Finding Optimal Paths Using Networks Without Learning—Unifying Classical Approaches

Kulvičius, Tomas; Herzog, Sebastian; Tamošiūnaitė, Minija; Wörgötter, Florentin

doi:10.1109/tnnls.2021.3089023

Cited by 5 publications

(4 citation statements)

References 36 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…These approaches, however, may not always give optimal solutions or may fail to find solutions at all, especially for larger graphs. Some other bioinspired neural networks were proposed [22], [23], [24], [25], [26], [27], [28] for solving path-planning problems. These approaches work on grid structures, where activity in the network is propagated from the source neuron to the neighboring neurons until activity propagation within the whole network is finished.…”

Section: A State Of the Artmentioning

confidence: 99%

Combining Optimal Path Search With Task-Dependent Learning in a Neural Network

Kulvicius,

Tamosiunaite,

Wörgötter

2024

IEEE Trans. Neural Netw. Learning Syst.

View full text Add to dashboard Cite

Finding optimal paths in connected graphs requires determining the smallest total cost for traveling along the graph's edges. This problem can be solved by several classical algorithms, where, usually, costs are predefined for all edges. Conventional planning methods can, thus, normally not be used when wanting to change costs in an adaptive way following the requirements of some task. Here, we show that one can define a neural network representation of path-finding problems by transforming cost values into synaptic weights, which allows for online weight adaptation using network learning mechanisms. When starting with an initial activity value of one, activity propagation in this network will lead to solutions, which are identical to those found by the Bellman-Ford (BF) algorithm. The neural network has the same algorithmic complexity as BF, and, in addition, we can show that network learning mechanisms (such as Hebbian learning) can adapt the weights in the network augmenting the resulting paths according to some task at hand. We demonstrate this by learning to navigate in an environment with obstacles as well as by learning to follow certain sequences of path nodes. Hence, the here-presented novel algorithm may open up a different regime of applications where path augmentation (by learning) is directly coupled with path finding in a natural way.

show abstract

Section: A State Of the Artmentioning

confidence: 99%

Combining Optimal Path Search With Task-Dependent Learning in a Neural Network

Kulvicius,

Tamosiunaite,

Wörgötter

2024

IEEE Trans. Neural Netw. Learning Syst.

View full text Add to dashboard Cite

show abstract

“…Some other bio-inspired neural networks were proposed [22], [23], [24], [25], [26], [27], [28] for solving path planning problems. These approaches work on grid structures, where activity in the network is propagated from the source neuron to the neighbouring neurons until activity propagation within the whole network is finished.…”

Section: State-of-the-artmentioning

confidence: 99%

Combining optimal path search with task-dependent learning in a neural network

Kulvičius¹,

Tamošiūnaitė²,

Wörgötter³

2022

Preprint

Self Cite

View full text Add to dashboard Cite

Finding optimal paths in connected graphs requires determining the smallest total cost for traveling along the graph's edges. This problem can be solved by several classical algorithms where, usually, costs are predefined for all edges. Conventional planning methods can, thus, normally not be used when wanting to change costs in an adaptive way following the requirements of some task. Here we show that one can define a neural network representation of path finding problems by transforming cost values into synaptic weights, which allows for online weight adaptation using network learning mechanisms. When starting with an initial activity value of one, activity propagation in this network will lead to solutions, which are identical to those found by the Bellman Ford algorithm. The neural network has the same algorithmic complexity as Bellman Ford and, in addition, we can show that network learning mechanisms (such as Hebbian learning) can adapt the weights in the network augmenting the resulting paths according to some task at hand. We demonstrate this by learning to navigate in an environment with obstacles as well as by learning to follow certain sequences of path nodes. Hence, the here-presented novel algorithm may open up a different regime of applications where path-augmentation (by learning) is directly coupled with path finding in a natural way.

