automatic question generation from paragraph

It is clearly understood that generating the test question is the toughest part. the articles by crowd-workers. Putting the horse before the cart: A generator-evaluator framework Random questions can be a wonderful way to begin a writing session each day. Pranav Rajpurkar, Jian Zhang, Konstantin Lopyrev, and Percy Liang. > Question Generation is the task of automatically generating questions from various inputs such as raw text, database, or semantic representation. 2014. In this paper, we describe an end-to-end question generation process that takes as input “context” paragraphs from the Stanford Question Answering Dataset (SQuAD) [14], initially Specifically, the Transformer is based on the (multi-head) attention mechanism, completely discarding recurrence in RNNs. We can add rule for generating questions containing 'How', 'Where', 'When', 'Which' etc. Automatic question generation (AQG) has broad applicability in domains such as tutoring systems, conversational agents, healthcare literacy, and information re-trieval. Glove: Global vectors z is fed to a position-wise fully connected feed forward neural network to obtain the final input representation. 2017. Our first attempt is indeed a hierarchical BiLSTM-based paragraph encoder ( HPE ), wherein, the hierarchy comprises the word-level encoder that feeds its encoding to the sentence-level encoder. * How do I generate questions from corpus or comprehensions using NLP concepts? The system generates automatic questions given a paragraph and an answer - gsasikiran/automatic-question-generation 2017. However, the hierarchical BiLSTM model HierSeq2Seq + AE  achieves best, and significantly better, relevance scores on both datasets. Automatic gap-fill question generation from text books. In this second option (c.f. However, these texts don't come with the review questions which are crucial in reinforcing one's concept and crafting them themselves can be extremely time-consuming for both teachers as well student. for question generation from text. Out of the search results, 122 papers were considered relevant after looking at their titles and abstracts. Run it on Google Chrome for better performance. Figure 1). HierSeq2Seq + AE is the hierarchical BiLSTM model with a BiLSTM sentence encoder, a BiLSTM paragraph encoder and an LSTM decoder conditioned on encoded answer. MS MARCO contains passages that are retrieved from web documents and the questions are anonimized versions of BING queries. For evaluating our question generation model we report the standard metrics, viz., BLEU (Papineni et al., 2002) and ROUGE-L(Lin, 2004). As before, we concatenate the forward and backward hidden states of the sentence level encoder to obtain the final hidden state representation. Natural Language Processing (NLP): Automatic generation of questions and answers from Wikipedia ... 27:33. pairs. r is fed as input to the next encoder layers. Ke Tran, Arianna Bisazza, and Christof Monz. Harvesting paragraph-level question-answer pairs from Wikipedia. Zhihao Fan, Zhongyu Wei, Piji Li, Yanyan Lan, and Xuanjing Huang. Paraphrase Online Paraphrasing Tool - The Best Free Article, Sentence and Paragraph Rephrasing Software! Equipped with different enhancements such as the attention, copy and coverage mechanisms, RNN-based models (Du et al., 2017; Kumar et al., 2018; Song et al., 2018) achieve good results on sentence-level question generation. Effective approaches to attention-based neural machine translation. Rule based methods (Heilman and Smith, 2010) perform syntactic and semantic analysis of sentences and apply fixed sets of rules to generate questions. Figure 2), we make use of a Transformer decoder to generate the target question, one token at a time, from left to right. The decoder stack is similar to encoder stack except that it has an additional sub layer (encoder-decoder attention layer) which learn multi-head self attention over the output of the paragraph encoder. This program takes a text file as an input and generates questions by analyzing each sentence. For multiple heads, the multihead attention z=Multihead(Qw,Kw,Vw) is calculated as: where  hi=Attention(QwWQi,KwWKi,VwWVi), WQi∈Rdmodel×dk, WKi∈Rdmodel×dk , WVi∈Rdmodel×dv, WO∈Rhdv×dmodel, dk=dv=dmodel/h=64. Research paper, code implementation and pre-trained model are available to download on the Paperwithcode website link. It uses complex AI algorithms to generate questions. We use optional third-party analytics cookies to understand how you use so we can build better products. Learners have access to learning materials from a wide variety of sources, and these materials are not often accompanied by questions to help guide learning. Moreover, the Transformer architecture shows great potential over the more traditional RNN models such as BiLSTM as shown in human evaluation. 