scholarly journals Unicoder-VL: A Universal Encoder for Vision and Language by Cross-Modal Pre-Training

2020 ◽  
Vol 34 (07) ◽  
pp. 11336-11344 ◽  
Author(s):  
Gen Li ◽  
Nan Duan ◽  
Yuejian Fang ◽  
Ming Gong ◽  
Daxin Jiang
Keyword(s):  
Large Scale ◽  
State Of The Art ◽  
Language Modeling ◽  
Context Aware ◽  
Commonsense Reasoning ◽  
Output Layer ◽  
The Cross ◽  
Cross Lingual ◽  
Vision And Language ◽  
Image Caption

We propose Unicoder-VL, a universal encoder that aims to learn joint representations of vision and language in a pre-training manner. Borrow ideas from cross-lingual pre-trained models, such as XLM (Lample and Conneau 2019) and Unicoder (Huang et al. 2019), both visual and linguistic contents are fed into a multi-layer Transformer (Vaswani et al. 2017) for the cross-modal pre-training, where three pre-trained tasks are employed, including Masked Language Modeling(MLM), Masked Object Classification(MOC) and Visual-linguistic Matching(VLM). The first two tasks learn context-aware representations for input tokens based on linguistic and visual contents jointly. The last task tries to predict whether an image and a text describe each other. After pretraining on large-scale image-caption pairs, we transfer Unicoder-VL to caption-based image-text retrieval and visual commonsense reasoning, with just one additional output layer. We achieve state-of-the-art or comparable results on both two tasks and show the powerful ability of the cross-modal pre-training.

scholarly journals WinoGrande

2021 ◽  
Vol 64 (9) ◽  
pp. 99-106
Author(s):  
Keisuke Sakaguchi ◽  
Ronan Le Bras ◽  
Chandra Bhagavatula ◽  
Yejin Choi
Keyword(s):  
High Performance ◽  
Large Scale ◽  
State Of The Art ◽  
Bias Reduction ◽  
Language Models ◽  
Systematic Bias ◽  
Commonsense Reasoning ◽  
Word Associations ◽  
Lower Accuracy ◽  
Key Steps

Commonsense reasoning remains a major challenge in AI, and yet, recent progresses on benchmarks may seem to suggest otherwise. In particular, the recent neural language models have reported above 90% accuracy on the Winograd Schema Challenge (WSC), a commonsense benchmark originally designed to be unsolvable for statistical models that rely simply on word associations. This raises an important question---whether these models have truly acquired robust commonsense capabilities or they rely on spurious biases in the dataset that lead to an overestimation of the true capabilities of machine commonsense. To investigate this question, we introduce WinoGrande, a large-scale dataset of 44k problems, inspired by the original WSC, but adjusted to improve both the scale and the hardness of the dataset. The key steps of the dataset construction consist of (1) large-scale crowdsourcing, followed by (2) systematic bias reduction using a novel AFLITE algorithm that generalizes human-detectable word associations to machine-detectable embedding associations. Our experiments demonstrate that state-of-the-art models achieve considerably lower accuracy (59.4%-79.1%) on WINOGRANDE compared to humans (94%), confirming that the high performance on the original WSC was inflated by spurious biases in the dataset. Furthermore, we report new state-of-the-art results on five related benchmarks with emphasis on their dual implications. On the one hand, they demonstrate the effectiveness of WINOGRANDE when used as a resource for transfer learning. On the other hand, the high performance on all these benchmarks suggests the extent to which spurious biases are prevalent in all such datasets, which motivates further research on algorithmic bias reduction.

scholarly journals KEPLER: A Unified Model for Knowledge Embedding and Pre-trained Language Representation

2021 ◽  
Vol 9 ◽  
pp. 176-194
Author(s):  
Xiaozhi Wang ◽  
Tianyu Gao ◽  
Zhaocheng Zhu ◽  
Zhengyan Zhang ◽  
Zhiyuan Liu ◽  
...  
Keyword(s):  
Link Prediction ◽  
Large Scale ◽  
State Of The Art ◽  
Source Code ◽  
Language Modeling ◽  
Unified Model ◽  
Factual Knowledge ◽  
Textual Information ◽  
Language Representation ◽  
Knowledge Graphs

