Knowledge-Based Artificial Neural Networks

By Geoffrey G. Towell, Jude W. Shavlik. 1994

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

Citations in the corpus

Deep learning in neural networks: An overview (2014, methodology): “A hybrid approach (Shavlik & Towell, 1989; Towell & Shavlik, 1994) initializes a potentially deep FNN through adomain theory in propositional logic, which may be acquired through explanationbased learning (DeJong & Mooney, 1986; Minton et al., 1989; Mitchell, Keller, & Kedar-Cabelli, 1986).”
Neural-Symbolic Computing: An Effective Methodology for Principled Integration of Machine Learning and Reasoning (2019, methodology): “KBANN[49] ✓ ✓ ✓”
End-to-end Differentiable Proving (2017, neutral): “Early proposals for neural-symbolic networks are limited to propositional rules ( e.g. , EBL-ANN [36], KBANN [37] and C-IL 2 P [38]).”
Knowledge Matters: Importance of Prior Information for Optimization (2013, background): “• Local Descent Hypothesis: The brain of a biological agent relies on approximate local descent and gradually improves itself while learning.”
Extracting comprehensible models from trained neural networks (1996, methodology): “Learning algorithms have been used to refine approximately-correct domain theories (Pazzani & Kibler, 1992; Ourston & Mooney, 1994; Towell & Shavlik, 1994).”
Model-Agnostic Linear Competitors - When Interpretable Models Compete and Collaborate with Black-Box Models (2019, background): “…body of continuous work on neural-symbolic or neural-expert systems (Garcez et al. 2015) pursued by a relatively small research community over the last two decades and has yielded several signiﬁcant results (McGarry et al. 1999, Garcez et al. 2012, Taha and Ghosh 1999, Towell and Shavlik 1994).”
Connectionist Theory Refinement: Genetically Searching the Space of Network Topologies (1997, unknown): “…because inductive learners that start with an approximately correct theory can achieve improved \generalization" (accuracy on examples not seen during training) with signi(cid:12)cantly fewer training examples (Ginsberg, 1990; Ourston & Mooney, 1994; Pazzani & Kibler, 1992; Towell & Shavlik, 1994).”
Interacting meaningfully with machine learning systems: Three experiments (2009, background): “Towell and Shavlik 1994, Haddawy et al. 2003, Langseth and Nielsen 2003, Yu 2007).”
Ensemble-based classifiers (2010, methodology): “Most research in this area has been concerned with combining empirical approaches with analytical methods (see for instance Towell and Shavlik 1994 .”
A Connectionist Computational Model for Epistemic and Temporal Reasoning (2006, background): “…from examples) into neural networks has been shown to provide a learning system that is more effective than purely symbolic or purely connectionist systems, especially when data are noisy Neural Computation 18, 1711–1738 (2006) C© 2006 Massachusetts Institute of Technology (Towell & Shavlik, 1994).”
Dimensions of Neural-symbolic Integration - A Structured Survey (2005, methodology): “The knowledge based artificial neural networks (KBANN) [Towell and Shavlik, 1994] are closely related to this approach, by using similar techniques to implement propositional logic formulae within neural networks, but with a focus on learning.”
Bridging the Gap: Converting Human Advice into Imagined Examples (2019, methodology): “Users provided advice in the form of simple if-then rules that were then converted into a knowledge-based neural network (Towell & Shavlik, 1994).”
Web mining in soft computing framework: relevance, state of the art and future directions (2002, methodology): “Information is derived by extracting rules from KBNNs [62].”
Lifelong Learning: A Case Study. (1995, background): “In some approaches, bias is obtained explicitly through the expertise of a human expert of the domain, communicated by symbolic if-then rules [33] [12] [65] [41] [40] [38].”
Active Learning (2012, background): “Such hybrid approaches of “explanation-based” and empirical, statistical machine learning have a long history, beginning in the mid-1990s with neural networks (Towell and Shavlik, 1994), and continuing through modern learning systems like SVMs (Kunapuli et al.”