489,839 Hits in 8.2 sec

Distributed knowledge by explanation networks

Y. Waern, R. Ramberg
2004 37th Annual Hawaii International Conference on System Sciences, 2004. Proceedings of the  
To conclude, the idea of distributing knowledge by support from an explanation network is fruitful and feasible.  ...  The explanation network idea has great knowledge acquisition power.  ...  This work has been supported by grants from the Swedish Board for Research in The Humanities and Social Sciences and the Swedish Board for Technical Development.  ... 
doi:10.1109/hicss.2004.1265332 dblp:conf/hicss/WaernR04 fatcat:jkozcsyrp5dcdjtsj6gy7wgmla

A Hybrid System Composed of Neural Networks and Genetic Algorithms

Dumitrescu Mihaela
2012 International Journal of Asian Business and Information Management  
In contrast to the systems mentioned above, neural networks are not generally able to offer explanations, because knowledge are not explicit, knowledge are transferred in the form of distributed weights  ...  Given the distributed way of knowledge representation, neural networks have some advantages in this respect.  ... 
doi:10.4018/jabim.2012100105 fatcat:5obtxkfkbzc3dmxil3iytcuf6y

The use of an explanation algorithm in a clinical event monitor

W R Hogan, M M Wagner
1999 Proceedings. AMIA Symposium  
Therefore, we decided to implement an advanced method for explanation (Suermondt's method for belief-network explanation).  ...  Clinical event monitors (CEMs) seek to improve patient care and reduce its cost by prompting clinicians to take actions that have these effects.  ...  Acknowledgements This work was supported by grants T15 LM/DE07059 and 5 R29 LM06233 from the National Library of Medicine.  ... 
pmid:10566365 pmcid:PMC2232853 fatcat:t726i4s2ofbm3p6a5clwc3bwoy

Handling Revision Inconsistencies: Creating Useful Explanations

Fabian Schmidt, Jorg Gebhardt, Rudolf Kruse
2015 2015 48th Hawaii International Conference on System Sciences  
In this article we examine the topic of creating and presenting explanations by discussing different components of inconsistencies and resulting consequences for explanations.  ...  initial distribution.  ...  One popular method for storing knowledge is by representing it in form of probability distributions and using the revision operation to incorporate new beliefs while keeping changes to the original knowledge  ... 
doi:10.1109/hicss.2015.447 dblp:conf/hicss/SchmidtGK15 fatcat:un4irznuo5cjfcd7fxunmauawu

Explaining Neural Networks Semantically and Quantitatively [article]

Runjin Chen, Hao Chen, Ge Huang, Jie Ren, Quanshi Zhang
2018 arXiv   pre-print
This paper presents a method to explain the knowledge encoded in a convolutional neural network (CNN) quantitatively and semantically.  ...  The analysis of the specific rationale of each prediction made by the CNN presents a key issue of understanding neural networks, but it is also of significant practical values in certain applications.  ...  Quantitative explanations α i y i for the male attribute.  ... 
arXiv:1812.07169v1 fatcat:e3d4cgdc6zhxndndvkkh24hhhq

On the Integration of Knowledge Graphs into Deep Learning Models for a More Comprehensible AI—Three Challenges for Future Research

Giuseppe Futia, Antonio Vetrò
2020 Information  
Within such a general direction, we identify three specific challenges for future research—knowledge matching, cross-disciplinary explanations and interactive explanations.  ...  A complementary approach is represented by symbolic AI, where symbols are elements of a lingua franca between humans and deep learning.  ...  As underlined by Rumelhart et al. [18] , the resulting knowledge consists of the connections between these computational units distributed throughout the network.  ... 
doi:10.3390/info11020122 fatcat:77ni2i6tdrhqxopw25vbybghi4

Page 197 of American Society of Civil Engineers. Collected Journals Vol. 118, Issue CP2 [page]

1992 American Society of Civil Engineers. Collected Journals  
In Neuroform the problem knowledge is distributed in the network weights and these weights create the desired mapping be- tween the system inputs and the outputs.  ...  In Neuroform the values of the network weights are determined by the network training and during the retrieval of the system outputs, all the weights are considered simultaneously (or in parallel).  ... 

