text stringlengths 41 31.4k |
|---|
<s>/FuturaBT-Medium /FuturaBT-MediumItalic /Futura-Light /Futura-LightOblique /GalliardITCbyBT-Bold /GalliardITCbyBT-BoldItalic /GalliardITCbyBT-Italic /GalliardITCbyBT-Roman /Garamond /Garamond-Bold /Garamond-BoldCondensed /Garamond-BoldCondensedItalic /Garamond-BoldItalic /Garamond-BookCondensed /Garamond-BookCondens... |
<s>/TimesNewRomanPSMT /Times-Roman /Times-RomanSC /Trajan-Bold /Trebuchet-BoldItalic /TrebuchetMS /TrebuchetMS-Bold /TrebuchetMS-Italic /Tunga-Regular /TwCenMT-Bold /TwCenMT-BoldItalic /TwCenMT-Condensed /TwCenMT-CondensedBold /TwCenMT-CondensedExtraBold /TwCenMT-CondensedMedium /TwCenMT-Italic /TwCenMT-Regular /Univer... |
<s>[600 600] /PageSize [612.000 792.000]>> setpagedevice</s> |
<s>See discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/336129284Assessment of Bangla Descriptive Answer Script DigitallyPreprint · September 2019DOI: 10.13140/RG.2.2.12203.39204CITATIONSREADS1835 authors, including:Some of the authors of this publication are als... |
<s>II dis-cusses previous works. Methodology discussed in Section III.Result analysis and discussion demonstrated in Section IV.And finally Section V refers the conclusion.II. PREVIOUS WORKSGenerally, answer scripts evaluation is a difficult task. Itbecomes harder for Bangla language. Various works are foundfor English... |
<s>ofFig 2 are taken in this step.Fig. 2. Sub-steps in the Keyword Generation1) Preprocessing: As answers collected from various do-mains and students contain many unnecessary words, stopsymbols, punctuation symbols, emotions symbols etc. Theseare noise for our unprocessed data. We remove these specialcharacters like #... |
<s>spelling checking is done in many proposed method. Butas we are working with Bangla language in which case suchresources are limited and we are using Akkhor Bangla Spelland Grammar Checker [12] in our proposed method, whichone is a tool to check spelling and grammatical mistakes.Hence we are proposing a simple keywo... |
<s>in fig 3 which is the graph of QuestionsVs Relative error.As resources in Bangla Language are hard to find and thereis no work found in past our work is just an initiative. But therelative error which is observed in our system can be acceptedas it is just the beginning. We hope to improve our work inthe future. In t... |
<s>See discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/331158767Question Bank Similarity Searching System (QB3S) Using NLP and InformationRetrieval TechniqueConference Paper · May 2019DOI: 10.1109/ICASERT.2019.8934449CITATIONSREADS4082 authors:Some of the author... |
<s>is represented in tabularform, need to strategically better searching techniques.We have used NLP tools [3] like, Tokenization or lexicalanalysis of bangla text corpus [4], StopWord removing,Rule-based sementic analysis or stemming to modify inflectedword and analyzed the structured data to build a dynamic B+tree cl... |
<s>course ofsecondary and higher secondary level of Bangladesh. Prob-lem based e-learning system (http://epbl.org) for interactivelearning of ICT course, practicing and model examinationfrom question bank and C programming, HTML, SQL, ERDlearning based on higher secondary book have already beenimplemented successfully.... |
<s>algorithm performs a linearsearch in every node read from the disk. The time complexityof each linear search is O(t) . Thus, the total time complexityof the B + tree search operation is O(tlogtn) and same timecomplexity for inserting and deleting [18].In case of higher number of chapter, it needs higher orderand for... |
<s>Number of documents with that term t in it.IDF (t) = ln(2)Where, N = Total number of documents And n = Number ofdocuments with term tWe have been developed an improved version of thisalgorithm where we used some pre-defined weighting factorbased on the classification of question structure.Weighting Factor,Where, w1>... |
