Filters








365,583 Hits in 7.7 sec

Active Learning for New Domains in Natural Language Understanding [article]

Stanislav Peshterliev, John Kearney, Abhyuday Jagannatha, Imre Kiss, Spyros Matsoukas
2019 arXiv   pre-print
We explore active learning (AL) for improving the accuracy of new domains in a natural language understanding (NLU) system.  ...  Additionally, case studies with human-in-the-loop AL on six new domains show 4.6%-9% improvement on an existing NLU system.  ...  Conclusions In this work, we focused on AL methods designed to select live data for manual annotation. The difference with prior work on AL is that we specifically target new domains in NLU.  ... 
arXiv:1810.03450v2 fatcat:advsnoleczbwzaws3i5x3scvw4

Active Learning for New Domains in Natural Language Understanding

Stanislav Peshterliev, John Kearney, Abhyuday Jagannatha, Imre Kiss, Spyros Matsoukas
2019 Proceedings of the 2019 Conference of the North  
We explore active learning (AL) for improving the accuracy of new domains in a natural language understanding (NLU) system.  ...  Additionally, case studies with human-in-the-loop AL on six new domains show 4.6%-9% improvement on an existing NLU system.  ...  For IVA, natural language understanding (NLU) is a main component (De Mori et al., 2008) , in conjunction with automatic speech recognition (ASR) and dialog management (DM).  ... 
doi:10.18653/v1/n19-2012 dblp:conf/naacl/PeshterlievKJKM19 fatcat:swrctrp2uravta2qr5cczvsum4

TULIPS - Teachable, Understanding Natural Language Input Problem-Solver

Michael G. Malkovsky
1975 International Joint Conference on Artificial Intelligence  
This paper describes the principal features of TULIPS program designed as a widely-oriented AI system that accepts natural language (NL) input.  ...  Having solved them TULIPS not only answers user's request but also learns new knowledge on its environment and improves its own mechanisms.  ...  It is obvious that a necessary feature of NL understanding (in the sense defined above) is using of data not only on a language but also on a current speaker, concret problem-domain and hearer's activities  ... 
dblp:conf/ijcai/Malkovsky75 fatcat:qkktn4r3cncuhlc25hu3l6f4vy

SkillNet-NLU: A Sparsely Activated Model for General-Purpose Natural Language Understanding [article]

Fan Zhang, Duyu Tang, Yong Dai, Cong Zhou, Shuangzhi Wu, Shuming Shi
2022 arXiv   pre-print
We evaluate on natural language understandings tasks and have the following findings.  ...  When learning for a new task, our approach precisely activates required skills and also provides an option to add new skills.  ...  to understand natural language questions, the skill to understand the meaning of texts in finance domain, etc.  ... 
arXiv:2203.03312v4 fatcat:g2o2g5kl5fbv3hrv4go7sxgtca

Robot Learning of Domain Specific Knowledge from Natural Language Sources [chapter]

Ines eh, Sandi Pohorec, Marjan Mernik, Milan Zorm
2010 Robot Learning  
The same is valid for learning of domain specific knowledge although the very first activity is the domain definition.  ...  In order to perform such a task it is vital that we are able to understand the natural language.  ...  /robot-learning-of-domain-specificknowledge-from-natural-language-sources  ... 
doi:10.5772/10249 fatcat:cjoy5bfxk5cffi75s542esjhre

enetCollect: A New European Network for combining Language Learning with Crowdsourcing Techniques

Rodrigo Agerri, Montse Maritxalar, Verena Lyding, Lionel Nicolas
2018 Revista de Procesamiento de Lenguaje Natural (SEPLN)  
framework for achieving a shared understanding of Language Learning and Crowdsourcing.  ...  This will allow to unlock the crowdsourcing potential available for language learning and to facilitate the development of prototypical experiments for the production of language learning material, such  ...  for language learning but also for domain specific learning (Science, History, etc).  ... 
dblp:journals/pdln/AgerriMLN18 fatcat:hc37h2lrqjcwfaorfhhdk6ou4m

Ontology Guided Semantic Self Learning Framework

Darshika N. Koggalahewa, Asoka S. Karunananda
2015 International Journal of Knowledge Engineering-IACSIT  
The research solved the problem of lack of understanding and capturing the knowledge from Natural Language sources and represents the captured knowledge in an automated and efficient manner.  ...  Ontology Guided Semantic Self Learning Framework is a software which is capable of learning from natural language sources.  ...  Knowledge acquisition is important for two basic activities such as learning and teaching. Learning is acquiring new knowledge, behaviours, skills, values or preferences.  ... 
doi:10.7763/ijke.2015.v1.5 fatcat:urnd23am4bggligwixp7yuj7lu