Abstract Pre-trained language representation models (PLMs) cannot well capture factual knowledge from text. In contrast, knowledge embedding (KE) methods can effectively represent the relational facts in knowledge graphs (KGs) with informative entity embeddings, but conventional KE models cannot take full advantage of the abundant textual information. In this paper, we propose a unified model for Knowledge Embedding and Pre-trained LanguagERepresentation (KEPLER), which can not only better integrate factual knowledge into PLMs but also produce effective text-enhanced KE with the strong PLMs. In KEPLER, we encode textual entity descriptions with a PLM as their embeddings, and then jointly optimize the KE and language modeling objectives. Experimental results show that KEPLER achieves state-of-the-art performances on various NLP tasks, and also works remarkably well as an inductive KE model on KG link prediction. Furthermore, for pre-training and evaluating KEPLER, we construct Wikidata5M1 , a large-scale KG dataset with aligned entity descriptions, and benchmark state-of-the-art KE methods on it. It shall serve as a new KE benchmark and facilitate the research on large KG, inductive KE, and KG with text. The source code can be obtained from https://github.com/THU-KEG/KEPLER.

scholarly journals WinoGrande: An Adversarial Winograd Schema Challenge at Scale

2020 ◽  
Vol 34 (05) ◽  
pp. 8732-8740 ◽  
Author(s):  
Keisuke Sakaguchi ◽  
Ronan Le Bras ◽  
Chandra Bhagavatula ◽  
Yejin Choi
Keyword(s):  
Large Scale ◽  
Human Performance ◽  
State Of The Art ◽  
Bias Reduction ◽  
Training Data ◽  
Language Models ◽  
Systematic Bias ◽  
Commonsense Reasoning ◽  
Word Associations ◽  
Key Steps

The Winograd Schema Challenge (WSC) (Levesque, Davis, and Morgenstern 2011), a benchmark for commonsense reasoning, is a set of 273 expert-crafted pronoun resolution problems originally designed to be unsolvable for statistical models that rely on selectional preferences or word associations. However, recent advances in neural language models have already reached around 90% accuracy on variants of WSC. This raises an important question whether these models have truly acquired robust commonsense capabilities or whether they rely on spurious biases in the datasets that lead to an overestimation of the true capabilities of machine commonsense.To investigate this question, we introduce WinoGrande, a large-scale dataset of 44k problems, inspired by the original WSC design, but adjusted to improve both the scale and the hardness of the dataset. The key steps of the dataset construction consist of (1) a carefully designed crowdsourcing procedure, followed by (2) systematic bias reduction using a novel AfLite algorithm that generalizes human-detectable word associations to machine-detectable embedding associations. The best state-of-the-art methods on WinoGrande achieve 59.4 – 79.1%, which are ∼15-35% (absolute) below human performance of 94.0%, depending on the amount of the training data allowed (2% – 100% respectively).Furthermore, we establish new state-of-the-art results on five related benchmarks — WSC (→ 90.1%), DPR (→ 93.1%), COPA(→ 90.6%), KnowRef (→ 85.6%), and Winogender (→ 97.1%). These results have dual implications: on one hand, they demonstrate the effectiveness of WinoGrande when used as a resource for transfer learning. On the other hand, they raise a concern that we are likely to be overestimating the true capabilities of machine commonsense across all these benchmarks. We emphasize the importance of algorithmic bias reduction in existing and future benchmarks to mitigate such overestimation.

scholarly journals Unsupervised cross-lingual model transfer for named entity recognition with contextualized word representations

PLoS ONE ◽  
2021 ◽  
Vol 16 (9) ◽  
pp. e0257230
Author(s):  
Huijiong Yan ◽  
Tao Qian ◽  
Liang Xie ◽  
Shanguang Chen
Keyword(s):  
Language Processing ◽  
Large Scale ◽  
Named Entity Recognition ◽  
Entity Recognition ◽  
Language Models ◽  
Competitive Performance ◽  
Named Entity ◽  
Model Transfer ◽  
The Cross ◽  
Cross Lingual

Named entity recognition (NER) is one fundamental task in the natural language processing (NLP) community. Supervised neural network models based on contextualized word representations can achieve highly-competitive performance, which requires a large-scale manually-annotated corpus for training. While for the resource-scarce languages, the construction of such as corpus is always expensive and time-consuming. Thus, unsupervised cross-lingual transfer is one good solution to address the problem. In this work, we investigate the unsupervised cross-lingual NER with model transfer based on contextualized word representations, which greatly advances the cross-lingual NER performance. We study several model transfer settings of the unsupervised cross-lingual NER, including (1) different types of the pretrained transformer-based language models as input, (2) the exploration strategies of the multilingual contextualized word representations, and (3) multi-source adaption. In particular, we propose an adapter-based word representation method combining with parameter generation network (PGN) better to capture the relationship between the source and target languages. We conduct experiments on a benchmark ConLL dataset involving four languages to simulate the cross-lingual setting. Results show that we can obtain highly-competitive performance by cross-lingual model transfer. In particular, our proposed adapter-based PGN model can lead to significant improvements for cross-lingual NER.

scholarly journals Documentary data and the study of past droughts: a global state of the art

