今日 cs.CL方向共计13篇文章。
推理分析(1篇)
[1]：Analogical Reasoning for Visually Grounded Language Acquisition
标题：基于视觉的语言习得中的类比推理
作者：Bo Wu, Haoyu Qin, Alireza Zareian, Carl Vondrick, Shih-Fu Chang
备注：12 pages
链接：https://arxiv.org/abs/2007.11668
摘要：Children acquire language subconsciously by observing the surrounding world and listening to descriptions. They can discover the meaning of words even without explicit language knowledge, and generalize to novel compositions effortlessly. In this paper, we bring this ability to AI, by studying the task of Visually grounded Language Acquisition (VLA). We propose a multimodal transformer model augmented with a novel mechanism for analogical reasoning, which approximates novel compositions by learning semantic mapping and reasoning operations from previously seen compositions. Our proposed method, Analogical Reasoning Transformer Networks (ARTNet), is trained on raw multimedia data (video frames and transcripts), and after observing a set of compositions such as "washing apple" or "cutting carrot", it can generalize and recognize new compositions in new video frames, such as "washing carrot" or "cutting apple". To this end, ARTNet refers to relevant instances in the training data and uses their visual features and captions to establish analogies with the query image. Then it chooses the suitable verb and noun to create a new composition that describes the new image best. Extensive experiments on an instructional video dataset demonstrate that the proposed method achieves significantly better generalization capability and recognition accuracy compared to state-of-the-art transformer models.自然语言生成(1篇)
[1]：Product Title Generation for Conversational Systems using BERT
标题：基于BERT的会话系统产品标题生成
作者：Mansi Ranjit Mane, Shashank Kedia, Aditya Mantha, Stephen Guo, Kannan Achan
链接：https://arxiv.org/abs/2007.11768
摘要：Through recent advancements in speech technology and introduction of smart devices, such as Amazon Alexa and Google Home, increasing number of users are interacting with applications through voice. E-commerce companies typically display short product titles on their webpages, either human-curated or algorithmically generated, when brevity is required, but these titles are dissimilar from natural spoken language. For example, "Lucky Charms Gluten Free Break-fast Cereal, 20.5 oz a box Lucky Charms Gluten Free" is acceptable to display on a webpage, but "a 20.5 ounce box of lucky charms gluten free cereal" is easier to comprehend over a conversational system. As compared to display devices, where images and detailed product information can be presented to users, short titles for products are necessary when interfacing with voice assistants. We propose a sequence-to-sequence approach using BERT to generate short, natural, spoken language titles from input web titles. Our extensive experiments on a real-world industry dataset and human evaluation of model outputs, demonstrate that BERT summarization outperforms comparable baseline models.情感分析(2篇)
[1]：NITS-Hinglish-SentiMix at SemEval-2020 Task 9: Sentiment Analysis For Code-Mixed Social Media Text
标题：NITS Hinglish SentiMix在SemEval-2020任务9：代码混合社交媒体文本的情感分析
作者：Subhra Jyoti Baroi, Nivedita Singh, Ringki Das, Thoudam Doren Singh
备注：In Proceedings of the 14th International Workshop on Semantic Evaluation (SemEval-2020), Barcelona, Spain, December. Association for Computational Linguistics
链接：https://arxiv.org/abs/2007.12081
摘要：Sentiment Analysis is the process of deciphering what a sentence emotes and classifying them as either positive, negative, or neutral. In recent times, India has seen a huge influx in the number of active social media users and this has led to a plethora of unstructured text data. Since the Indian population is generally fluent in both Hindi and English, they end up generating code-mixed Hinglish social media text i.e. the expressions of Hindi language, written in the Roman script alongside other English words. The ability to adequately comprehend the notions in these texts is truly necessary. Our team, rns2020 participated in Task 9 at SemEval2020 intending to design a system to carry out the sentiment analysis of code-mixed social media text. This work proposes a system named NITS-Hinglish-SentiMix to viably complete the sentiment analysis of such code-mixed Hinglish text. The proposed framework has recorded an F-Score of 0.617 on the test data.
