Pub Date : 2024-04-04DOI: 10.1007/978-3-031-44505-7_35
Hern'an Ceferino V'azquez, Jorge Sanchez, Rafael Carrascosa
{"title":"Integrating Hyperparameter Search into Model-Free AutoML with Context-Free Grammars","authors":"Hern'an Ceferino V'azquez, Jorge Sanchez, Rafael Carrascosa","doi":"10.1007/978-3-031-44505-7_35","DOIUrl":"https://doi.org/10.1007/978-3-031-44505-7_35","url":null,"abstract":"","PeriodicalId":430111,"journal":{"name":"Learning and Intelligent Optimization","volume":"11 12","pages":"523-536"},"PeriodicalIF":0.0,"publicationDate":"2024-04-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"140741237","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-04-08DOI: 10.48550/arXiv.2304.04051
George Watkins, G. Montana, Juergen Branke
The graph colouring problem consists of assigning labels, or colours, to the vertices of a graph such that no two adjacent vertices share the same colour. In this work we investigate whether deep reinforcement learning can be used to discover a competitive construction heuristic for graph colouring. Our proposed approach, ReLCol, uses deep Q-learning together with a graph neural network for feature extraction, and employs a novel way of parameterising the graph that results in improved performance. Using standard benchmark graphs with varied topologies, we empirically evaluate the benefits and limitations of the heuristic learned by ReLCol relative to existing construction algorithms, and demonstrate that reinforcement learning is a promising direction for further research on the graph colouring problem.
{"title":"Generating a Graph Colouring Heuristic with Deep Q-Learning and Graph Neural Networks","authors":"George Watkins, G. Montana, Juergen Branke","doi":"10.48550/arXiv.2304.04051","DOIUrl":"https://doi.org/10.48550/arXiv.2304.04051","url":null,"abstract":"The graph colouring problem consists of assigning labels, or colours, to the vertices of a graph such that no two adjacent vertices share the same colour. In this work we investigate whether deep reinforcement learning can be used to discover a competitive construction heuristic for graph colouring. Our proposed approach, ReLCol, uses deep Q-learning together with a graph neural network for feature extraction, and employs a novel way of parameterising the graph that results in improved performance. Using standard benchmark graphs with varied topologies, we empirically evaluate the benefits and limitations of the heuristic learned by ReLCol relative to existing construction algorithms, and demonstrate that reinforcement learning is a promising direction for further research on the graph colouring problem.","PeriodicalId":430111,"journal":{"name":"Learning and Intelligent Optimization","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-04-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"128805016","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-02-26DOI: 10.48550/arXiv.2302.13225
Hui Wang, Abdallah Saffidine, T. Cazenave
Recent work proposed the UCTMAXSAT algorithm to address Maximum Satisfiability Problems (MaxSAT) and shown improved performance over pure Stochastic Local Search algorithms (SLS). UCTMAXSAT is based on Monte Carlo Tree Search but it uses SLS instead of purely random playouts. In this work, we introduce two algorithmic variations over UCTMAXSAT. We carry an empirical analysis on MaxSAT benchmarks from recent competitions and establish that both ideas lead to performance improvements. First, a nesting of the tree search inspired by the Nested Monte Carlo Search algorithm is effective on most instance types in the benchmark. Second, we observe that using a static flip limit in SLS, the ideal budget depends heavily on the instance size and we propose to set it dynamically. We show that it is a robust way to achieve comparable performance on a variety of instances without requiring additional tuning.
{"title":"Towards Tackling MaxSAT by Combining Nested Monte Carlo with Local Search","authors":"Hui Wang, Abdallah Saffidine, T. Cazenave","doi":"10.48550/arXiv.2302.13225","DOIUrl":"https://doi.org/10.48550/arXiv.2302.13225","url":null,"abstract":"Recent work proposed the UCTMAXSAT algorithm to address Maximum Satisfiability Problems (MaxSAT) and shown improved performance over pure Stochastic Local Search algorithms (SLS). UCTMAXSAT is based on Monte Carlo Tree Search but it uses SLS instead of purely random playouts. In this work, we introduce two algorithmic variations over UCTMAXSAT. We carry an empirical analysis on MaxSAT benchmarks from recent competitions and establish that both ideas lead to performance improvements. First, a nesting of the tree search inspired by the Nested Monte Carlo Search algorithm is effective on most instance types in the benchmark. Second, we observe that using a static flip limit in SLS, the ideal budget depends heavily on the instance size and we propose to set it dynamically. We show that it is a robust way to achieve comparable performance on a variety of instances without requiring additional tuning.","PeriodicalId":430111,"journal":{"name":"Learning and Intelligent Optimization","volume":"41 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-02-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133248216","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-07-26DOI: 10.48550/arXiv.2207.12688
Efthyvoulos Drousiotis, P. Spirakis
Decision trees (DT) are highly famous in machine learning and usually acquire state-of-the-art performance. Despite that, well-known variants like CART, ID3, random forest, and boosted trees miss a probabilistic version that encodes prior assumptions about tree structures and shares statistical strength between node parameters. Existing work on Bayesian DT depends on Markov Chain Monte Carlo (MCMC), which can be computationally slow, especially on high dimensional data and expensive proposals. In this study, we propose a method to parallelise a single MCMC DT chain on an average laptop or personal computer that enables us to reduce its run-time through multi-core processing while the results are statistically identical to conventional sequential implementation. We also calculate the theoretical and practical reduction in run time, which can be obtained utilising our method on multi-processor architectures. Experiments showed that we could achieve 18 times faster running time provided that the serial and the parallel implementation are statistically identical.