2018 ◽  
Vol 14 (12) ◽  
pp. 1915-1960 ◽  
Author(s):  
Rudolf Brázdil ◽  
Andrea Kiss ◽  
Jürg Luterbacher ◽  
David J. Nash ◽  
Ladislava Řezníčková
Keyword(s):  
Large Scale ◽  
State Of The Art ◽  
Drought Indices ◽  
Documentary Evidence ◽  
Climatic Trends ◽  
Instrumental Observations ◽  
Spatio Temporal ◽  
Epigraphic Evidence ◽  
Administrative Evidence

Abstract. The use of documentary evidence to investigate past climatic trends and events has become a recognised approach in recent decades. This contribution presents the state of the art in its application to droughts. The range of documentary evidence is very wide, including general annals, chronicles, memoirs and diaries kept by missionaries, travellers and those specifically interested in the weather; records kept by administrators tasked with keeping accounts and other financial and economic records; legal-administrative evidence; religious sources; letters; songs; newspapers and journals; pictographic evidence; chronograms; epigraphic evidence; early instrumental observations; society commentaries; and compilations and books. These are available from many parts of the world. This variety of documentary information is evaluated with respect to the reconstruction of hydroclimatic conditions (precipitation, drought frequency and drought indices). Documentary-based drought reconstructions are then addressed in terms of long-term spatio-temporal fluctuations, major drought events, relationships with external forcing and large-scale climate drivers, socio-economic impacts and human responses. Documentary-based drought series are also considered from the viewpoint of spatio-temporal variability for certain continents, and their employment together with hydroclimate reconstructions from other proxies (in particular tree rings) is discussed. Finally, conclusions are drawn, and challenges for the future use of documentary evidence in the study of droughts are presented.

Enhanced context-aware recommendation using topic modeling and particle swarm optimization

2021 ◽  
pp. 1-16
Author(s):  
Ibtissem Gasmi ◽  
Mohamed Walid Azizi ◽  
Hassina Seridi-Bouchelaghem ◽  
Nabiha Azizi ◽  
Samir Brahim Belhaouari
Keyword(s):  
Topic Modeling ◽  
Latent Dirichlet Allocation ◽  
State Of The Art ◽  
Weighting Function ◽  
Contextual Factors ◽  
Pearson Correlation ◽  
Correlation Coefficients ◽  
Pso Algorithm ◽  
Context Aware ◽  
Proposed Model

Context-Aware Recommender System (CARS) suggests more relevant services by adapting them to the user’s specific context situation. Nevertheless, the use of many contextual factors can increase data sparsity while few context parameters fail to introduce the contextual effects in recommendations. Moreover, several CARSs are based on similarity algorithms, such as cosine and Pearson correlation coefficients. These methods are not very effective in the sparse datasets. This paper presents a context-aware model to integrate contextual factors into prediction process when there are insufficient co-rated items. The proposed algorithm uses Latent Dirichlet Allocation (LDA) to learn the latent interests of users from the textual descriptions of items. Then, it integrates both the explicit contextual factors and their degree of importance in the prediction process by introducing a weighting function. Indeed, the PSO algorithm is employed to learn and optimize weights of these features. The results on the Movielens 1 M dataset show that the proposed model can achieve an F-measure of 45.51% with precision as 68.64%. Furthermore, the enhancement in MAE and RMSE can respectively reach 41.63% and 39.69% compared with the state-of-the-art techniques.

scholarly journals VIPPrint: Validating Synthetic Image Detection and Source Linking Methods on a Large Scale Dataset of Printed Documents

2021 ◽  
Vol 7 (3) ◽  
pp. 50
Author(s):  
Anselmo Ferreira ◽  
Ehsan Nowroozi ◽  
Mauro Barni
Keyword(s):  
Large Scale ◽  
State Of The Art ◽  
Child Pornography ◽  
Forensic Analysis ◽  
Synthetic Image ◽  
Image Detection ◽  
Face Images ◽  
Large Scale Dataset ◽  
Scanned Images ◽  
Analysis Of The Images

The possibility of carrying out a meaningful forensic analysis on printed and scanned images plays a major role in many applications. First of all, printed documents are often associated with criminal activities, such as terrorist plans, child pornography, and even fake packages. Additionally, printing and scanning can be used to hide the traces of image manipulation or the synthetic nature of images, since the artifacts commonly found in manipulated and synthetic images are gone after the images are printed and scanned. A problem hindering research in this area is the lack of large scale reference datasets to be used for algorithm development and benchmarking. Motivated by this issue, we present a new dataset composed of a large number of synthetic and natural printed face images. To highlight the difficulties associated with the analysis of the images of the dataset, we carried out an extensive set of experiments comparing several printer attribution methods. We also verified that state-of-the-art methods to distinguish natural and synthetic face images fail when applied to print and scanned images. We envision that the availability of the new dataset and the preliminary experiments we carried out will motivate and facilitate further research in this area.

