9,754 Hits in 2.5 sec

Learning Heuristic Search via Imitation [article]

Mohak Bhardwaj, Sanjiban Choudhury, Sebastian Scherer
2017 arXiv   pre-print
We present SaIL, an efficient algorithm that trains heuristic policies by imitating "clairvoyant oracles" - oracles that have full information about the world and demonstrate decisions that minimize search  ...  Our approach paves the way forward for learning heuristics that demonstrate an anytime nature - finding feasible solutions quickly and incrementally refining it over time.  ...  We demonstrate that we are able to learn heuristic policies with widely varying characteristics simply by training on different data distributions (Section 4.3).  ... 
arXiv:1707.03034v1 fatcat:wvaf7se5kndohlisqlhphj7bra

Learning to Speed Up Query Planning in Graph Databases [article]

Mohammad Hossain Namaki, F A Rezaur Rahman Chowdhury, Md Rakibul Islam, Janardhan Rao Doppa, Yinghui Wu
2018 arXiv   pre-print
Subsequently, we learn greedy search control knowledge to imitate the search behavior of the target query plans.  ...  In this paper, we study the problem of learning to speedup query planning in graph databases towards the goal of improving the computational-efficiency of query processing via training queries.We present  ...  heuristic // Pruning 14: end while 15: return state-action pair sequence from s0 to s * Learning Greedy Policies via Imitation Learning Our goal is to learn a greedy policy (Π select , Π f etch ) that  ... 
arXiv:1801.06766v1 fatcat:2yexjo3eyjfmlorogc3sqawi7a

HC-Search: A Learning Framework for Search-based Structured Prediction

J.R. Doppa, A. Fern, P. Tadepalli
2014 The Journal of Artificial Intelligence Research  
Given a structured input, the framework uses a search procedure guided by a learned heuristic H to uncover high quality candidate outputs and then employs a separate learned cost function C to select a  ...  Guided by this decomposition, we minimize the overall loss in a greedy stage-wise manner by first training H to quickly uncover high quality outputs via imitation learning, and then training C to correctly  ...  Following our staged learning approach, one could start with learning a heuristic via exact imitation of the oracle search.  ... 
doi:10.1613/jair.4212 fatcat:ixnlsxzwurcv5kaejmsyu2x37e

A Survey for Solving Mixed Integer Programming via Machine Learning [article]

Jiayi Zhang and Chang Liu and Junchi Yan and Xijun Li and Hui-Ling Zhen and Mingxuan Yuan
2022 arXiv   pre-print
Then, we advocate further promoting the different integration of machine learning and MIP and introducing related learning-based methods, which can be classified into exact algorithms and heuristic algorithms  ...  Like other CO problems, the human-designed heuristic algorithms for MIP rely on good initial solutions and cost a lot of computational resources.  ...  Heuristic Algorithms Though MIP can be solved via B&B exactly, it is time and resource-consuming due to its NP-hard nature.  ... 
arXiv:2203.02878v1 fatcat:wiezy5ilird3fgbf3qdjzj4wmu

ℋC-search for structured prediction in computer vision

Michael Lam, Janardhan Rao Doppa, Sinisa Todorovic, Thomas G. Dietterich
2015 2015 IEEE Conference on Computer Vision and Pattern Recognition (CVPR)  
We complement this search operator by applying the DAGGER algorithm to robustly train the search heuristic so it learns from its previous mistakes.  ...  Search-based methods provide an alternative that has the potential to achieve higher performance. These methods learn to control a search procedure that constructs and evaluates candidate solutions.  ...  Exact Imitation Learning: In exact imitation learning, the heuristic function H is learned by "imitating" the search trajectory of the oracle heuristic.  ... 
doi:10.1109/cvpr.2015.7299126 dblp:conf/cvpr/LamDTD15 fatcat:cag6ahnl4bealizhguaahvtupm

Learning to Search via Retrospective Imitation [article]

Jialin Song, Ravi Lanka, Albert Zhao, Aadyot Bhatnagar, Yisong Yue, Masahiro Ono
2019 arXiv   pre-print
We study the problem of learning a good search policy for combinatorial search spaces.  ...  We propose retrospective imitation learning, which, after initial training by an expert, improves itself by learning from retrospective inspections of its own roll-outs.  ...  In this paper, we take a learning approach to finding an effective search heuristic.  ... 
arXiv:1804.00846v4 fatcat:pw4vb2mi6ffppoct3wsdmceo4y

Data-driven Planning via Imitation Learning [article]

Sanjiban Choudhury, Mohak Bhardwaj, Sankalp Arora, Ashish Kapoor, Gireeja Ranade, Sebastian Scherer, Debadeepta Dey
2017 arXiv   pre-print
We hence present a novel data-driven imitation learning framework to efficiently train planning policies by imitating a clairvoyant oracle - an oracle that at train time has full knowledge about the world  ...  In this paper, we address the problem of enabling planners to adapt their search strategies by inferring such good actions in an efficient manner using only the information uncovered by the search up until  ...  Solving POMDP via Imitation of a Clairvoyant Oracle To examine the applicability of imitation learning in the POMDP framework, we compare the loss function (9) to the action value function (5) .  ... 
arXiv:1711.06391v1 fatcat:2nltaus5bfhdldnuequtw3c4fe