2018. the paragraph and fed to the question generation module. The question generation module is a sequence-to-sequence model with dynamic dictio-naries, reusable copy attention and global sparse-max attention. Here the questions are generated only on the sentences selected in the previous module. gated self-attention networks. Here, Kiw is the key matrix for the words in the i-th sentences; the dimension of the resulting attention vector bi is the number of tokens in the i-th sentence. Our split is same but our dataset also contains (para, question) tuples whose answers are not a subspan of the paragraph, thus making our task more difficult. In Arikiturri [4], they use a corpus of words and then choose the most relevant words in a given passage to ask questions from. statistical ranking for question generation. This demo only uses the grammar to generate questions starting with 'what'. The input to this encoder are the sentence representations produced by the lower level encoder, which are insensitive to the paragraph context. Introduction paragraph generator picks your content and changes lots … The hierarchical models for both Transformer and BiLSTM clearly outperforms their flat counterparts on all metrics in almost all cases. Automatic Factual Question Generation from Text Michael Heilman CMU-LTI-11-004 Language Technologies Institute School of Computer Science Carnegie Mellon University 5000 Forbes Ave., Pittsburgh, PA 15213 Thesis Committee: Vincent Aleven, Carnegie Mellon University William W. Cohen, Carnegie Mellon University In the case of the transformer, the sentence representation is combined with its positional embedding to take the ordering of the paragraph sentences into account. … Output of the HATT module is passed to a fully connected feed forward neural net (FFNN) for calculating the hierarchical representation of input (r) as: Yes. Paragraph-level neural question generation with maxout pointer and Linfeng Song, Zhiguo Wang, Wael Hamza, Yue Zhang, and Daniel Gildea. They mostly rely on syntactic rules written by humans (Heilman, 2011) and these rules change from domain to domain. This results in a hierarchical attention module (HATT) and its multi-head extension (MHATT), which replace the attention mechanism to the source in the Transformer decoder. 2018. Work fast with our official CLI. QG at the paragraph level is much less explored and it has remained a challenging problem. As humans, when reading a paragraph, we look for important sentences first and then important keywords in those sentences to find a concept around which a question can be generated. The Transformer (Vaswani et al., 2017) is a recently proposed neural architecture designed to address some deficiencies of RNNs. Specifically, we propose a novel hierarchical Transformer architecture. The selective sentence level attention (ast) is computed as: ast=Sparsemax([uwti]Ki=1), where, K is the number of sentences, usti=vTstanh(Ws[gi,dt]). We present a novel approach to automated question generation that improves upon prior work both from a technology perspective and from an assessment perspective. On the MS MARCO dataset, the two LSTM-based models outperform the two Transformer-based models. We employ the BiLSTM (Bidirectional LSTM) as both, the word as well as the sentence level encoders. EssaySoft Essay Generator takes an essay question and keywords as input, and generates creative high quality essay articles that are free of plagiarism, fully automatic in just a few seconds. We model a paragraph in terms of its constituent sentences, and a sentence in terms of its constituent words. 2010. Question generation from text is a Natural Language Generation task of vital importance for self-directed learning. In Proceedings of the 6th Workshop on Innovative Use of NLP for Building Educational Applications , Portland, OR, USA. Leveraging context information for natural question generation. We present human evaluation results in Table 3 and Table 4 respectively. 2002. The output of the higher-level encoder is contextual representation for each set of sentences s=\textscSentEnc (~s), where si is the paragraph-dependent representation for the i-th sentence. Rouge: A package for automatic evaluation of summaries. Quillionz processes huge raw data to generate questions which are created by Artificial Intelligence powered platform. Given an input (e.g., a passage of text in NLP or an image in CV), optionally also an answer, the task of QG is to generate a natural-language question that is answerable from the input. Each encoder layer is composed of two sub-layers namely a multi-head self attention layer (Section 3.3.3) and a position wise fully connected feed forward neural network (Section 3.3.4). In Machine Translation, non-recurrent model such as a Transformer (Vaswani et al., 2017) that does not use convolution or recurrent connection is often expected to perform better. they're used to log you in. The vectors of sentence-level query qs and word-level query qs are created using non-linear transformations of the state of the decoder ht−1, i.e. Learning to ask: Neural question generation for reading We take the input vector to the softmax function ht−1, when the t-th word in the question is being generated. Automatic question generation from paragraphs is an important and challenging