scholarly journals SketchGNN: Semantic Sketch Segmentation with Graph Neural Networks

2021 ◽  
Vol 40 (3) ◽  
pp. 1-13
Author(s):  
Lumin Yang ◽  
Jiajie Zhuang ◽  
Hongbo Fu ◽  
Xiangzhi Wei ◽  
Kun Zhou ◽  
...  
Keyword(s):  
Neural Network ◽  
Neural Networks ◽  
Network Architecture ◽  
Large Scale ◽  
State Of The Art ◽  
Semantic Segmentation ◽  
Structure Information ◽  
Graph Neural Networks ◽  
Node Labels ◽  
Point Level

We introduce SketchGNN , a convolutional graph neural network for semantic segmentation and labeling of freehand vector sketches. We treat an input stroke-based sketch as a graph with nodes representing the sampled points along input strokes and edges encoding the stroke structure information. To predict the per-node labels, our SketchGNN uses graph convolution and a static-dynamic branching network architecture to extract the features at three levels, i.e., point-level, stroke-level, and sketch-level. SketchGNN significantly improves the accuracy of the state-of-the-art methods for semantic sketch segmentation (by 11.2% in the pixel-based metric and 18.2% in the component-based metric over a large-scale challenging SPG dataset) and has magnitudes fewer parameters than both image-based and sequence-based methods.

scholarly journals ShadingNet: Image Intrinsics by Fine-Grained Shading Decomposition

2021 ◽  
Author(s):  
Anil S. Baslamisli ◽  
Partha Das ◽  
Hoang-An Le ◽  
Sezer Karaoglu ◽  
Theo Gevers
Keyword(s):  
Neural Network ◽  
Large Scale ◽  
State Of The Art ◽  
Image Decomposition ◽  
Natural Environments ◽  
Decomposition Algorithms ◽  
Ambient Light ◽  
Fine Grained ◽  
Large Scale Dataset ◽  
Direct Illumination

AbstractIn general, intrinsic image decomposition algorithms interpret shading as one unified component including all photometric effects. As shading transitions are generally smoother than reflectance (albedo) changes, these methods may fail in distinguishing strong photometric effects from reflectance variations. Therefore, in this paper, we propose to decompose the shading component into direct (illumination) and indirect shading (ambient light and shadows) subcomponents. The aim is to distinguish strong photometric effects from reflectance variations. An end-to-end deep convolutional neural network (ShadingNet) is proposed that operates in a fine-to-coarse manner with a specialized fusion and refinement unit exploiting the fine-grained shading model. It is designed to learn specific reflectance cues separated from specific photometric effects to analyze the disentanglement capability. A large-scale dataset of scene-level synthetic images of outdoor natural environments is provided with fine-grained intrinsic image ground-truths. Large scale experiments show that our approach using fine-grained shading decompositions outperforms state-of-the-art algorithms utilizing unified shading on NED, MPI Sintel, GTA V, IIW, MIT Intrinsic Images, 3DRMS and SRD datasets.

App2Vec: Context-Aware Application Usage Prediction

2021 ◽  
Vol 15 (6) ◽  
pp. 1-21
Author(s):  
Huandong Wang ◽  
Yong Li ◽  
Mu Du ◽  
Zhenhui Li ◽  
Depeng Jin
Keyword(s):  
Dirichlet Process ◽  
Service Providers ◽  
State Of The Art ◽  
Representation Learning ◽  
Context Aware ◽  
Challenging Problem ◽  
Performance Gap ◽  
Bayesian Mixture Model ◽  
Bayesian Mixture ◽  
Spatio Temporal

Both app developers and service providers have strong motivations to understand when and where certain apps are used by users. However, it has been a challenging problem due to the highly skewed and noisy app usage data. Moreover, apps are regarded as independent items in existing studies, which fail to capture the hidden semantics in app usage traces. In this article, we propose App2Vec, a powerful representation learning model to learn the semantic embedding of apps with the consideration of spatio-temporal context. Based on the obtained semantic embeddings, we develop a probabilistic model based on the Bayesian mixture model and Dirichlet process to capture when , where , and what semantics of apps are used to predict the future usage. We evaluate our model using two different app usage datasets, which involve over 1.7 million users and 2,000+ apps. Evaluation results show that our proposed App2Vec algorithm outperforms the state-of-the-art algorithms in app usage prediction with a performance gap of over 17.0%.

Sign in / Sign up

Export Citation Format

Share Document