Interorganizational imitation heuristics arising from cognitive frames

Ralitza Nikolaeva
2014 Journal of Business Research  
imitation heuristics.  ...  The literature on organizational imitation mostly disregards its cognitive aspect. Yet, imitation is a cognitive heuristic for complex strategic decisions.  ...  Thus, when the adoption issue is viewed as an opportunity, but the non-adoption is perceived as a threat, social learning would occur either via conformist transmission or via prestige-biased transmission  ... 
doi:10.1016/j.jbusres.2014.03.001 fatcat:nzv2rsvkyjan3py4bzk22beyhy

Bounded Suboptimal Search with Learned Heuristics for Multi-Agent Systems

Markus Spies, Marco Todescato, Hannes Becker, Patrick Kesper, Nicolai Waniek, Meng Guo
Building upon recent results on applying deep learning to learn generalized reactive policies, we propose to learn heuristics by imitation learning.  ...  However, directly applying learned heuristics in search algorithms such as A∗ breaks optimality guarantees, since learned heuristics are not necessarily admissible.  ...  Multi-Agent Policy via Imitation Learning As mentioned earlier, we first train a DNN to imitate expert behaviors.  ... 
doi:10.1609/aaai.v33i01.33012387 fatcat:26zm7zxebzfrppybpflukmi66q

Learning TSP Requires Rethinking Generalization [article]

Chaitanya K. Joshi, Quentin Cappart, Louis-Martin Rousseau, Thomas Laurent
2021 arXiv   pre-print
Towards leveraging transfer learning to solve large-scale TSPs, this paper identifies inductive biases, model architectures and learning algorithms that promote generalization to instances larger than  ...  While state-of-the-art Machine Learning approaches perform closely to classical solvers when trained on trivially small sizes, they are unable to generalize the learnt policy to larger instances of practical  ...  Models can be trained to imitate optimal solvers via supervised learning or by minimizing the length of TSP tours via reinforcement learning.  ... 
arXiv:2006.07054v3 fatcat:fsxtrv2tzveabftlxzjuzpdura

Extracting Relations between Non-Standard Entities using Distant Supervision and Imitation Learning

Isabelle Augenstein, Andreas Vlachos, Diana Maynard
2015 Proceedings of the 2015 Conference on Empirical Methods in Natural Language Processing  
We propose to ameliorate this issue by jointly training the named entity classifier and the relation extractor using imitation learning which reduces structured prediction learning to classification learning  ...  Compared to using FIGER and Stanford NER, average precision is 10 points and 19 points higher with our imitation learning approach.  ...  Also referred to as search-based structured prediction or learning to search.  ... 
doi:10.18653/v1/d15-1086 dblp:conf/emnlp/AugensteinVM15 fatcat:txoviervmzhz3fkvsph4h4eboe

A General Large Neighborhood Search Framework for Solving Integer Linear Programs [article]

Jialin Song, Ravi Lanka, Yisong Yue, Bistra Dilkina
2020 arXiv   pre-print
We show that one can learn a good neighborhood selector using imitation and reinforcement learning techniques.  ...  We focus on solving integer programs, and ground our approach in the large neighborhood search (LNS) paradigm, which iteratively chooses a subset of variables to optimize while leaving the remainder fixed  ...  Learning to Search. In learning to search, one typically operates within the framework of a search heuristic, and trains a local decision policy from training data.  ... 
arXiv:2004.00422v3 fatcat:w4rfod67pfgqhlquol3mbamvju

Heuristic Search Planning with Deep Neural Networks using Imitation, Attention and Curriculum Learning [article]

Leah Chrestien, Tomas Pevny, Antonin Komenda, Stefan Edelkamp
2021 arXiv   pre-print
This paper presents a network model to learn a heuristic capable of relating distant parts of the state space via optimal plan imitation using the attention mechanism, which drastically improves the learning  ...  Learning a well-informed heuristic function for hard task planning domains is an elusive problem.  ...  Neural search policies (Gomoluch et al. 2020) rely on parameter learning to generate heuristics during the actual search process.  ... 
arXiv:2112.01918v1 fatcat:6ihocezymngurk2bcgzvdfpbcm

Adaptive Imitation Scheme for Memetic Algorithms [chapter]

Ehsan Shahamatnia, Ramin Ayanzadeh, Rita A. Ribeiro, Saeid Setayeshi
2011 IFIP Advances in Information and Communication Technology  
These algorithms are similar in nature to genetic algorithms as they follow evolutionary strategies, but they also incorporate a refinement phase during which they learn about the problem and search space  ...  The efficiency of these algorithms depends on the nature and architecture of the imitation operator used.  ...  the exploration capacities of underlying GA and the exploitation capabilities of local search heuristic.  ... 
doi:10.1007/978-3-642-19170-1_12 fatcat:yxt5pyb5ljcvdmvxzxzvze6gfi

Learning Generalized Reactive Policies using Deep Neural Networks [article]

Edward Groshev, Maxwell Goldstein, Aviv Tamar, Siddharth Srivastava, Pieter Abbeel
2018 arXiv   pre-print
We show that our approach can also be used to automatically learn a heuristic function that can be used in directed search algorithms.  ...  heuristic functions with minimal human input.  ...  As seen in Figure 3b the learned GRP heuristic significantly outperforms the Manhattan heuristic in both greedy search and A* search, on the 9x9 problems.  ... 
arXiv:1708.07280v3 fatcat:fybtlstknbck7popa3lbgxgob4
« Previous Showing results 1 — 15 out of 9,754 results