A Graph Grammar Approach to Generate Neural Network Topologies

Vempati, Chaitanya; Campbell, Matthew I.

doi:10.1115/detc2007-34588

Cited by 7 publications

(3 citation statements)

References 0 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In addition to the double-pushout method, a final step has been added wherein the graph is repaired by attaching or deleting dangling arcs in order keep a complete graph (Nagl, 1976). (Campbell, 2008) Grammars have a number of applications, from implementing a MEMS fabrication sequence (Jawalkar & Campbell, 2007) to developing neural networks (Vempati & Campbell, 2007) to more strictly mechanical application such as epicyclic gear trains (Schmidt & Chase, December 2000) and clocks (Starling & Shea, 2002). Our focus is on the possibilities of grammars to assist in the creation of gear trains including simple, compound, bevel, and worm gear trains, ignoring epicyclic gear train configurations and helical gears.…”

Section: Chapter 2: Backgroundmentioning

confidence: 99%

Automated Synthesis and Optimization of Gear Train Topologies

Swantner

Campbell

2009

Volume 5: 35th Design Automation Conference, Parts a and B

Self Cite

View full text Add to dashboard Cite

A method for automating the design of simple and compound gear trains using graph grammars is described. The resulting computational tool removes the tedium for engineering designers searching through the immense number of possible gear choices and combinations by hand. The variables that are automatically optimized by the computational tool include the gear dimensions as well as the location of the gears in space. The gear trains are optimized using a three-step process. The first step is a tree-search based on a language of gear rules which represent all possible gear train configurations. The second step optimizes the discrete values such as number of teeth through an exhaustive search of a gear catalog. The final step is a gradient-based algorithm which optimizes the non-discrete variables such as angles and lengths in the positioning of the gears. The advantage of this method is that the graph grammar allows all possible simple and compound gear trains to be included in the search space while the method of optimization ensures the optimal candidate for a given problem is chosen with the tuned parameters.

show abstract

Section: Chapter 2: Backgroundmentioning

confidence: 99%

Automated Synthesis and Optimization of Gear Train Topologies

Swantner

Campbell

2009

Volume 5: 35th Design Automation Conference, Parts a and B

Self Cite

View full text Add to dashboard Cite

show abstract

“…However, existing methods need be extend in these aspects: a) algorithm universality [2,3]; b) application range [4,5]; c) intuitive or concise model expression [6,7,8], d) model quality and computation performances [9]; e) a good appearance. In this paper, a novel automatic generation algorithm for NM is proposed based on the above aspects.…”

Section: Introductionmentioning

confidence: 99%

Automatic generation algorithms of network models for complex systems

Guo

Xiao

Han

2010

The 2nd International Conference on Information Science and Engineering

View full text Add to dashboard Cite

Presently, lots of theory achievements have been gotten on the modeling of complex systems. However, it is necessary that theory achievements are applied into practice. In addition, it is difficult that models of complex systems are automatically constructed. So, a novel "draw-net" algorithm based on a structure space is presented to realize automatic network model graph generation for complex systems. The structure space is designed with a special data structure, which is used to save the basic information of the network model. Then different spaces are defined to manage the classified information according to certain operating sequence. The generation model is based on classification spaces, and these works contain the computing and the drawing of graphic objects. In the generation process, the iterative calculation is used; the whole model is produced with nodes, relationships, and parameters eventually. Finally, an example on making network models is used to illustrate the proposed algorithm. It shows this algorithm is feasible, and can satisfy realistic requirements. Simultaneously, it is discovered that the auto-generation arithmetic has a good universality, and can be widely extended in the practices.

show abstract

“…[43] and [23], but also neural networks [51]. Because graph grammars modify a valid graph into another valid graph, each state can be analyzed.…”

Section: Computational Synthesismentioning

confidence: 99%

Knowledge engineering for design synthesis support

Schotborgh¹

View full text Add to dashboard Cite

A Graph Grammar Approach to Generate Neural Network Topologies

Cited by 7 publications

References 0 publications

Automated Synthesis and Optimization of Gear Train Topologies

Automated Synthesis and Optimization of Gear Train Topologies

Automatic generation algorithms of network models for complex systems

Knowledge engineering for design synthesis support

Contact Info

Product

Resources

About