Inference and reasoning in a Bayesian knowledge-intensive CBR system

Hoda Nikpour, Agnar Aamodt
2021 Progress in Artificial Intelligence  
AbstractThis paper presents the inference and reasoning methods in a Bayesian supported knowledge-intensive case-based reasoning (CBR) system called BNCreek.  ...  The semantic network inference methods and the CBR method are employed to handle the difficulties of inferencing and reasoning in uncertain domains.  ...  Pål Skalle for providing the drilling network model, and preparing the drilling cases, and Prof. Helge Langseth and Dr. Frode Sørmo for their useful suggestions.  ... 
doi:10.1007/s13748-020-00223-1 fatcat:fj5vslq6ovb4bja76jvsehpjju

Neuro-Symbolic Approaches for Knowledge Representation in Expert Systems

Ioannis Hatzilygeroudis, Jim Prentzas
2005 International Journal of Hybrid Intelligent Systems  
Section 2 discusses background knowledge focusing on the advantages and disadvantages of symbolic rules and neural networks.  ...  We then stress the point that not all hybrid neuro-symbolic approaches can be accommodated by existing categories.  ...  Acknowledgements This work was supported by the Research Committee of the University of Patras, Greece, Program "Karatheodoris", project No 2788.  ... 
doi:10.3233/his-2004-13-401 fatcat:uzemw427gjg6vj4tjqy6biddim

An intelligent problem solving environment for designing explanation models and for diagnostic reasoning in probabilistic domains [chapter]

Jörg Folckers, Claus Möbus, Olaf Schröder, Heinz-Jürgen Thole
1996 Lecture Notes in Computer Science  
This differs from existing reasoning systems based on Bayesian networks, i.e. in medical domains, which contain a built-in knowledge base that may be used but not created or modified by the learner.  ...  Uncertainty is handled by the Bayesian network approach. Thus the modelling task for the learner consists of creating a Bayesian network for the problem at hand.  ...  In this case, it is the uncertainty of knowledge, which we chose to handle by the Bayesian network approach.  ... 
doi:10.1007/3-540-61327-7_133 fatcat:4pxeqn3cbnbl7khl7ue6lsupxa

Correlation DialTone-Building Internet—Based Distributed Event Correlation Services [chapter]

Gabriel Jakobson, Girish Pathak
2000 Lecture Notes in Computer Science  
We describe process/knowledge representation models and implementation of the distributed correlation service using CORBA, XML, Java, and model-based reasoning technologies.  ...  We present the component-based architecture of the distributed event correlation service and its component services: correlation subscription, data mediation, event parsing, event correlation, event delivery  ...  Explanation Service The Explanation Service (Figure 4 ) is used to analyze conclusions or situations recognized by the Correlation Engine.  ... 
doi:10.1007/10722515_9 fatcat:som6dimg7ze23ath3tg3igohu4

The split-up system

John Zeleznikow, Andrew Stranieri
1995 Proceedings of the fifth international conference on Artificial intelligence and law - ICAIL '95  
Argument structures proposed by Toulmin can be used to represent legal knowledge in a manner that enables rulebased reasoning to be integrated with neural networks.  ...  Because explanations are at least as important as conclusions, we iHustrate the use of Toulrnin structures in the generation of explanations for conclusions reached by either mle sets or neural networks  ...  The knowledge that each expert brings to the task is represented within the knowledge base by Toulmin structures.  ... 
doi:10.1145/222092.222235 dblp:conf/icail/ZeleznikowS95 fatcat:oxq7yzjf2rbwnpbdudojxuptoe

Neurosymbolic AI: The 3rd Wave [article]

Artur d'Avila Garcez, Luis C. Lamb
2020 arXiv   pre-print
We focus on research that integrates in a principled way neural network-based learning with symbolic knowledge representation and logical reasoning.  ...  for network models.  ...  In neural-symbolic computation, knowledge learned by a neural network can be represented symbolically. Reasoning takes place either symbolically or within the network in distributed form.  ... 
arXiv:2012.05876v2 fatcat:m5g4kvcvo5epzmlsmft7vj5hva

Page 934 of Psychological Abstracts Vol. 80, Issue 2 [page]

1993 Psychological Abstracts  
In neural networks, knowledge is encoded in numeric parameters and distributed all over the system. This paper dis- cusses the ability of neural networks to generate explanations.  ...  Connectionist semantic networks (i.e., connectionist systems with an explicit conceptual hierarchy) be- long to a class of artificial neural networks that can be extended by an explanation component that  ... 

Introspective Learning by Distilling Knowledge from Online Self-explanation [article]

Jindong Gu and Zhiliang Wu and Volker Tresp
2020 arXiv   pre-print
We start by investigating the effective components of the knowledge transferred from the teacher network to the student network.  ...  Motivated by the conclusion, we propose an implementation of introspective learning by distilling knowledge from online self-explanations.  ...  Networks Learned from Peer Networks In our proposed algorithm, we trains a network with knowledge in explanations created on the same network.  ... 
arXiv:2009.09140v1 fatcat:z2j222u77rb23fetcxaoqhe6n4
« Previous Showing results 1 — 15 out of 489,839 results