<s>used as a measure of divergencebetween the vectors, and cosine of the angle is used as thenumeric similarity [23]. Cosine has the nice property that itis 1.0 for identical vectors and 0.0 for orthogonal vector. If~D is the document vector and ~Q is the query vector, then thesimilarity of document D to query Q (or sc... |
<s>techniques arenot efficient in finding the similarity of questions in a QuestionBank developed for PBeL systems for many reasons: i) thedocument structure is different in nature that the conventionaltext documents, and ii) the contents of the different parts ofthe documents have different weights.In this paper, we h... |
<s>, pp. 537542, 2013.[16] A. S. M. L. Hoque, G. M. M. Bashir, and M. R. Uddin, Equivalenceof problems in problem based e-learning of database, in Technology forEducation (T4E), 2014 IEEE Sixth International Conference on , pp.106109, IEEE, 2014.[17] G. M. M. Bashir and A. S. M. L. Hoque, An effective learning andteach... |
<s>Microsoft Word - 16nov QA paper.docxSee discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/329972144Improving Answer Extraction For Bangali Q/A System Using Anaphora-Cataphora ResolutionConference Paper · December 2018DOI: 10.1109/CIET.2018.8660888CITATIONSREADS... |
<s>a word net for the exact answer prediction by semantic matching with anaphora-cataphora resolution. Word net is referred as a lexical database which contains words with their synonyms and relational words, noun, adjective, verbs with grouping [2]. There are few works done for Bangla QA system. In Bangla, there are s... |
<s>the question words. The flowchart of our system is shown in below: Fig. 1. Flow chart of our proposed QA system In this system, a big problem was faced. For example- 1st sentence: । 2nd sentence: । Question: According to our system, answer would be— “” which is not expected. It is needed to replace “ ” with “ ”. Thi... |
<s>As it contains “ , it gets lower priority. Details can be found in Syntactical rules. Total 3 TABLE II. “ ” TAG: Words Score Reason 2 Single word 2 Single word Total 4 We can see that, the total score or priority of “ ” tag is 3 and “ ” tag is 4. So, we’ll consider the upper priority. Here, the tag of “ ” is the upp... |
<s>is in singular form. So, it indicates “ ”. For example, 1st sentence: ।2nd sentence: । Here, “ ” and “ ” are two nouns from the first sentence and “ ” and “ ” are pronouns from second sentence. There is a “ ” with “ ”. So, in this sentence, “ ” is somehow connected with “ ”. “ ” is the noun who is taking action. Thi... |
<s>(DMA) can do this type of processing. It also needs some basic knowledge. Just like, 1) If someone do something bad, he will feel guilt later. 2) If a person thanks someone, he will welcome him. 3) Fast food is dangerous for health. Etc. Semantic Memory (SM) can be used to save and use the basic knowledge or static ... |
<s>bengali." Proceedings of the 51st Annual Meeting of the Association for Computational Linguistics (Volume 2: Short Papers). Vol. 2. 2013. [14] Tazakka, Tazbeea, Md Asifuzzaman, and Sabir Ismail. "Anaphora Resolution in Bangla Language." International Journal of Computer Applications 154.9 (2016). [15] Sanjay Chatter... |
<s>Bangla Intelligence Question Answering System Based on Mathematics and Statistics2019 22nd International Conference on Computer and Information Technology (ICCIT), 18-20 December 2019 978-1-7281-5842-6/19/$31.00 ©2019 IEEE Bangla Intelligence Question Answering System Based on Mathematics and Statistics Md. Kowsher ... |
<s>Cosine similarity, Jaccard similarity and Naïve-Bayes. These methods aid to establish the relations between users‟ questions and information. The contributions are summarized as follows: • We have introduced mathematical and statistical procedures for BIQAS for information retrieval. • For the pre-processing of data... |
<s>[11]. Cai Dongfeng and Cui Huan discovered a web-based Chinese Automatic Question Answering System which uses Google Web services API [12]. Liu Hongshen, Qin Feng, Chen Xiaoping, Tao Tao, et al. proposed a teaching mode software which can keep the attendance of the students and also can keep the answer of any studen... |