[2]：HCMS at SemEval-2020 Task 9: A Neural Approach to Sentiment Analysis for Code-Mixed Texts
标题：SemEval-2020任务9：代码混合文本情感分析的神经方法
作者：Aditya Srivastava, V. Harsha Vardhan
备注：6 pages, 2 figures, 4 tables, math equations, to be published in the proceedings of the 14th International Workshop on Semantic Evaluation (SemEval) 2020, Association for Computational Linguistics (ACL). Code for the paper is available atthis https URL. Data and task description is available atthis https URL
链接：https://arxiv.org/abs/2007.12076
摘要：Problems involving code-mixed language are often plagued by a lack of resources and an absence of materials to perform sophisticated transfer learning with. In this paper we describe our submission to the Sentimix Hindi-English task involving sentiment classification of code-mixed texts, and with an F1 score of 67.1%, we demonstrate that simple convolution and attention may well produce reasonable results.模型(1篇)
[1]：Effects of Language Relatedness for Cross-lingual Transfer Learning in Character-Based Language Models
标题：基于特征的语言模型中语言关联性对跨语言迁移学习的影响
作者：Mittul Singh, Peter Smit, Sami Virpioja, Mikko Kurimo
链接：https://arxiv.org/abs/2007.11648
摘要：Character-based Neural Network Language Models (NNLM) have the advantage of smaller vocabulary and thus faster training times in comparison to NNLMs based on multi-character units. However, in low-resource scenarios, both the character and multi-character NNLMs suffer from data sparsity. In such scenarios, cross-lingual transfer has improved multi-character NNLM performance by allowing information transfer from a source to the target language. In the same vein, we propose to use cross-lingual transfer for character NNLMs applied to low-resource Automatic Speech Recognition (ASR). However, applying cross-lingual transfer to character NNLMs is not as straightforward. We observe that relatedness of the source language plays an important role in cross-lingual pretraining of character NNLMs. We evaluate this aspect on ASR tasks for two target languages: Finnish (with English and Estonian as source) and Swedish (with Danish, Norwegian, and English as source). Prior work has observed no difference between using the related or unrelated language for multi-character NNLMs. We, however, show that for character-based NNLMs, only pretraining with a related language improves the ASR performance, and using an unrelated language may deteriorate it. We also observe that the benefits are larger when there is much lesser target data than source data.其他(8篇)
[1]：Health, Psychosocial, and Social issues emanating from COVID-19 pandemic based on Social Media Comments using Natural Language Processing
标题：COVID-19大流行引起的健康、心理社会和社会问题基于使用自然语言处理的社交媒体评论
作者：Oladapo Oyebode, Chinenye Ndulue, Ashfaq Adib, Dinesh Mulchandani, Banuchitra Suruliraj, Fidelia Anulika Orji, Christine Chambers, Sandra Meier, Rita Orji
链接：https://arxiv.org/abs/2007.12144
摘要：The COVID-19 pandemic has caused a global health crisis that affects many aspects of human lives. In the absence of vaccines and antivirals, several behavioural change and policy initiatives, such as physical distancing, have been implemented to control the spread of the coronavirus. Social media data can reveal public perceptions toward how governments and health agencies across the globe are handling the pandemic, as well as the impact of the disease on people regardless of their geographic locations in line with various factors that hinder or facilitate the efforts to control the spread of the pandemic globally. This paper aims to investigate the impact of the COVID-19 pandemic on people globally using social media data. We apply natural language processing (NLP) and thematic analysis to understand public opinions, experiences, and issues with respect to the COVID-19 pandemic using social media data. First, we collect over 47 million COVID-19-related comments from Twitter, Facebook, YouTube, and three online discussion forums. Second, we perform data preprocessing which involves applying NLP techniques to clean and prepare the data for automated theme extraction. Third, we apply context-aware NLP approach to extract meaningful keyphrases or themes from over 1 million randomly selected comments, as well as compute sentiment scores for each theme and assign sentiment polarity based on the scores using lexicon-based technique. Fourth, we categorize related themes into broader themes. A total of 34 negative themes emerged, out of which 15 are health-related issues, psychosocial issues, and social issues related to the COVID-19 pandemic from the public perspective. In addition, 20 positive themes emerged from our results. Finally, we recommend interventions that can help address the negative issues based on the positive themes and other remedial ideas rooted in research.