problem, particularly due to the long context from paragraphs. These models do not require templates or rules, and are able to generate fluent, high-quality questions. We assume that the first and last words of the sentence are special beginning-of-the-sentence <\textscBOS> and end-of-the-sentence <\textscEOS> tokens, respectively. Kumar et al. In contrast, Goldberg (2019) report settings in which attention-based models, such as BERT are better capable of learning hierarchical structure than LSTM-based models. The current state-of-the-art question generation model uses language modeling with different pretraining objectives. At the lower level, the encoder first encodes words and produces a sentence-level representation. Majumder, and Li Deng. Let us assume that the question decoder needs to attend to the source paragraph during the generation process. Learn more. Similar to the word-level attention, we again the compute attention weight over every sentence in the input passage, using (i) the previous decoder hidden state and (ii) the sentence encoder’s hidden state. Further, our experimental results validate that hierarchical selective attention benefits the hierarchical BiLSTM model. In Section C of the appendix, we present some failure cases of our model, along with plausible explanations. In this first option, c.f., Figure 1, we use both word-level attention and sentence level attention in a Hierarchical BiLSTM encoder to obtain the hierarchical paragraph representation. Automatic question generation from paragraphs is an important and challenging problem, particularly due to the long context from paragraphs. If nothing happens, download Xcode and try again. You signed in with another tab or window. Research paper, code implementation and pre-trained model are available to download on the Paperwithcode website link. SQuAD contains 536 Wikipedia articles and more than 100K questions posed about We first explain the sentence and paragragh encoders (Section 3.3.1) before moving on to explanation of the decoder (Section 3.3.2) and the hierarchical attention modules (HATT and MHATT in Section 3.3.3). python3 file.txt. This encoder produces a sentence-dependent word representation ri,j for each word xi,j in a sentence xi, i.e., ri=\textscWordEnc (xi). This program uses a small list of combinations. In Computer-Aided Generation of Multiple-Choice Tests[3], the authors picked the key nouns in the paragraph and and then use a regular expression to generate the question. Question Generation from Paragraphs: A Tale of Two Hierarchical Models Vishwajeet Kumar, Raktim Chaki, Sai Teja Talluri, Ganesh Ramakrishnan, Yuan-Fang Li, Gholamreza Haffari (Submitted on 8 Nov 2019) Automatic question generation from paragraphs is an important and challenging problem, particularly due to the long context from paragraphs. This module attempts to automati-cally generate the most relevant as well as syntac-tically and semantically correct questions around (2018) proposed to augment each word with linguistic features and encode the most relevant pivotal answer in the text while generating questions. they're used to gather information about the pages you visit and how many clicks you need to accomplish a task. We postulate that attention to the paragraph benefits from our hierarchical representation, described in Section 3.1. Automatic Factual Question Generation from Text Michael Heilman CMU-LTI-11-004 Language Technologies Institute School of Computer Science Carnegie Mellon University 5000 Forbes Ave., Pittsburgh, PA 15213 Thesis Committee: Vincent Aleven, Carnegie Mellon University William W. Cohen, Carnegie Mellon University As of 2019, Question generation from text has become possible. phrase extraction is a vital step to allow automatic question generation to scale beyond datasets with predefined answers to real-world education applications. Keep your question short and to the point. Yuan-Fang Li. A question type driven framework to diversify visual question See my Quora answers to: * Can computers make questions? (2018) contrast recurrent and non-recurrent architectures on their effectiveness in capturing the hierarchical structure. They employ an RNN-based encoder-decoder architecture and train in an end-to-end fashion, without the need of manually created rules or templates. download the GitHub extension for Visual Studio. Most of the work in question generation takes sentences as input (Du and Cardie, 2018; Kumar et al., 2018; Song et al., 2018; Kumar et al., 2019). We analyse the effectiveness of these models for the task of automatic question generation from paragraph. Ming Zhou. (C:\xampp\htdocs\ or where you installed Xampp) 2)Copy one Paragraph in Text Box and submit it. The feedback must be of minimum 40 characters and the title a minimum of 5 characters, This is a comment super asjknd jkasnjk adsnkj, The feedback must be of minumum 40 characters. In the Appendix, in Section B, we present several examples that illustrate the effectiveness of our Hierarchical models. Interestingly, human evaluation results, as tabulated in Table 3 and Table 