<s>algorithms are urged to obtain the relationship between the questions and answers. But Cosine Similarity deals with vectors. In this case, the documents and questions transmit to vectors using the TF-IDF model. In order to minimize the execution time and space complexity, we have used SVD techniques. Authorized lice... |
<s>the unknown words, we have established a corpus of the suffix of Bengali language; for instances: (te), (che), (yer), etc. The longest common suffix has been removed from the last position of an unknown word. Thus we obtain the lemma or root of an unknown word. Synonym words indicate the exact or nearly the same mea... |
<s>to remove the stop words from the inserted documents and questions. We have compiled 74 topics as informative documents like as hall information, department information, teacher information, library, NSTU nature, bus schedules etc. of Noakhali Science and Technology University (NSTU) that is our third corpus as ques... |
<s>An advance it can be shaped with the assist of Deep Learning algorithms such as Recurrent Neural Network (RNN) by processing the BNLP. REFERENCES [1] Clark, Peter, et al. "Think you have solved question answering? try arc, the ai2 reasoning challenge." arXiv preprint arXiv:1803.05457 (2018). [2] Asiaee, Amir Hosein.... |
<s>Design and Development of Question Answering System in Bangla Language from Multiple Documents1st International Conference on Advances in Science, Engineering and Robotics Technology 2019 (ICASERT 2019) 978-1-7281-3445-1/19/$31.00 ©2019 IEEE Design and Development of Question Answering System in Bangla Language from... |
<s>extraction from question, N-grams formation from keywords for approximate matching, retrieving n-best answers and generating specific answers by question type and evaluation of performance and correctness. Details of the steps are following: A. Removing Stop Words In the proposed system necessary intelligence about ... |
<s>from wikipedia(Bangla) and several questions given below have been set from these documents. Question 1: “কক্সবাজার থানা প্রথম কদব প্রবতষ্ঠিত হয়? (When Cox’s Bazar Thana has been first established?)” Answer 1: “১৮৫৪ োল” (“1854 year”) Answer 2: “োল এবাং নপৌেভা” (“Year and Municipality”) Question 2: “কম্পিউটাদর ... |
<s>কদব আববষৃ্কত হয়? 1 2 1 0.5 1 গণকযি বক? 1 1 1 1 1 মাইদরাপ্রদেের উদ্ভাবক নকান প্রবতিান? 0 5 1 0 0 কম্পিউটাদর প্রথম বাাংলা নলখা েম্ভব হয় কখন? 1 2 1 0.5 1 বাাংলা ওয়াডবপ্রদেবোং েফটওয়যার উদ্ভাবন কদর কারা? 1 1 1 1 1 মাইদরােফট উইদডাজ’ এর েদে বযবহাদরর জনয ইন্টারদফে ‘ববজয়’ কদব উদ্ভাববত হয়? 1 2 1 0.5 1 Average: 0.5 0.69 In Ta... |
<s>3rd Workshop on South and Southeast Asian Natural Language Processing, 2012. [5] P. Pakray, P. Bhaskar, S. Banerjee, B. C. Pal, S. Bandyopadhyay, A. Gelbukh, "A Hybrid Question Answering System based on Information Retrieval,” CLEF (Notebook Papers/Labs/Workshop), 2011. [6] L. Zhenqiu, "Design of automatic question ... |
<s>Galley Proof 12/04/2019; 14:42 File: idt–1-idt180074.tex; BOKCTP/duyan p. 1Intelligent Decision Technologies -1 (2019) 1–16 1DOI 10.3233/IDT-180074IOS PressA novel Bengali Language Query ProcessingSystem (BLQPS) in medical domainKailash Pati Mandala,∗, Prasenjit Mukherjeea, Baisakhi Chakrabortya and Atanu Chattopadh... |
<s>36guages. Bengali language has been used in important 37states of India such as West Bengal, Tripura, Assam 38and Andaman Nicobar Islands. The national language 39of Bangladesh is also Bengali. The proposed system 40ISSN 1872-4981/19/$35.00 © 2019 – IOS Press and the authors. All rights reservedrrected proof version... |
<s>set of Prolog programming logic. Finally 99the answer is fetched by executing the Prolog logic 100as discussed in [4]. The EFLEX system is an effi- 101cient database interface system that consists of ana- 102lyzer, mapper and translator. The analyzer interprets 103the given natural language query for the mapper. The... |
