Article
Computer Science, Artificial Intelligence
Linnan Wang, Saining Xie, Teng Li, Rodrigo Fonseca, Yuandong Tian
Summary: This paper proposes a method called Latent Action Neural Architecture Search (LaNAS) to improve the efficiency of neural architecture search. By learning actions to partition the search space into regions with different performance metrics, LaNAS achieves significantly better sample efficiency compared to baseline methods. Experimental results demonstrate the superiority of LaNAS in various tasks.
IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE
(2022)
Correction
Chemistry, Multidisciplinary
Tarak K. Patra, Troy D. Loeffler, Subramanian K. R. S. Sankaranarayanan
Summary: Correction for the paper "Accelerating copolymer inverse design using monte carlo tree search" by Tarak K. Patra et al. in Nanoscale. The correction provides accurate information and data by addressing the errors in the original paper.
Correction
Chemistry, Multidisciplinary
Tarak K. Patra, Troy D. Loeffler, Subramanian K. R. S. Sankaranarayanan
Summary: This correction article addresses the paper 'Accelerating copolymer inverse design using monte carlo tree search' by Tarak K. Patra et al., published in Nanoscale, 2020, 12, 23653-23662, and provides clarification on the authors' work and the journal where it was published.
Article
Biochemistry & Molecular Biology
Mihaly Varadi, Damian Bertoni, Paulyna Magana, Urmila Paramval, Ivanna Pidruchna, Malarvizhi Radhakrishnan, Maxim Tsenkov, Sreenath Nair, Milot Mirdita, Jingi Yeo, Oleg Kovalevskiy, Kathryn Tunyasuvunakool, Agata Laydon, Augustin Zidek, Hamish Tomlinson, Dhavanthi Hariharan, Josh Abrahamson, Tim Green, John Jumper, Ewan Birney, Martin Steinegger, Demis Hassabis, Sameer Velankar
Summary: The AlphaFold Database Protein Structure Database (AlphaFold DB) has expanded significantly since its initial release in 2021, now containing over 214 million predicted protein structures. Powered by the AlphaFold2 artificial intelligence (AI) system, the database has integrated its predictions into primary data resources such as PDB, UniProt, Ensembl, InterPro, and MobiDB. This manuscript details the enhancements made to data archiving, including the addition of model organisms, global health proteomes, Swiss-Prot integration, and curated protein datasets. The access mechanisms of AlphaFold DB, from direct file access to advanced queries using Google Cloud Public Datasets, are also discussed, along with improvements and added services since its release, such as enhancements to the Predicted Aligned Error viewer and the 3D viewer customization options.
NUCLEIC ACIDS RESEARCH
(2023)
Article
Automation & Control Systems
Haokai Hong, Min Jiang, Gary G. Yen
Summary: The large-scale multiobjective optimization problem (LSMOP) involves optimizing multiple conflicting objectives and hundreds of decision variables. Existing algorithms often focus on improving performance but pay little attention to improving insensitivity. We propose an evolutionary algorithm based on Monte Carlo tree search to improve the performance and insensitivity of solving LSMOPs.
IEEE TRANSACTIONS ON CYBERNETICS
(2023)
Article
Multidisciplinary Sciences
Victor-Alexandru Darvariu, Stephen Hailes, Mirco Musolesi
Summary: We address the issue of goal-directed graph construction, aiming to find a set of edges that maximally improve a global objective function in a given starting graph. This problem is prevalent in transportation and infrastructure networks that are critical to society. By formulating it as a deterministic Markov decision process, we propose improvements over the standard UCT algorithm to solve this problem efficiently and enhance attack resilience in networks. Our approach achieves significant advancements in efficiency and scalability compared to previous reinforcement learning methods.
PROCEEDINGS OF THE ROYAL SOCIETY A-MATHEMATICAL PHYSICAL AND ENGINEERING SCIENCES
(2023)
Article
Computer Science, Artificial Intelligence
Mattia Crippa, Pier Luca Lanzi, Fabio Marocchi
Summary: The study applies an extension of Monte Carlo Tree Search called SP-MCTS to Sokoban, introducing two extensions to address challenges specific to the game. Evaluation of domain-independent enhancements shows some improvements in MCTS performance. Results suggest that while SP-MCTS can perform similarly to IDA* with more iterations, IDA* remains the better solver due to its ability to integrate domain knowledge.