{"title":"Single MCMC Chain Parallelisation on Decision Trees","authors":"Efthyvoulos Drousiotis, P. Spirakis","doi":"10.48550/arXiv.2207.12688","DOIUrl":"https://doi.org/10.48550/arXiv.2207.12688","url":null,"abstract":"Decision trees (DT) are highly famous in machine learning and usually acquire state-of-the-art performance. Despite that, well-known variants like CART, ID3, random forest, and boosted trees miss a probabilistic version that encodes prior assumptions about tree structures and shares statistical strength between node parameters. Existing work on Bayesian DT depends on Markov Chain Monte Carlo (MCMC), which can be computationally slow, especially on high dimensional data and expensive proposals. In this study, we propose a method to parallelise a single MCMC DT chain on an average laptop or personal computer that enables us to reduce its run-time through multi-core processing while the results are statistically identical to conventional sequential implementation. We also calculate the theoretical and practical reduction in run time, which can be obtained utilising our method on multi-processor architectures. Experiments showed that we could achieve 18 times faster running time provided that the serial and the parallel implementation are statistically identical.","PeriodicalId":430111,"journal":{"name":"Learning and Intelligent Optimization","volume":"9 4","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-07-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"120821035","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-05-07DOI: 10.48550/arXiv.2205.03739
E. Agapaki
Airports are constantly facing a variety of hazards and threats from natural disasters to cybersecurity attacks and airport stakeholders are confronted with making operational decisions under irregular conditions. We introduce the concept of the foundational twin, which can serve as a resilient data platform, incorporating multiple data sources and enabling the interaction between an umbrella of twins. We then focus on providing data sources and metrics for each foundational twin, with an emphasis on the environmental airport twin for major US airports.
{"title":"Airport Digital Twins for Resilient Disaster Management Response","authors":"E. Agapaki","doi":"10.48550/arXiv.2205.03739","DOIUrl":"https://doi.org/10.48550/arXiv.2205.03739","url":null,"abstract":"Airports are constantly facing a variety of hazards and threats from natural disasters to cybersecurity attacks and airport stakeholders are confronted with making operational decisions under irregular conditions. We introduce the concept of the foundational twin, which can serve as a resilient data platform, incorporating multiple data sources and enabling the interaction between an umbrella of twins. We then focus on providing data sources and metrics for each foundational twin, with an emphasis on the environmental airport twin for major US airports.","PeriodicalId":430111,"journal":{"name":"Learning and Intelligent Optimization","volume":"199 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-05-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"115524830","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-04-11DOI: 10.48550/arXiv.2204.04960
A. Erzin, R. Plotnikov, I. Ladygin
The Constraint Shortest Path (CSP) problem is as follows. An $n$-vertex graph is given, each edge/arc assigned two weights. Let us call them"cost"and"length"for definiteness. Finding a min-cost upper-bounded length path between a given pair of vertices is required. The problem is NP-hard even when the lengths of all edges are the same. Therefore, various approximation algorithms have been proposed in the literature for it. The constraint on path length can be accounted for by considering one edge weight equals to a linear combination of cost and length. By varying the multiplier value in a linear combination, a feasible solution delivers a minimum to the function with new weights. At the same time, Dijkstra's algorithm or its modifications are used to construct the shortest path with the current weights of the edges. However, with insufficiently large graphs, this approach may turn out to be time-consuming. In this article, we propose to look for a solution, not in the original graph but specially constructed hierarchical structures (HS). We show that the shortest path in the HS is constructed with $O(m)$-time complexity, where $m$ is the number of edges/arcs of the graph, and the approximate solution in the case of integer costs and lengths of the edges is found with $O(mlog n)$-time complexity. The a priori estimate of the algorithm's accuracy turned out to depend on the parameters of the problem and can be significant. Therefore, to evaluate the algorithm's effectiveness, we conducted a numerical experiment on the graphs of roads of megalopolis and randomly constructed unit-disk graphs (UDGs). The numerical experiment results show that in the HS, a solution close to optimal one is built 10--100 times faster than in the methods which use Dijkstra's algorithm to build a min-weight path in the original graph.