<s>for knowledge modeling which in-155creases the correct responses of user’s query that is ex-156plained in [10]. In [11] deals with structural ambiguity157of a Bengali sentence. The given sentence is tokenized158by the Tokenizer. Then the validator checks grammati-159cal mistake in the sentence. Whereas the En-conver... |
<s>out 217POS type andBengali meaning of the token. TheGram- 218matical Rule Generator contains a set of pre-defined 219production which useful to form a correct parse tree. 220The Conversion Unit converts the English parse tree 221into corresponding Bengali parse tree. Finally Ben- 222gali sentence prints. This system... |
<s>আেছ?281(nothyea o purulyeki ki haspathal ache?)282That means in English Language283What are hospitals available in Nadia and Purulia?2843.1.2. Query tokenization285The proposed system reads the query and slices286it into meaningful linguistic units called tokens after287Fig. 1. Data flow diagram of BLQPS.Fig. 2. Nat... |
<s>tagging. 336The BLQPS selects every token from the tokenized 337query and compares with each word of interrogative’s 338word list. If token matches with any word of interroga- 339tive word’s list then the score of the selected token will 340be same as score of the interrogative’s word list i.e. 2. 341Otherwise token... |
<s>Department িবভাগ (Department) dept_id hos_id NULL8 Department শাখা (Department) dept_id hos_id NULL9 Department অংশ (Department) dept_id hos_id NULL10 Department দ র (Department) dept_id hos_id NULL11 Doctor ডা ার (Doctor) doc_id hos_id dept_id12 Doctor িচিক�সক (Doctor) doc_id hos_id dept_id13 Doctor বদ (Doctor) doc... |
<s>department dept_name িবভাগ নাম (Department’s Name) dept_id hos_idTable 7Hospital table of default databaseHos_id Hos_name Hos_add Hos_district Hos_state1 মুিশদাবাদ জলা হাসপাতাল (Murshidabad District Hospital) লালেগালা (Lalgola) মুিশদাবাদ (Murshidabad) পি মব (West Bengal)2 নিদয়া জলা হাসপাতাল (Nadia District Hospital)... |
<s>entity 3 (table 3)… entity n (table n)WHERE475condition 1 and condition 2 and condition 3… and con-476dition n. SELECT, FROM and WHERE clauses are477fixed in the structure of a SQL query for retrieving data,478that why the proposed system has to find attribute(s),479entities and condition(s). There are several cases... |
<s>table, entity_id 545field is the primary key. Entity_name field contains 546participating entity name corresponding to table names 547of default database like hospital, doctor and depart- 548ment. The synonyms field contains all possible syn- 549onyms of the entities in Bengali language. The pri- 550mary_key field c... |
<s>in Ben- 604gali transcript has followed IPA notation as per Help: 605IPA/Bengali given in website https://en.wikipedia.org/ 606wiki/Help:IPA/Bengali 6074.1. Step i 608The Bengali Language Query Processing System 609(BLQPS) is a domain specific natural language query 610processing system. The system has been designed... |
<s>and Nadia’s Hospital?), has been tokenized and666shown in Table 10. After POS tagging the BLQPSmaps667known and unknown token or word by considering in-668dex of query string array(token array) and score ar-669ray as well as score value of the score array. The non-670consecutive unknown word generates only one patte... |
<s>table.730vi. নিদয়ার হাসপাতােলর (nothyar haspathaler) – This731will not be selected.732vii. অি (osthi) (Orthopedic) – This will not be se-733lected.734viii. িচিক�সা (chykitsa) (Treatment) – This will not be735selected.736ix. িবভাগ (bybhag) (Department) – This will be se-737lected as entity from enttities table and co... |
<s>phase. 805After POS tagging known and unknown tokens 806have been given in Fig. 6. 807Above example discussed in Section 3(iii) has 808been taken in the Fig. 6. 809iii. If token taken one at a time= n number of pattern 810will be generated. 811rrected proof versionGalley Proof 12/04/2019; 14:42 File: idt–1-idt180074... |
<s>BLQPS does not contain such type ofData Dictionary.iii) The time complexity has been computed here.7 M.M. Anwar et al., Syn-tax Analysis and Mach-ine Translation of BanglaSentences [13]i) Syntax Analysis and Machine Translation basedsystem works on pre-defined grammar rules.ii) This system uses trained corpus to ide... |