EXPERT SYSTEMS WITH APPLICATIONS
(2022)
Article
Computer Science, Artificial Intelligence
Hok-Leung Kung, Shu-Jun Yang, Kuo-Chan Huang
Summary: Workflow computing is crucial in scientific and engineering fields, and workflow scheduling is a challenging research problem. This paper introduces a novel workflow scheduling approach based on Monte Carlo Tree Search (MCTS), which improves workflow execution schedules through the development of new mechanisms. Experimental results demonstrate significant performance advantages in terms of execution makespan compared to previous methods.
CONNECTION SCIENCE
(2022)
Article
Engineering, Civil
Shuojie Mo, Xiaofei Pei, Chaoxian Wu
Summary: Reinforcement learning has shown its decision-making ability in autonomous driving, but traditional training methods often lead to unsafe behaviors. This paper proposes a RL-based method combined with risk state estimation and safe policy search to reduce unsafe behaviors, resulting in faster convergence and safer behaviors in several random overtake scenarios.
IEEE TRANSACTIONS ON INTELLIGENT TRANSPORTATION SYSTEMS
(2022)
Article
Chemistry, Multidisciplinary
Okan Asik, Fatma Basak Aydemir, Huseyin Levent Akin
Summary: This paper proposes a new scalable planning approach by integrating decoupled planning and Monte Carlo Tree Search for cooperative multi-agent planning problems. The authors show the coordination and action synchronization problems caused by the separation of individual action updates and propose stochastic action selection policies and a combined method to address them. They empirically demonstrate the effectiveness of these methods.
APPLIED SCIENCES-BASEL
(2023)
Article
Computer Science, Information Systems
Shun-Chieh Chang
Summary: This study discusses various bandwidth allocation methods for network slicing and proposes a new algorithm based on Monte Carlo Tree Search to handle randomness and optimization. Simulation results show that the proposed algorithm improves the overall throughput by around 10% compared to previous algorithms, and it provides a better evaluation of the performance gap relative to the optimal solution.
Article
Computer Science, Information Systems
Xiao Fu, Hangyu Deng, Xin Yuan, Jinglu Hu
Summary: The study presents a feasible monophonic music generation framework that improves musical coherence by simulating subsequent trends for each predicted note. The framework includes three steps: predicting potential candidates, modeling and evaluating subsequent trends, and selecting the best candidate. Monte-Carlo tree search algorithm and smoothed polynomial upper confidence trees algorithm are used for accuracy and efficiency. The framework exhibits a better sense of musicality compared to note-by-note sequence prediction models.
IEEE TRANSACTIONS ON MULTIMEDIA
(2023)
Review
Computer Science, Artificial Intelligence
Maciej Swiechowski, Konrad Godlewski, Bartosz Sawicki, Jacek Mandziuk
Summary: Monte Carlo Tree Search (MCTS) is a powerful method for designing game-playing bots or solving sequential decision problems. It requires problem-dependent modification or integration with other techniques in more complex games and practical domains.
ARTIFICIAL INTELLIGENCE REVIEW
(2023)
Article
Automation & Control Systems
Qi Wang, Yongsheng Hao, Jie Cao
Summary: This paper introduces a novel framework called OmegaZero, based on Alphago Zero, which is trained through self-play without requiring expert experience or labeled data. The method is divided into two stages: the first stage learns node representations and memory history trajectories using GAT and GRU, while the second stage employs MCTS and deep RL to search for solution space and train the model.