show abstract

“…2022 3713-3726 Heiden, U., see Hong, D., TNNLS Nov. 2022 6518-6531 Helaoui, M., see Hu, X., TNNLS Aug. 2022 3923-3937 Heng, P., see Liu, W., TNNLS Sept. 2022 4785-4799 Heo, B., see Ro, Y., TNNLS Sept. 2022 4648-4660 Herzog, S., see Kulvicius, T., TNNLS Dec. 2022 Huang, T., Deng, L., Jiang, T., Vivone, G., and Chanussot, J., Hyper-spectral Image Super-Resolution via Deep Spatiospectral Attention Convolutional Neural Networks; TNNLS Dec. 2022 7251-7265 Hu, L., see Wang, K., TNNLS May 2022 2159-2167 Hu, P., see Zhen, L., TNNLS Feb. 2022 798-810 Hu, Q., Zhang, H., Gao, F., Xing, C., and An, J., Hu, X., Liu, Z., Yu, X., Zhao, Y., Chen, W., Hu, B., Du, X., Li, X., Helaoui, M., Wang, W., and Ghannouchi, F.M., Convolutional Neural Network for Behavioral Modeling and Predistortion of Wideband Power Amplifiers; TNNLS Aug. 2022 3923-3937 Hu, X., see Bian, Y., TNNLS Dec. 2022 7928-7936 Hu, Y., see Guo, J., TNNLS July 2022 3157-3170 Hu, Y., see Ji, Q., TNNLS Oct. 2022 5681-5693 Hu, Y., Subagdja, B., Tan, A., and Yin, Q., Vision-Based Topological Mapping and Navigation With Self-Organizing Neural Networks; TNNLS 7101-7113 Hu, Z., Ren, H., and Shi, P., Zhao, R., Bi, L., Zhang, D., and Lu, C., Neural Embedding Singular Value Decomposition for Collaborative Filtering; TNNLS Oct. 2022 6021-6029 Huang, T., see He, X., 7818-7828 Huang, T., see Hu, J., 7251-7265 Huang, T., see Dong, T., TNNLS Dec. 2022 7266-7276 Huang, W., see Wang, Q., 1414-1428 Huang, W., Wu, H., Chen, Q., Luo, C., Zeng, S., Li, T., and Huang, Y., TNNLS Sept. 2022 4945-4959 Nguang, S.K., see Peng, Z., TNNLS Aug. 2022 4043-4055 Nguang, S.K., see Yan, S., 6905-6915 Nguyen, T., Roy, S., and Meunier, J., SmithNet: Strictness on Motion-Texture Coherence for Anomaly Detection; TNNLS June 2022 2287-2300 Ni, J., Huang, Z., Yu, C., Lv, D., and Wang, C., Comparative Convolutional Dynamic Multi-Attention Recommendation Model; TNNLS Aug. 2022 3510-3521 Ni, X., see Wen, S., TNNLS July 2022 3109-3119 Ni, Y., see Wang, Z., TNNLS Dec. 2022 7610-7620 Ni, Z., see Mu, C., TNNLS Sept. 2022 4437-4450 Nicolaou, M.A., see Oldfield, J., TNNLS Aug.…”

mentioning

confidence: 99%

“…Kulvicius, T., TNNLS Dec. 2022 7877-7887 Wu, B., see Zhao, S., TNNLS Feb. 2022 473-493 Wu, D., Shang, M., Luo, X., and Wang, Z., An L 1 -and-L 2 -Norm-Oriented Latent Factor Model for Recommender Systems; TNNLS Oct. 2022 5775-5788 Wu, D., Nie, F., Dong, X., Wang, R., and Li, X., Parameter-Free Consensus Embedding Learning for Multiview Graph-Based Clustering; TNNLS Dec.…”

mentioning

confidence: 99%

2022 Index IEEE Transactions on Neural Networks and Learning Systems Vol. 33

2022

IEEE Trans. Neural Netw. Learning Syst.

View full text Add to dashboard Cite

This index covers all technical items-papers, correspondence, reviews, etc.-that appeared in this periodical during 2022, and items from previous years that were commented upon or corrected in 2022. Departments and other items may also be covered if they have been judged to have archival value.The Author Index contains the primary entry for each item, listed under the first author's name. The primary entry includes the coauthors' names, the title of the paper or other item, and its location, specified by the publication abbreviation, year, month, and inclusive pagination. The Subject Index contains entries describing the item under all appropriate subject headings, plus the first author's name, the publication abbreviation, month, and year, and inclusive pages. Note that the item title is found only under the primary entry in the Author Index.

show abstract

Finding Optimal Paths Using Networks Without Learning—Unifying Classical Approaches

Cited by 5 publications

References 36 publications

Combining Optimal Path Search With Task-Dependent Learning in a Neural Network

Combining Optimal Path Search With Task-Dependent Learning in a Neural Network

Combining optimal path search with task-dependent learning in a neural network

2022 Index IEEE Transactions on Neural Networks and Learning Systems Vol. 33

Contact Info

Product

Resources

About