<s>zw + st}∴ f(n) ∼= n+ n× n× n+n(n+ 1)n(n+ 1){n× n+ n× n+ n× n}= n+ n3 +n(n+ 1){1 + n2 + n2 + n2}= n+ n3 +n(n+ 1){3n2 + 1}= n+ n3 +(n2 + n){3n2 + 1}= n+ n3 +3n43n33n45n3(3n4 + 5n3 + n2 + 3n) = O(n4)5. General features based comparative study of842the proposed system with similar type system843The comparative study of ... |
<s>processed as best is an adjec-894tive. Hence, queries with qualitative terms defined by895adjectives cannot be processed which is a limitation of896the BLQPS.897From the time complexity analysis it is found that898time complexity is in O(n4) which works well when899the number of unknown words/tokens are few. How-900... |
<s>of Bangla sentences. International Journal 969of Computer Science and Network Security 2009; 9(8): 317- 970326. 971[14] MahmudMR,AfrinM, RazzaqueMA, et al. A rule based Ben- 972gali stemmer. International Conference on Advances in Com- 973puting, Communications and Informatics 2014. p. 2750-2756. 974doi: 10.1109/ICA... |
<s>GESJ: Computer Science and Telecommunications 2018|No.2(54) ISSN 1512-1232 AUTOMATIC QUESTION AND ANSWER GENERATION FROM BENGALI AND ENGLISH TEXTS Shudipta Sharma1, Muhammad Kamal Hossen2, Md. Sajjatul Islam3, Md. Shahnur Azad Chowdhury4, Md. Jiabul Hoque5 1,2Chittagong University of Engineering & Technology, 3Chitt... |
<s>great eye on the field of Natural Language Processing (NLP). Someone do researches on text processing, others on text understanding, text summarizing, finding answers for given questions, and generating questions, etc. All of these tasks are a great deal with NLP. Some of the works that are done earlier on question ... |
<s>system; however, it gives a basis to implement a complete QA system for Bengali. The implementation was involved with the generation of questions from the medical domain only. They also considered only simple questions (‘Wh’ questions). Pakray et al. [9] developed a keyword based multilingual restricted domain quest... |
<s>parts such as “We, the students of your school, are fully responsible for this”. This is extracted to “We are fully responsible for this”. Three subtasks are followed to accomplish the goal. They are: (i) Removal of Stop Words and Relative Clauses: The words whose absences don’t make any significant change in the se... |
<s>for the selected answer phrase, (ii) Main verb decomposition, (iii) Subject-auxiliary inversion, and (iv) Replacement of answer phrase with question phrase and placing at the beginning of the sentence. These steps are shown in Figure 4. GESJ: Computer Science and Telecommunications 2018|No.2(54) ISSN 1512-1232 Figur... |
<s>to replace the answer phrase with the proper question phrase. The step by step process of generating questions from Bengali text is illustrated in Figure 7 with an example. Figure 7: Steps of Question Generation from Bengali Sentences 3.3.1 Simple Sentence Extraction First, the complex and compound sentences are tra... |
<s>using some Bengali and English texts as input. A list of some of these texts is given in Table 3. Table 3. Input Texts in Bengali and English Title of the Text Contents Language Cricket Team ২০১৫ সাল বাংলােদশ ি�েকেটর �ণর যুগ িছল। িক� ২০১৭ সােল ছ� হািরেয় েফেল। মাশরািফ িবন মুতর ্ার েনতৃে� বাংলােদেশর েপস েবািলং আ�মেণ ন... |
<s>The organization is financed by assessed and voluntary contributions from its member states. English 4.1 Question Generation Task The number of generated questions by human and the proposed system for the texts listed in Table 3 is provided in Table 4. On the basis of the number of generated questions, we have evalu... |
<s>of the Generated Questions from the ‘Flood’ Texts Questions Generated by the QA System Correctness কত সােলর বনযায় েদেশর ১৬ি ে্লার ১৮৪০০ েহ�র ্িমর ফসল ন� হয়? Ok এখােন বািষরক বৃি�পােতর পিরমাণ কত িমিলিমটার? No বাংলােদশ িক? Ok কতি আআ র্ ািতক নদনসহ ৭০০ি নদন এ েদেশ ্ােলর মেতা িব�ার কের আেছ? No কতি নদনর উউসসল ভারেত অবিসত? ... |