[2]：Deep Learning based, end-to-end metaphor detection in Greek language with Recurrent and Convolutional Neural Networks
标题：基于深度学习的递归和卷积神经网络的希腊语端到端隐喻检测
作者：Konstantinos Perifanos, Eirini Florou, Dionysis Goutsos
链接：https://arxiv.org/abs/2007.11949
摘要：This paper presents and benchmarks a number of end-to-end Deep Learning based models for metaphor detection in Greek. We combine Convolutional Neural Networks and Recurrent Neural Networks with representation learning to bear on the metaphor detection problem for the Greek language. The models presented achieve exceptional accuracy scores, significantly improving the previous state of the art results, which had already achieved accuracy 0.82. Furthermore, no special preprocessing, feature engineering or linguistic knowledge is used in this work. The methods presented achieve accuracy of 0.92 and F-score 0.92 with Convolutional Neural Networks (CNNs) and bidirectional Long Short Term Memory networks (LSTMs). Comparable results of 0.91 accuracy and 0.91 F-score are also achieved with bidirectional Gated Recurrent Units (GRUs) and Convolutional Recurrent Neural Nets (CRNNs). The models are trained and evaluated only on the basis of the training tuples, the sentences and their labels. The outcome is a state of the art collection of metaphor detection models, trained on limited labelled resources, which can be extended to other languages and similar tasks.
[3]：AI4D -- African Language Dataset Challenge
标题：AI4D——非洲语言数据集挑战
作者：Kathleen Siminyu, Sackey Freshia, Jade Abbott, Vukosi Marivate
链接：https://arxiv.org/abs/2007.11865
摘要：As language and speech technologies become more advanced, the lack of fundamental digital resources for African languages, such as data, spell checkers and Part of Speech taggers, means that the digital divide between these languages and others keeps growing. This work details the organisation of the AI4D - African Language Dataset Challenge, an effort to incentivize the creation, organization and discovery of African language datasets through a competitive challenge. We particularly encouraged the submission of annotated datasets which can be used for training task-specific supervised machine learning models.
[4]：Applying GPGPU to Recurrent Neural Network Language Model based Fast Network Search in the Real-Time LVCSR
标题：GPGPU在递归神经网络中的应用基于语言模型的实时快速网络搜索
作者：Kyungmin Lee, Chiyoun Park, Ilhwan Kim, Namhoon Kim, Jaewon Lee
备注：4 pages, 2 figures, Interspeech2015(Accepted)
链接：https://arxiv.org/abs/2007.11794
摘要：Recurrent Neural Network Language Models (RNNLMs) have started to be used in various fields of speech recognition due to their outstanding performance. However, the high computational complexity of RNNLMs has been a hurdle in applying the RNNLM to a real-time Large Vocabulary Continuous Speech Recognition (LVCSR). In order to accelerate the speed of RNNLM-based network searches during decoding, we apply the General Purpose Graphic Processing Units (GPGPUs). This paper proposes a novel method of applying GPGPUs to RNNLM-based graph traversals. We have achieved our goal by reducing redundant computations on CPUs and amount of transfer between GPGPUs and CPUs. The proposed approach was evaluated on both WSJ corpus and in-house data. Experiments shows that the proposed approach achieves the real-time speed in various circumstances while maintaining the Word Error Rate (WER) to be relatively 10% lower than that of n-gram models.