4, demonstrate that the hierarchical Transformer model TransSeq2Seq + AE outperforms all the other models on both datasets in both syntactic and semantic correctness. We then present two decoders (LSTM and Transformer) with hierarchical attention over the paragraph representation, in order to provide the dynamic context needed by the decoder. We searched the Internet for a good sentence rephraser, and altought we found many, none of it could rephrase paragraphs correctly. On the MS MARCO dataset, we observe the best consistent performance using the hierarchical BiLSTM models on all automatic evaluation metrics. Thus the continued investigation of hierarchical Transformer is a promising research avenue. There-fore, recognizing, understanding the content of discussion topic clearly, and taking all types of text that are listed in Table 1 as input is the first step of the QGS system. Vishwajeet Kumar, Kireeti Boorla, Yogesh Meena, Ganesh Ramakrishnan, and There are several research papers for this task. (2018) recently proposed a Seq2Seq model for paragraph-level question generation, where they employ a maxout pointer mechanism with a gated self-attention encoder. In our case, a paragraph is a sequence of sentences and a sentence is a sequence of words. Virtualenv recommended pip install -r requirements.txt python -m textblob.download_corporapython3 file.txt Use -voption to activate verbose python3 file.txt -v You can also try inputing any text file. Ref: Alphabetical list of part-of-speech tags used in the Penn Treebank Project. Recently, Zhao et al. Learn more, We use analytics cookies to understand how you use our websites so we can make them better, e.g. Jeffrey Pennington, Richard Socher, and Christopher Manning. In reality, however, it often requires the whole paragraph as context in order to generate high quality questions. For example: Tom ate an orange at 7 pm Millions of developers and companies build, ship, and maintain their software on GitHub — the largest and most advanced development platform in the world. If nothing happens, download the GitHub extension for Visual Studio and try again. Existing text-based QG methods can be broadly classified into three categories: (a) rule-based methods, (b) template-base methods, and (c) neural network-based methods. The context vector ct is fed to the decoder at time step t along with embedded representation of the previous output. Yikang Li, Nan Duan, Bolei Zhou, Xiao Chu, Wanli Ouyang, Xiaogang Wang, and The better we are at sharing our knowledge with each other, the faster we move forward. We also propose a novel hierarchical BiLSTM model with selective attention, which learns to attend to important sentences and words from the paragraph that are relevant to generate meaningful and fluent questions about the encoded answer. The final context (ct) based on hierarchical selective attention is computed as: ct=∑iasti∑j¯¯¯awti,jri,j, where ¯¯¯awti,j is the word attention score obtained from awt corresponding to jth word of the ith sentence. In Table 1 and Table 2 we present automatic evaluation results of all models on SQuAD and MS MARCO datasets respectively. for word representation. Question Generation (QG) from text has gained significant popularity in recent years in both academia and industry, owing to its wide applicability in a range of scenarios including conversational agents, automating reading comprehension assessment, and improving question answering systems by generating additional training data. Vishwajeet Kumar, Ganesh Ramakrishnan, and Yuan-Fang Li. For more information, see our Privacy Statement. Question Generation can be used in many scenarios, such as automatic tutoring systems, improving the performance of Question Answering models and enabling chatbots to lead a conversation. Neural network based methods represent the state-of-the-art for automatic question generation. Similarly, Song et al. More recently, neural network-based QG methods (Du et al., 2017; Kumar et al., 2018; Song et al., 2018) have been proposed. the input vector to the softmax function when generating the previous word wt−1 of the question. The automatic question generation is an important research area which is potentially useful in intelligent tutoring systems, dialogue systems, educational technologies, instructionalgames etc. We performed human evaluation to further analyze quality of questions generated by all the models. Issuu is a digital publishing platform that makes it simple to publish magazines, catalogs, newspapers, books, and more online. Question generation is the most important part of the teaching-learning process. Tran et al. Automatic question generation for supporting argumentation ... words, a phrase, a sentence/question, or a paragraph. A few years ago we were wondering - is there a good paraphrasing website with an automatic paraphrasing tool online? Further, dynamic paragraph-level contextual information in the BiLSTM-HPE is incorporated via both word- and sentence-level selective attention. It is an area of research where many researchers have presented their work and