<s>by the existing QA systems. The performance and accuracy of the system are evaluated by inputting different texts both in Bengali and English. We are able to increase the performance of the system up to 78.33% and the percentage of accuracy of the generated questions is 76% on average. The performance and accuracy o... |
<s>International Conference on Bangla Speech and Language Processing(ICBSLP), 27-28 September, 2019Bengali Question Answering System for FactoidQuestions: A statistical approachSourav SarkerComputer Science and EngineeringShahjalal University ofScience and TechnologySylhet-3114, Bangladeshsourav39@student.sust.eduSyeda... |
<s>some aregeneralized.AnswerBus is an open-domain question answeringsystem where the information related to the answer isretrieved from the web in sentence level. The authors usedfive search engines (Google, Yahoo, Altavista, WiseNutand Yahoo News) to extract the web documents whichcontain the answers to the questions... |
<s>built our knowledge base based onthe website which carries the information solely aboutShahjalal University of Science & Technology (SUST) likewww.sust.edu, en.wikipedia.org/wiki/Shahjalal_University_of_Science_and_Technology etc. andnews from different web portals. We also collected some ofthe paragraphs about SUST... |
<s>an answer.We wanted to build a question answering system forBengali language on a closed domain and not having enoughresources to build it, we had to work from the scratch. Wehave divided our methodology to answer a factoid questionin four basic parts:1) Data Preparation2) Question Classification3) Document Categori... |
<s>formula:Accuracy =TP + TNP +Nwhere,P = total number of positive samplesN = total number of negative samplesTP (True Positive) = number of cases when the actualclass of the sample is true and also predicted as trueTN (True Negative) = number of cases when the actualclass of the sample is false and also predicted as f... |
<s>J. Ko, J. Betteridge, M. W.Bilotti, A. H. Schlaikjer, and E. Nyberg, “Javelin iii: Cross-lingualquestion answering from japanese and chinese documents.” in NTCIR,2007.[5] B. F. Green Jr, A. K. Wolf, C. Chomsky, and K. Laughery, “Baseball:an automatic question-answerer,” in Papers presented at the May 9-11, 1961, wes... |
<s>See discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/340351877IMPLEMENTATION OF AN AUTOMATIC QUESTION ANSWERING SYSTEM USINGMACHINE LEARNING SAKIF AHMED ABIRPreprint · September 2019DOI: 10.13140/RG.2.2.13210.39368CITATIONSREADS1931 author:Sakif Ahmed AbirUniv... |
<s>sincere gratitude to my project supervisors, Khan Raqib Mahmud Sir and Dr. Abul Kalam Al Azad sir for giving me the opportunity to work with them in this project, I was able to overcome all the obstacles that I encountered in these enduring three months of my project. In fact, they always gave me immense hope every ... |
<s>2.3.5 Sequence to Sequence Architecture………………………………….. 12 2.3.6 Chat-Bot for collage Management System Using A.I………………... 12 2.3.7 Product-Aware Answer Generation………………………………….. 13 vii | P a g e 3. BACKGROUND………………………………………………………… 14 3.1 Natural Language Processing (NLP)……………………………………. 14 3.2 Machine Learning – ML……... |
<s>Snapshot of the train.py file……………………………………………………. 69 Figure B.1: Snapshot of the Database………………………………………………..………. 70 Figure B.2: Snapshot of the Database………………………………………………..………. 71 Figure B.3: Snapshot of the Database………………………………………………..………. 71 Figure B.4: Snapshot of the Database………………………………………………..………. 72 Figure ... |
<s>automatic question answering system has been such accepted issue due to increasing number of websites, online shopping sites and telecommunication services. To reduce the human labour and flexible along with quick response privilege different type of approaches have been taken to develop a system for automatic quest... |