An Explanation Tool to Support Learning of Basic Programming

Carlos Tobar, Claudio Oliveira, Ricardo De Freitas
2006 Proceedings. Frontiers in Education. 36th Annual Conference  
Using techniques of approximate natural language understanding, the system tries to recognize explanation contents to send the result to a module that classifies explanations as correct, incorrect, or  ...  The domains considered ideal for ITS contain easily represented issues in computational form and allow the interaction type between student and ITS be limited to a restricted set of words, symbols, and  ...  Natural Language Understanding The module for approximate natural language understanding tries to perform the following operations based on the explanation generated by a student: • Lexical analysis -it  ... 
doi:10.1109/fie.2006.322578 fatcat:itrcpuhcg5cdferilnblvfasae

Metaphors of Learning Chinese Students Live by

Can Xiong
2015 Journal of Language Teaching and Research  
The application of metaphor into education language provides us with a better understanding of the learning activity.  ...  In this paper, embedded epistemological assumptions on the nature, the practice and the aims of learning are conveyed via metaphorical models.  ...  The application of metaphor into education language provides us with a better understanding of the educational activity. This paper tries to focus on the learning aspect of the whole activity.  ... 
doi:10.17507/jltr.0604.23 fatcat:prsai4fm7jf3nl73z3i2a3xrj4

Teaching Natural Language Processing (NLP) Using Ontology Based Education Design

Zobia Rehman, Stefania Kifor
2015 Balkan Region Conference on Engineering and Business Education  
Teaching via ontology is in practice since last two decades. Natural Language Processing (NLP) is a combination of computation and linguistic and is often hard to teach.  ...  Visual and symbolic representation of knowledge is easy to understand for both teachers and students.  ...  Apart from all these issues an instructor has to make some room in his lectures for programming as the NLP friendly Python (a programming language) is always new to students.  ... 
doi:10.1515/cplbu-2015-0024 fatcat:rwjzgqyvazafnh543u472ip2cm

ADVANCES IN KNOWLEDGE ACQUISITION AND REPRESENTATION

LAWRENCE B. HOLDER, ZDRAVKO MARKOV, INGRID RUSSELL
2006 International journal on artificial intelligence tools  
support understanding and reasoning.  ...  At the heart of this integration is the need for performance feedback throughout the process to guide the selection of alternative methods, the support for human interaction in the process, and the definition  ...  s article in this issue describes one such scenario in which knowledge representation is used to support natural language understanding.  ... 
doi:10.1142/s0218213006003016 fatcat:u4ftdbe34nbq7fs3lrq45awbv4

Machine Common Sense Concept Paper [article]

David Gunning
2018 arXiv   pre-print
The absence of common sense prevents intelligent systems from understanding their world, behaving reasonably in unforeseen situations, communicating naturally with people, and learning from new experiences  ...  Recent advances in machine learning have resulted in new AI capabilities, but in all of these applications, machine reasoning is narrow and highly specialized.  ...  Joshua Alspector for his expertise in deep learning and insights into its potential for achieving machine common sense; and Ms.  ... 
arXiv:1810.07528v1 fatcat:pntbtkr7tfglxbdznb4t3inhsm

Guidelines for Efficient and Effective End-User Development of Mashups [chapter]

Saeed Aghaee, Cesare Pautasso
2013 Lecture Notes in Computer Science  
Mashups provide a high potential for EUD activities on the Web.  ...  Users on the Web can tap into a vast resource of off-theshelf components in order to rapidly compose new lightweight software applications called mashups.  ...  The mashup descriptions written in natural language are readable and understandable by any English-speaking end user.  ... 
doi:10.1007/978-3-642-38706-7_23 fatcat:epjxkzqhqbf4vdn5mpamq6sb4u

ANDERSON TAXONOMY-BASED INTENSIVE TEST EVALUATION TOOL FOR SENIOR HIGH SCHOOL

Abduh Khairani Nurul
2019 Ideas  
Evaluation is one of the components covered in the education system and listening is one of the language skills that needs maximum attention.  ...  (C5), and create (C6) The results of the test instruments made were intended for class X high school students in the form of multiple choice tests in the amount of twenty questions.  ...  Although in reality, language is naturally verbal and manifested in speaking and understanding the conversation.  ... 
doi:10.24256/ideas.v7i1.725 fatcat:aaoxgsfkobfxbhclspzpqiecvm

Basic Artificial Intelligence Research at the Georgia Institute of Technology

Ashok K. Goel, Ronald C. Arkin, Kurt P. Eiselt, Janet L. Kolodner, Daryl T. Lawton, Ashwin Ram
1991 The AI Magazine  
The AQUA project explores these ideas in a natural language-understanding domain.  ...  Question-Driven Understanding Ram is investigating research issues in the areas of creative understanding, machine learning, and abduction in a natural language-understanding domain.  ...  Subscription price: $249 for 12 issues. Applied AI News Continued from page 16  ... 
doi:10.1609/aimag.v12i2.897 dblp:journals/aim/GoelAEKLR91 fatcat:xh7x2fh2ezg4dkv6sd7wpk6yle
« Previous Showing results 1 — 15 out of 365,583 results