[5]：Revealing semantic and emotional structure of suicide notes with cognitive network science
标题：用认知网络科学揭示自杀笔记的语义和情感结构
作者：Andreia Sofia Teixeira, Szymon Talaga, Trevor James Swanson, Massimo Stella
链接：https://arxiv.org/abs/2007.12053
摘要：Understanding the cognitive and emotional perceptions of people who commit suicide is one of the most sensitive scientific challenges. There are circumstances where people feel the need to leave something written, an artifact where they express themselves, registering their last words and feelings. These suicide notes are of utmost importance for better understanding the psychology of suicidal ideation. This work gives structure to the linguistic content of suicide notes, revealing interconnections between cognitive and emotional states of people who committed suicide. We build upon cognitive network science, psycholinguistics and semantic frame theory to introduce a network representation of the mindset expressed in suicide notes. Our cognitive network representation enables the quantitative analysis of the language in suicide notes through structural balance theory, semantic prominence and emotional profiling. Our results indicate that the emotional syntax connecting positively- and negatively-valenced terms gives rise to a degree of structural balance that is significantly higher than null models where the affective structure was randomized. We show that suicide notes are affectively compartmentalized such that positive concepts tend to cluster together and dominate the overall network structure. A key positive concept is "love", which integrates information relating the self to others in ways that are semantically prominent across suicide notes. The emotions populating the semantic frame of "love" combine joy and trust with anticipation and sadness, which connects with psychological theories about meaning-making and narrative psychology. Our results open new ways for understanding the structure of genuine suicide notes informing future research for suicide prevention.
[6]：SBAT: Video Captioning with Sparse Boundary-Aware Transformer
标题：SBAT：具有稀疏边界感知变换器的视频字幕
作者：Tao Jin, Siyu Huang, Ming Chen, Yingming Li, Zhongfei Zhang
备注：Appearing at IJCAI 2020
链接：https://arxiv.org/abs/2007.11888
摘要：In this paper, we focus on the problem of applying the transformer structure to video captioning effectively. The vanilla transformer is proposed for uni-modal language generation task such as machine translation. However, video captioning is a multimodal learning problem, and the video features have much redundancy between different time steps. Based on these concerns, we propose a novel method called sparse boundary-aware transformer (SBAT) to reduce the redundancy in video representation. SBAT employs boundary-aware pooling operation for scores from multihead attention and selects diverse features from different scenarios. Also, SBAT includes a local correlation scheme to compensate for the local information loss brought by sparse operation. Based on SBAT, we further propose an aligned cross-modal encoding scheme to boost the multimodal interaction. Experimental results on two benchmark datasets show that SBAT outperforms the state-of-the-art methods under most of the metrics.
[7]：Clustering of Social Media Messages for Humanitarian Aid Response during Crisis
标题：危机期间人道主义援助响应的社交媒体信息聚集
作者：Swati Padhee, Tanay Kumar Saha, Joel Tetreault, Alejandro Jaimes
备注：6 pages, 1 figure. Research work was done while Swati was interning at Dataminr Inc. and presented at the AI for Social Good, Harvard CRCS Workshop 2020 (this https URL)
链接：https://arxiv.org/abs/2007.11756
摘要：Social media has quickly grown into an essential tool for people to communicate and express their needs during crisis events. Prior work in analyzing social media data for crisis management has focused primarily on automatically identifying actionable (or, informative) crisis-related messages. In this work, we show that recent advances in Deep Learning and Natural Language Processing outperform prior approaches for the task of classifying informativeness and encourage the field to adopt them for their research or even deployment. We also extend these methods to two sub-tasks of informativeness and find that the Deep Learning methods are effective here as well.
[8]：Integrating Image Captioning with Rule-based Entity Masking
标题：图像字幕与基于规则的实体屏蔽的集成
作者：Aditya Mogadala, Xiaoyu Shen, Dietrich Klakow
链接：https://arxiv.org/abs/2007.11690
摘要：Given an image, generating its natural language description (i.e., caption) is a well studied problem. Approaches proposed to address this problem usually rely on image features that are difficult to interpret. Particularly, these image features are subdivided into global and local features, where global features are extracted from the global representation of the image, while local features are extracted from the objects detected locally in an image. Although, local features extract rich visual information from the image, existing models generate captions in a blackbox manner and humans have difficulty interpreting which local objects the caption is aimed to represent. Hence in this paper, we propose a novel framework for the image captioning with an explicit object (e.g., knowledge graph entity) selection process while still maintaining its end-to-end training ability. The model first explicitly selects which local entities to include in the caption according to a human-interpretable mask, then generate proper captions by attending to selected entities. Experiments conducted on the MSCOCO dataset demonstrate that our method achieves good performance in terms of the caption quality and diversity with a more interpretable generating process than previous counterparts.