<s>THESIS CONTRIBUTION For better understanding the internal working process and different functions of Machine Learning is to be needed as an active area of research. As our main focus is to create a system for Answering Questions from user automatically by training. So, we will contribute through the thesis that we w... |
<s>aims at selecting answer spans that a) match the expected answer type (a time as indicated by “When”) and b) are close to important question words [26]. The actual answer “1688-1692” would easily be extracted by such a heuristic. Early work on chatbots [21] relied on handcrafted templates or heuristic rules to do re... |
<s>is the combination of all the fields of Natural Language Processing (NLP). With the advent of smart personal assistants like Siri, Google Chrome, and Cortana [25], we may hope for the fulfillment of the dream of Colby. “Before there were computers, we could distinguish persons from non-persons on the basis of an abi... |
<s>chatbots are those which can reply to a limited number of subjects. A very good example would be goal based chatbot. An open domain chatbot does not have any knowledge base so it has to generate new sentence for each interaction. Since it has no goal so the users can take the conversation to anywhere. Often unrelate... |
<s>cloze-style machine comprehension tasks [19]. 11 | P a g e 2.3.3 E-Commerce Question-Answering In recent years, product aware question answering has received several attention. Most of existing strategies aim at extracting relevant sentences from input text to answer the given question propose a framework for opinio... |
<s>text, key-value memory network (KVMN) is purposed to store this type of data. He et al. incorporate copying and retrieving knowledge from knowledge base stored in KVMN to generate natural answers within an encoder-decoder framework. This uses a KVMN to store the translate history which gives model the opportunity to... |
<s>advantages and limitations. Also we will see the general work flow diagram. 3.1 NATURAL LANGUAGE PROCESSING - NLP Natural Language Processing (NLP) is a research area of Artificial Intelligence (AI) which focus in the study and development of systems that allows the communication between a person and a machine throu... |
<s>as used for many other types of classifiers. In the statistics and computer science literature, naive Bayes models are known under a variety of names, including simple Bayes and independence Bayes. All these names reference the use of Bayes' theorem in the classifier's decision rule, but naïve Bayes is not (necessar... |
<s>quickly and deterministically. This means that Naive Bayes classifiers can perform well even with high-dimensional data points and/or a large number of data points. Figure 3.1: Workflow of Naive Bayes Classifier 3.4 SEARCH ALGORITHM Artificial Intelligence is the study of building agents that act rationally. Most of... |
<s>C for the response. C is a function of ConfidenceValue(t, r). The r with the highest score C is the response which will be sent to the output adapter. To find a good response you would calculate the score for multiple responses and choose the one with the highest score. “Selecting a potential response from a set of ... |
<s>knowledge of possible responses as it interacts with humans and other sources of informative data. An untrained instance of ChatterBot starts off with no knowledge of how to communicate. Each time a user enters a statement, the library saves the text that they entered and the text that the statement was in response ... |
<s>of strings where the list represents a conversation. In this case, the order of each response is based on its placement in a given conversation or the list of string. 1. Import Chatterbot and Trainer Class Library 2. Set the trainer as List Data Trainer 3. Provide the list of strings which you want to provide for tr... |
<s>designed to allow a chat bot to have a versatile method of receiving or retrieving input from a given source. It is required to add in parameters to specify the input and output terminal adapter. The input terminal adapter simply reads the user’s input from the terminal. 4.3.2 INPUT ADAPTERS The Chatterbot’s input a... |
