efficient exploration
Recently Published Documents


TOTAL DOCUMENTS

235
(FIVE YEARS 96)

H-INDEX

22
(FIVE YEARS 6)

Author(s):  
Runjia Ji ◽  
Wensi Zhang ◽  
Yongxin Pan ◽  
Wei Lin

Magnetosome gene clusters (MGCs), which are responsible for magnetosome biosynthesis and organization in magnetotactic bacteria (MTB), are the key to deciphering the mechanisms and evolutionary origin of magnetoreception, organelle biogenesis, and intracellular biomineralization in bacteria. Here, we report the development of MagCluster, a Python stand-alone tool for efficient exploration of MGCs from large-scale (meta)genomic data.


2021 ◽  
Vol 2021 ◽  
pp. 1-15
Author(s):  
Xiaogang Ruan ◽  
Peng Li ◽  
Xiaoqing Zhu ◽  
Hejie Yu ◽  
Naigong Yu

Developing artificial intelligence (AI) agents is challenging for efficient exploration in visually rich and complex environments. In this study, we formulate the exploration question as a reinforcement learning problem and rely on intrinsic motivation to guide exploration behavior. Such intrinsic motivation is driven by curiosity and is calculated based on episode memory. To distribute the intrinsic motivation, we use a count-based method and temporal distance to generate it synchronously. We tested our approach in 3D maze-like environments and validated its performance in exploration tasks through extensive experiments. The experimental results show that our agent can learn exploration ability from raw sensory input and accomplish autonomous exploration across different mazes. In addition, the learned policy is not biased by stochastic objects. We also analyze the effects of different training methods and driving forces on exploration policy.


2021 ◽  
Vol 7 (1) ◽  
Author(s):  
Bowen Lei ◽  
Tanner Quinn Kirk ◽  
Anirban Bhattacharya ◽  
Debdeep Pati ◽  
Xiaoning Qian ◽  
...  

AbstractBayesian optimization (BO) is an indispensable tool to optimize objective functions that either do not have known functional forms or are expensive to evaluate. Currently, optimal experimental design is always conducted within the workflow of BO leading to more efficient exploration of the design space compared to traditional strategies. This can have a significant impact on modern scientific discovery, in particular autonomous materials discovery, which can be viewed as an optimization problem aimed at looking for the maximum (or minimum) point for the desired materials properties. The performance of BO-based experimental design depends not only on the adopted acquisition function but also on the surrogate models that help to approximate underlying objective functions. In this paper, we propose a fully autonomous experimental design framework that uses more adaptive and flexible Bayesian surrogate models in a BO procedure, namely Bayesian multivariate adaptive regression splines and Bayesian additive regression trees. They can overcome the weaknesses of widely used Gaussian process-based methods when faced with relatively high-dimensional design space or non-smooth patterns of objective functions. Both simulation studies and real-world materials science case studies demonstrate their enhanced search efficiency and robustness.


2021 ◽  
Author(s):  
Ali Abou Taka ◽  
Hector Corzo ◽  
Aurora Pribram-Jones ◽  
Hrant Hratchian

△SCF methods have proven to be reliable computational tools for the assignment and interpretation of photoelectron spectra of isolated molecules. These results have increased the interest in △SCF techniques for electronic excited states based on improved algorithms that prevent convergence to ground states. In this work, one of these △SCF improved algorithms is studied to demonstrate its ability to explore the molecular properties for excited states. Results from △SCF calculations for a set of representative molecules are compared with results obtained using time-dependent density functional theory and single substitution configuration interaction method. For the △SCF calculations, the efficacy of a spin-purification technique is explored to remedy some of the spin-contamination presented in some of the SCF solutions. The obtained results suggest that the proposed projection-based SCF scheme, in many cases, alleviates the spin--contamination present in the SCF single determinants, and provides a computational alternative for the efficient exploration of the vibrational properties of excited states molecules.