ENGINEERING APPLICATIONS OF ARTIFICIAL INTELLIGENCE
(2021)
Article
Computer Science, Artificial Intelligence
Pengfei Ding, Guanfeng Liu, Yan Wang, Kai Zheng, Xiaofang Zhou
Summary: An attributed dynamic graph is a graph with multiple dynamic attributes associated with each edge. People can specify multiple constraints in the attributes to illustrate their requirements in attributed dynamic graph based applications. However, the multi-constrained temporal path discovery in such graphs is a challenging NP-complete problem. The existing methods adopt reinforcement learning with Monte Carlo tree search, but they are not applicable in real-time applications due to the expensive cost of query time and storage space. To address this issue, a new Adaptive Monte Carlo Tree Search algorithm (A-MCTS) is proposed, which dynamically adjusts the priority of historical records and utilizes an adaptive replay memory structure to improve the performance and reduce the required discovery experience.
IEEE TRANSACTIONS ON KNOWLEDGE AND DATA ENGINEERING
(2023)
Article
Biochemistry & Molecular Biology
Patrick Bryant, Arne Elofsson
JOURNAL OF MOLECULAR BIOLOGY
(2020)
Article
Multidisciplinary Sciences
Patrick Bryant, Arne Elofsson
Article
Biochemistry & Molecular Biology
Gabriele Pozzati, Petras Kundrotas, Arne Elofsson
Summary: This study tested different interface prediction methods for scoring docking solutions and found that contact-based interface prediction is the best method. The limitations of using interface predictions as constraints in a docking protocol were also discussed.
PROTEINS-STRUCTURE FUNCTION AND BIOINFORMATICS
(2022)
Correction
Multidisciplinary Sciences
Patrick Bryant, Gabriele Pozzati, Arne Elofsson
NATURE COMMUNICATIONS
(2022)
Article
Multidisciplinary Sciences
P. Bryant, G. Pozzati, A. Elofsson
Summary: Predicting the structure of protein complexes is extremely challenging. In this study, the authors utilize AlphaFold2 combined with optimized multiple sequence alignments to model interacting protein complexes, achieving state-of-the-art accuracy in predicting both the interaction status and mode of proteins.
NATURE COMMUNICATIONS
(2022)
Article
Biochemistry & Molecular Biology
David F. Burke, Patrick Bryant, Inigo Barrio-Hernandez, Danish Memon, Gabriele Pozzati, Aditi Shenoy, Wensi Zhu, Alistair S. Dunham, Pascal Albanese, Andrew Keller, Richard A. Scheltema, James E. Bruce, Alexander Leitner, Petras Kundrotas, Pedro Beltrao, Arne Elofsson
Summary: In this study, the authors investigate the use of AlphaFold2 to predict protein structures in the human protein-protein interactome and discuss its limitations. They demonstrate the high confidence of the predicted models and identify potential mechanisms for disease mutations. Additionally, they show the application of predicted binary complexes in expanding our understanding of human cell biology.
NATURE STRUCTURAL & MOLECULAR BIOLOGY
(2023)
Article
Biochemistry & Molecular Biology
Mehmet Akdel, Douglas E. Pires, Eduard Porta Pardo, Jurgen Janes, Arthur O. Zalevsky, Balint Meszaros, Patrick Bryant, Lydia L. Good, Roman A. Laskowski, Gabriele Pozzati, Aditi Shenoy, Wensi Zhu, Petras Kundrotas, Victoria Ruiz Serra, Carlos H. M. Rodrigues, Alistair S. Dunham, David Burke, Neera Borkakoti, Sameer Velankar, Adam Frost, Jerome Basquin, Kresten Lindorff-Larsen, Alex Bateman, Andrey Kajava, Alfonso Valencia, Sergey Ovchinnikov, Janani Durairaj, David B. Ascher, Janet M. Thornton, Norman E. Davey, Amelie Stein, Arne Elofsson, Tristan Croll, Pedro Beltrao
Summary: This study evaluates the performance of AlphaFold2 in structural biology applications and finds that it performs well and can partially replace experimentally determined structures, which is of great significance for life science research.
NATURE STRUCTURAL & MOLECULAR BIOLOGY
(2022)
Article
Genetics & Heredity
Patrick Bryant, Arne Elofsson
Summary: Changes in DNA methylation are strongly correlated with age, but there are non-linear relationships. Significant methylation changes in different tissues have different relationships with ageing. Assuming linear correlations may miss important biological markers.
NAR GENOMICS AND BIOINFORMATICS
(2022)