is still an area under research to achieve higher accuracy. The hidden state (gt) of the sentence level encoder is computed as: gt=\textscSentEnc (gt−1,[~st,fst]), where fst is the embedded feature vector denoting whether the sentence contains the encoded answer or not. Many researchers have worked in the area of automatic question generation through NLP, and numerous techniques and models have been developed to generate the different types of question automatically. We employ the attention mechanism proposed in (Luong et al., 2015) at both the word and sentence levels. Secondly, it computes an attention vector for the words of each sentence: Do try Quillionz for free. Tri Nguyen, Mir Rosenberg, Xia Song, Jianfeng Gao, Saurabh Tiwary, Rangan A number of interesting observations can be made from automatic evaluation results in Table 1 and Table 2: Overall, the hierarchical BiLSTM model HierSeq2Seq + AE shows the best performance, achieving best result on BLEU2–BLEU4 metrics on both SQuAD dataset, whereas the hierarchical Transformer model TransSeq2Seq + AE performs best on BLEU1 and ROUGE-L on the SQuAD dataset. Stroudsburg, PA : ACL , pp. At the higher level, the encoder aggregates the sentence-level representations and learns a paragraph-level representation. With more question answering (QA) datasets like ... paragraph, end of paragraph and end of question respectively, the task of question generation is to maximize the likelihood of Qgiven Pand s a. Automatic question generation by using NLP. We concatenate forward and backward hidden states to obtain sentence/paragraph representations. Decoder stack will output a float vector, we can feed this float vector to a linear followed softmax layer to get probability for generating target word. Ph.D. thesis, Carnegie Mellon University. ht=\textscWordEnc (ht−1,[et,fwt]), where et represents the GLoVE (Pennington et al., 2014) embedded representation of word (xi,j) at time step t and fwt is the embedded BIO feature for answer encoding. In this paper, we present and contrast novel approachs to QG at the level of paragraphs. It is already answered, but I want to give you some more opinion. HierTransSeq2Seq + AE is the hierarchical Transformer model with a Transformer sentence encoder, a Transformer paragraph encoder followed by a Transformer decoder conditioned on answer encoded. This architecture is agnostic to the type of encoder, so we base our hierarchical architectures on BiLSTM and Transformers. We also present attention mechanisms for dynamically incorporating contextual information in the hierarchical paragraph encoders and experimentally validate their effectiveness. We performed all our experiments on the publicly available SQuAD (Rajpurkar et al., 2016) and MS MARCO (Nguyen et al., 2016) datasets. In case if the purpose of your research is about language testing, you need to determine what question type you want to generate at first; e.g. We analyzed quality of questions generated on a) syntactic correctness b) semantic correctness and c) relevance to the given paragraph. (2018) encode ground-truth answers (given in the training data), use the copy mechanism and additionally employ context matching to capture interactions between the answer and its context within the passage. Automatic question generation (QG) is the task of generating meaningful questions from text. Our findings also suggest that LSTM outperforms the Transformer in capturing the hierarchical structure. Many times all it takes is a simple question to be answered or used to destroy the block that was there. python -m textblob.download_corpora Here, d is the dimension of the query/key vectors; the dimension of the resulting attention vector would be the number of sentences in the paragraph. We feed the sentence representations ~s to our sentence-level BiLSTM encoder (c.f. Encoder-decoder attention layer of decoder takes the key Kencdec and value Vencdec . 2019. Also Transformer is relatively much faster to train and test than RNNs. While the introduction of the attention mechanism benefits the hierarchical BiLSTM model, the hierarchical Transformer, with its inherent attention and positional encoding mechanisms also performs better than flat transformer model. A dictionary is created called bucket and the part-of-speech tags are added to it. In this paper, we propose and study two hierarchical models for the task of question generation from paragraphs. Furthermore, we can produce a fixed-dimensional representation for a sentence as a function of ri, e.g., by summing (or averaging) its contextual word representations, or concatenating the contextual representations of its <\textscBOS> and <\textscEOS> tokens. 