{"title":"Constrained Shortest Path and Hierarchical Structures","authors":"A. Erzin, R. Plotnikov, I. Ladygin","doi":"10.48550/arXiv.2204.04960","DOIUrl":"https://doi.org/10.48550/arXiv.2204.04960","url":null,"abstract":"The Constraint Shortest Path (CSP) problem is as follows. An $n$-vertex graph is given, each edge/arc assigned two weights. Let us call them\"cost\"and\"length\"for definiteness. Finding a min-cost upper-bounded length path between a given pair of vertices is required. The problem is NP-hard even when the lengths of all edges are the same. Therefore, various approximation algorithms have been proposed in the literature for it. The constraint on path length can be accounted for by considering one edge weight equals to a linear combination of cost and length. By varying the multiplier value in a linear combination, a feasible solution delivers a minimum to the function with new weights. At the same time, Dijkstra's algorithm or its modifications are used to construct the shortest path with the current weights of the edges. However, with insufficiently large graphs, this approach may turn out to be time-consuming. In this article, we propose to look for a solution, not in the original graph but specially constructed hierarchical structures (HS). We show that the shortest path in the HS is constructed with $O(m)$-time complexity, where $m$ is the number of edges/arcs of the graph, and the approximate solution in the case of integer costs and lengths of the edges is found with $O(mlog n)$-time complexity. The a priori estimate of the algorithm's accuracy turned out to depend on the parameters of the problem and can be significant. Therefore, to evaluate the algorithm's effectiveness, we conducted a numerical experiment on the graphs of roads of megalopolis and randomly constructed unit-disk graphs (UDGs). The numerical experiment results show that in the HS, a solution close to optimal one is built 10--100 times faster than in the methods which use Dijkstra's algorithm to build a min-weight path in the original graph.","PeriodicalId":430111,"journal":{"name":"Learning and Intelligent Optimization","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-04-11","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129991981","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-01-17DOI: 10.1007/978-3-031-24866-5_27
Dan Halbersberg, M. Halevi, M. Salhov
{"title":"Search and Score-Based Waterfall Auction Optimization","authors":"Dan Halbersberg, M. Halevi, M. Salhov","doi":"10.1007/978-3-031-24866-5_27","DOIUrl":"https://doi.org/10.1007/978-3-031-24866-5_27","url":null,"abstract":"","PeriodicalId":430111,"journal":{"name":"Learning and Intelligent Optimization","volume":"255 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-01-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133529867","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2021-11-08DOI: 10.1007/978-3-030-92121-7_8
R. Dio, A. Galligo, Angelos Mantzaflaris, B. Mauroy
{"title":"Spirometry-based airways disease simulation and recognition using Machine Learning approaches","authors":"R. Dio, A. Galligo, Angelos Mantzaflaris, B. Mauroy","doi":"10.1007/978-3-030-92121-7_8","DOIUrl":"https://doi.org/10.1007/978-3-030-92121-7_8","url":null,"abstract":"","PeriodicalId":430111,"journal":{"name":"Learning and Intelligent Optimization","volume":"61 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-11-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"121946200","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2021-06-16DOI: 10.1007/978-3-031-24866-5_23
Xiaomeng Dong, Michael Potter, Gaurav Kumar, Yun-Chan Tsai, V. R. Saripalli, T. Trafalis
{"title":"Optimizing Data Augmentation Policy Through Random Unidimensional Search","authors":"Xiaomeng Dong, Michael Potter, Gaurav Kumar, Yun-Chan Tsai, V. R. Saripalli, T. Trafalis","doi":"10.1007/978-3-031-24866-5_23","DOIUrl":"https://doi.org/10.1007/978-3-031-24866-5_23","url":null,"abstract":"","PeriodicalId":430111,"journal":{"name":"Learning and Intelligent Optimization","volume":"37 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-06-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131924441","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2021-06-14DOI: 10.1007/978-3-030-92121-7_4
Jakob Bossek, Aneta Neumann, F. Neumann
{"title":"Exact Counting and Sampling of Optima for the Knapsack Problem","authors":"Jakob Bossek, Aneta Neumann, F. Neumann","doi":"10.1007/978-3-030-92121-7_4","DOIUrl":"https://doi.org/10.1007/978-3-030-92121-7_4","url":null,"abstract":"","PeriodicalId":430111,"journal":{"name":"Learning and Intelligent Optimization","volume":"36 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-06-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"124047990","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：