for traffic monitoring are promising countermeasures, which allow to systematically quantify and predict local traffic flow in order to to support municipal traffic planning decisions. In this paper, we present a novel open benchmark dataset, containing 2.5 hours of stereo audio recordings of 4718 vehicle passing events captured with both high-quality sE8 and medium-quality MEMS microphones. This dataset is well suited to evaluate the use-case of deploying audio classification algorithms to embedded sensor devices with restricted microphone quality and hardware processing power. In addition, this paper provides a detailed review of recent acoustic traffic monitoring (ATM) algorithms as well as the results of two benchmark experiments on vehicle type classification and direction of movement estimation using four state-of-the-art convolutional neural network architectures.

Neural IR models have often been studied in homogeneous and narrow settings, which has considerably limited insights into their generalization capabilities. To address this, and to allow researchers to more broadly establish the effectiveness of their models, we introduce BEIR (Benchmarking IR), a heterogeneous benchmark for information retrieval. We leverage a careful selection of 17 datasets for evaluation spanning diverse retrieval tasks including open-domain datasets as well as narrow expert domains. We study the effectiveness of nine state-of-the-art retrieval models in a zero-shot evaluation setup on BEIR, finding that performing well consistently across all datasets is challenging. Our results show BM25 is a robust baseline and Reranking-based models overall achieve the best zero-shot performances, however, at high computational costs. In contrast, Dense-retrieval models are computationally more efficient but often underperform other approaches, highlighting the considerable room for improvement in their generalization capabilities. In this work, we extensively analyze different retrieval models and provide several suggestions that we believe may be useful for future work. BEIR datasets and code are available at https://github.com/UKPLab/beir.

We translate a closed text that is known in advance into a severely low resource language by leveraging massive source parallelism. In other words, given a text in 124 source languages, we translate it into a severely low resource language using only ~1,000 lines of low resource data without any external help. Firstly, we propose a systematic method to rank and choose source languages that are close to the low resource language. We call the linguistic definition of language family Family of Origin (FAMO), and we call the empirical definition of higher-ranked languages using our metrics Family of Choice (FAMC). Secondly, we build an Iteratively Pretrained Multilingual Order-preserving Lexiconized Transformer (IPML) to train on ~1,000 lines (~3.5%) of low resource data. To translate named entities correctly, we build a massive lexicon table for 2,939 Bible named entities in 124 source languages, and include many that occur once and covers more than 66 severely low resource languages. Moreover, we also build a novel method of combining translations from different source languages into one. Using English as a hypothetical low resource language, we get a +23.9 BLEU increase over a multilingual baseline, and a +10.3 BLEU increase over our asymmetric baseline in the Bible dataset. We get a 42.8 BLEU score for Portuguese-English translation on the medical EMEA dataset. We also have good results for a real severely low resource Mayan language, Eastern Pokomchi.

When Ekman arrived in the tropics of Okapa, he ran experiments to assess how the Fore recognized emotions. Because the Fore had minimal contact with Westerners and mass media, Ekman had theorized that their recognition and display of core expressions would prove that such expressions were universal. He would show them flash cards of facial expressions and see if they described the emotion as he did. In Ekman's own words, "All I was doing was showing funny pictures." But Ekman had no training in Fore history, language, culture, or politics.

Artificial intelligence is changing the world we live in. It will redefine the workplace and have significant implications for everything we do, probably by the end of this decade. Some AI applications are already a part of our everyday lives, such as intelligent car navigation systems. So, what is artificial intelligence? AI can be defined as'the ability of machines to perform tasks that normally require human intelligence'. AI has in fact been around for several decades.

Computer science unplugged (CS Unplugged, or just "Unplugged") is a pedagogy for teaching computational ideas to grade-school students without using a computer.a It was developed in the early 1990s as a necessity when working with computers in the classroom was not usually practical, but it still finds widespread adoption as a supplement to computer-based lessons, even where devices are readily available. This appears as a contradiction to some (if you are teaching computer science, why not spend as much time as possible on a computer?), Unfortunately, Unplugged can also be used to justify poor decisions by treating it as a complete curriculum in itself--a teacher who does not have the time or support to extend themselves in new curriculum content might rely on Unplugged as "enough," or administrators might justify a lack of funding by suggesting that schools use Unplugged teaching instead of buying devices. The Unplugged approach is widely used, mentioned in dozens of research papers about CS education, has been translated into many languages, and is widely used in teacher professional development.1

On Earth, avoiding collisions is a key priority for traffic cops, air traffic controllers, and the parents of toddlers. It is no different in space--and perhaps even more critical--given that objects orbiting the Earth are moving at more than 17,000 m.p.h., which means that even very small objects less than a centimeter in diameter have caused damage to the International Space Station, the Space Shuttle, and satellites. In fact, the U.S. National Aeronautics and Space Administration (NASA) estimates there are more than 500,000 such objects orbiting the Earth that are larger than a marble, and at least a million smaller pieces of debris that cannot be tracked. Based on the growing number of commercial and government launches of spacecraft, satellites, and even space stations, the number of objects that will need to be catalogued, tracked, and managed is expected to grow significantly in the coming years. And the solutions to this issue are fraught with both technical and political challenges.

As CEO of Hitachi Vantara, Gajen helps solve clients' problems by bringing to bear Hitachi's unrivaled industrial expertise across sectors. You might not think saving the world's tropical rainforests is a data challenge, but the urgent task of protecting the last remaining two million square miles of forest is precisely that. What is more, the challenge holds vital lessons for anyone tackling a data project with seemingly insurmountable odds. Logging, much of it illegal, strips the planet of more than 32 million acres of natural forest every year. If you ever imagined literally trying to find a needle in a haystack, then you might be able to contemplate what it is like to find a chainsaw in forested areas the size of Australia.