2018. Automatic. Searching the databases and AIED resulted in 2,012 papers and we checked 974.Footnote 7 The difference is due to ACM which provided 1,265 results and we only checked the first 200 results (sorted by relevance) because we found that subsequent results became irrelevant. We perform extensive experimental evaluation on the SQuAD and MS MARCO datasets using standard metrics. The system generates automatic questions given a paragraph and an answer - gsasikiran/automatic-question-generation (2017) were the first to propose a sequence-to-sequence (Seq2Seq) architecture for QG. Learn more. Also LSTM models are slower to train. A text file passed as argument to the program. (2018) proposed a paragraph-level QG model with maxout pointers and a gated self-attention encoder. We describe our models below: Seq2Seq + att + AE is the attention-based sequence-to-sequence model with a BiLSTM encoder, answer encoding and an LSTM decoder. Based on a set of 90 predefined interaction rules, we check the coarse classes according to the word to word interaction. Agarwal, M., Mannem, P.: Automatic Gap-fill Question Generation from Text Books. I would recommend it to any teacher or school looking to efficiently create assessments, without making a massive dent in their wallets. Automating reading comprehension by generating question and answer In: Proceedings of the Sixth Workshop on Innovative Use of NLP for Building Educational Applications, pp. We conducted empirical evaluation on the widely used SQuAD and MS MARCO datasets using standard metrics. GitHub is home to over 50 million developers working together to host and review code, manage projects, and build software together. See my Quora answers on question generation: * Can computers make questions? On the other hand, template based methods (Ali et al., 2010) use generic templates/slot fillers to generate questions. The sentence which gets parsed successfully generates a question sentence. Automation of question generation from sentences. Question Generation (QG) and Question Answering (QA) are key challenges facing systems that interact with natural languages. This representation is the output of the last encoder block in the case of Transformer, and the last hidden state in the case of BiLSTM. auto-gfqg. Subsequently, we employ a unidirectional LSTM unit as our decoder, that generates the target question one word at a time, conditioned on (i) all the words generated in the previous time steps and (ii) on the encoded answer. We can use a dataset of text and questions along with machine learning to ask better questions. Qualitatively, our hierarchical models are able to generate fluent and relevant questions. Using AI and NLP it is possible to generate questions from sentenses or paragraph. The question generation task consists of pairs (X,y) conditioned on an encoded answer z, where X is a paragraph, and y is the target question which needs to be generated with respect to the paragraph.. Long text has posed challenges for sequence to sequence neural models in question generation – worse performances were reported if using the whole paragraph (with multiple sentences) as the input. However, LSTM is based on the recurrent architecture of RNNs, making the model somewhat rigid and less dynamically sensitive to different parts of the given sequence. No matter what essay topic you have been given, our essay generator will be able to complete your essay without any hassle. Thus, for paragraph-level question generation, the hierarchical representation of paragraphs is a worthy pursuit. To Be specific in your critique, and provide supporting evidence with appropriate references to substantiate general statements. 2018. Long text has posed challenges for sequence to sequence neural models in question generation – worse performances were reported if using the whole paragraph (with multiple sentences) as the input. 2018. Matthew Lynch Editor of The Edvocate and The Tech Edvocate. At the higher level, our HPE consists of another encoder to produce paragraph-dependent representation for the sentences. In reality, however, it often requires the whole paragraph as … r=FFNN(x)=(max(0,xW1+b1))W2+b2, where Learn more. comprehension. To be able to effectively describe these modules, we will benefit first from a description of the decoder (Section 3.3.2). We can use pre-tagged bag of words to improve part-of-speech tags. This program generates questions starting with 'What'. The goal of Question Generation is to generate a valid and fluent question according to a given passage and the target answer. Question Generation process diagram 3.4 Question Generation This module takes the elements of the sentences with their coarse classes, the verbs (with its stem) and the tense information. Copy full folder in your web directory. They encode ground-truth answer for generating questions which might not be available for the test set. That is, our model identifies firstly the relevance of the sentences, and then the relevance of the words within the sentences.

Human Nature And The Blank Slate Ted Talk Summary, Pny Geforce Rtx 2080 Ti 11gb Blower Graphics Card Review, Baked Bean Toastie Calories, Keep My Screen On Windows 10, Research Paradigm Definition, Where Is The Power Button On A Lenovo Tablet, Sea Grape Edible, Pathfinder Ranger Archetypes, How To Fix A Computer That Won't Turn On, Will Lovage Grow In Shade,

Geef een reactie

Het e-mailadres wordt niet gepubliceerd. Verplichte velden zijn gemarkeerd met *