<s>number of items that are shared between the sets. In the denominator, we 1. BestMatch logic adapter is a logic adapter that returns a response based on known responses to the closet matches to the input statement. 2. Import the Unicode literals and the logic adapter library. 3. In Get method, it takes a statement st... |
<s>4.4 ENVIRONMENT SETUP Natural Language Processing (NLP) techniques such as Natural Language Toolkit (NLTK) for Python can be applied to analyze speech, and intelligent responses can be found by designing an engine to provide appropriate human-like responses[1]. For the setup of Bengali chatbot, we installed Python 3... |
<s>also integrated this system in a personal blog. In the process we made a small Bengali corpus. Since the responses are as good as its knowledge base so a lot of work has to be done to enhance the knowledge base. Topic-wise data can be fed to our system for the enhancement of its knowledge base. But during building t... |
<s>different types of questions. In case of generating output there are two things, one is understanding the question from input of the user. And another one is giving answer from system learning as trained. Our system can successfully give answer to any question but not in Bengali. After successfully understanding the... |
<s>samples, we can see our QAs gives a similar reply like Neural Conversational Machine (NCM) whereas in comparison with a related work Cleverbot our QAs has outwitted it in most of the instance. Amongst the many limitations, the lack of a coherent personality makes it difficult for our system to pass the Turing test [... |
<s>base. Topic-wise data can be fed to our system for the enhancement of its knowledge base. But during building the knowledge base, the developer must provide a knowledge base free from errors. 45 | P a g e 6.2 LIMITATIONS QA systems can enable users to access the knowledge in a natural way by asking natural language ... |
<s>Boston/Drdrecht/London, pp. 5-8. [6] Hao Wang, Zhengdong Lu, Hang Li, and Enhong Chen. 2013. A dataset for research on short text conversations. In EMNLP, pages 935–945. [7] Greenwood M. and Gaizauskas R. Using a Named Entity Tagger to Generalise Surface Matching Text Patterns for Question Answering. In Proceedings ... |
<s>M. (1950). Computing machinery and intelligence. Mind, 59(236), 433-460. [26] Vinyals, O., & Le, Q. (2015). A neural conversational model. arXiv preprint arXiv:1506.05869. [27] Weizenbaum, J. (1966). ELIZA—a computer program for the study of natural language communication between man and machine. Communications of t... |
<s>Appendix B, we have shown some snapshots of our dataset. Figure B.1: Snapshot of the Dataset 71 | P a g e Figure B.2: Snapshot of the Dataset Figure B.3: Snapshot of the Dataset 72 | P a g e Figure B.4: Snapshot of the Dataset Figure B.5: Snapshot of the Dataset 73 | P a g e Figure B.6: Snapshot of the Dataset Figur... |
<s>untitledDetection of Semantic Errors from Simple Bangla Sentences K. M. Azharul Hasan, Muhammad Hozaifa Computer Science and Engineering Department Khulna University of Engineering & Technology Khulna 9203, Bangladesh. azhasan@gmail.com, hozaifa.moaj@gmail.com Sanjoy Dutta Computer Science and Engineering Department... |
<s>two fold; 1. Whether the verb with the subject has a well formed structure with semantic compatibility (SV relation) 2. Whether the Object and Verb (OV relation) has semantic compatibility. To establish these two relations, we have created classification table of Bangla verbs on the basis of tense, person and also a... |
<s>relation): If there is a well-established semantic bond between an Object of the sentence and verb then there is a true OV relation between the object and verb otherwise the relation is false. OV relation is true when there is a true SV relation in that sentence. For example “মানষু ভাত খায়” has a true OV relation an... |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.