2021 ◽  
Author(s):  
Ruben Staub ◽  
Stephan Steinmann

Model Hamiltonians based on the so-called cluster expansion (CE), which consist of a linear fit of parameters corresponding to geometric patterns, provide an efficient and rigorous means to quickly evaluate the energy of diverse arrangements of adsorbate mixtures on reactive surfaces as typically relevant for heterogeneous catalysis. However, establishing the model Hamiltonian is a tedious task, requiring the construction and optimization of many geometries. Today, most of these geometries are constructed by hand, based on chemical intuition or random choices. Hence, the quality of the training set is unlikely to be optimal and its construction is not reproducible. Herein, we propose a reformulation of the construction of the training set as a strategy-based game, aiming at an efficient exploration of the relevant patterns constituting the model Hamiltonian. Based on this reformulation, we exploit a typical active learning solution for machine-learning such a strategy game: an upper confidence tree (UCT) based framework. However, in contrast to standard games, evaluating the true score is computationally expensive, as it requires a costly geometry optimization. Hence, we augment the UCT with a pre-exploration step inspired by the variance-based Design of Experiments (DoE) methods. This novel mixed UCT+DoE framework allows to automatically construct a well adapted training set, minimizing computational cost and user-intervention. As a proof of principle, we apply our UCT+DoE approach on the CO oxidation reaction on Pd(111), for which a relevant model Hamiltonian has been established previously. The results demonstrate the effectiveness of the custom built UCT and its significant benefits on a DoE-based approach.


2021 ◽  
Author(s):  
Divya Karade

Computer-aided drug design (CADD) techniques continue to struggle to provide a useful advance in the area of drug development due to the difficulties in an efficient exploration of the vast drug-like chemical space to uncover new chemical compounds with desired biological properties. Other challenges that users must overcome in order to fully use the potential of CADD tools and techniques include a lack of completely autonomous methods, the necessity for retraining even after deployment, and their lack of interpretability. To solve this issue, we created the ‘Custom ML Tools’ integrated within the framework of ‘AIDrugAPP’. ‘Custom ML Tools’ includes four modules: ‘Mol Identifier’, ‘DesCal’, ‘AutoDL’, and ‘Auto-Multi-ML’ which give users free access to molecular identification using SMILES and compound names, similarity search, descriptor calculation, the building of ML/DL QSAR models, and their usage in predicting new data. The study demonstrates the potential of the novel tool for computational investigations in drug discovery research. The WebApp with its modules has therefore been made available for public use at: https://sars-covid-app.herokuapp.com/


Author(s):  
J Víctor Moreno-Mayar

Abstract Present-day and ancient population genomic studies from different study organisms have rapidly become accessible to diverse research groups worldwide. Unfortunately, as datasets and analyses become more complex, researchers with less computational experience often miss their chance to analyse their own data. We introduce FrAnTK, a user-friendly toolkit for computation and visualisation of allele frequency-based statistics in ancient and present-day genome variation datasets. We provide fast, memory-efficient tools that allow the user to go from sequencing data to complex exploratory analyses and visual representations with minimal data manipulation. Its simple usage and low computational requirements make FrAnTK ideal for users that are less familiar with computer programming carrying out large-scale population studies.


Biomolecules ◽  
2021 ◽  
Vol 11 (10) ◽  
pp. 1505
Author(s):  
Zita Harmat ◽  
Dániel Dudola ◽  
Zoltán Gáspári

Ensemble-based structural modeling of flexible protein segments such as intrinsically disordered regions is a complex task often solved by selection of conformers from an initial pool based on their conformity to experimental data. However, the properties of the conformational pool are crucial, as the sampling of the conformational space should be sufficient and, in the optimal case, relatively uniform. In other words, the ideal sampling is both efficient and exhaustive. To achieve this, specialized tools are usually necessary, which might not be maintained in the long term, available on all platforms or flexible enough to be tweaked to individual needs. Here, we present an open-source and extendable pipeline to generate initial protein structure pools for use with selection-based tools to obtain ensemble models of flexible protein segments. Our method is implemented in Python and uses ChimeraX, Scwrl4, Gromacs and neighbor-dependent backbone distributions compiled and published previously by the Dunbrack lab. All these tools and data are publicly available and maintained. Our basic premise is that by using residue-specific, neighbor-dependent Ramachandran distributions, we can enhance the efficient exploration of the relevant region of the conformational space. We have also provided a straightforward way to bias the sampling towards specific conformations for selected residues by combining different conformational distributions. This allows the consideration of a priori known conformational preferences such as in the case of preformed structural elements. The open-source and modular nature of the pipeline allows easy adaptation for specific problems. We tested the pipeline on an intrinsically disordered segment of the protein Cd3ϵ and also a single-alpha helical (SAH) region by generating conformational pools and selecting ensembles matching experimental data using the CoNSEnsX+ server.


Sign in / Sign up

Export Citation Format

Share Document